JP2005121450A - Device for inspecting defect, and method therefor, and method of working inner face of cylindrical object - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an inspection method and an inspection device for inspecting quickly a defect such as a cavity, a flaw and a stain, even when a machining oil or a cleaning liquid is deposited on an inner face of an inspected cylindrical object such as an engine cylinder. <P>SOLUTION: This device/method is provided with a cleaning part 3 for cleaning a residual liquid left in the inner face of the cylindrical object 1, an optical system 4 for detecting a two-dimensional image of the inner face of the cylindrical object cleaned by the cleaning part, an image processing part 8 for acquiring information of the cavity defect, the defect in the vicinity of a contour in a worked hole, the worked flaw defect, the stain and the like, based on the two-dimensional image detected by the optical system, and an output part for outputting the information of the defect acquired by the image processing part. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention eliminates defects such as cast holes, scratches, dirt, etc. on the inner surface of a cylindrical object formed by casting or the like and then machined as necessary, such as engine cylinders and automobile running system parts. The present invention relates to a defect inspection apparatus and method for inspecting, and a cylindrical object inner surface processing method.

As conventional techniques for inspecting the damage of the inner surface of the cylindrical body, JP-A-3-226659 (Patent Document 1), JP-A-8-128960 (Patent Document 2) and JP-A-2002-22671 (Patent Document) Document 3) is known.
In Patent Literature 1, a mirror disposed inside a cylinder to be inspected, a one-dimensional CCD camera that captures an image by reflecting the inner surface of the cylinder with the mirror, and the one-dimensional CCD camera and the mirror are integrated. A cylindrical inner surface inspection apparatus is described that includes a rotational drive unit that rotates about a central axis of the cylinder.

Patent Document 2 includes a conical mirror body and a CCD camera, an image pickup unit that moves relative to the inside of the cylinder in the longitudinal direction and images the entire inner surface of the cylinder, and is transferred from the image pickup unit. There is described a cylindrical body inspection machine that includes an image input unit that develops a circular area image signal and an image processing unit that performs image processing on the image signal developed by the image input unit.
Patent Document 3 describes a cylindrical inner wall surface inspection apparatus including a conical mirror formed in multiple stages and an imaging camera that captures an optical image reflected by the multi-stage conical mirror.

JP-A-3-226659

JP-A-8-128960 JP 2002-22671 A

However, in the above prior arts 1 to 3, consideration is given to the inspection in the state where the processing oil is attached to the inner surface of the cylinder and the inspection of the defect in the state where the cleaning liquid is attached after cleaning the processing oil. Was not.
Moreover, the point which detects the defect around the hole processed into the cylindrical surface was not considered.
Moreover, the point which detects the processing flaw on the surface and the plating peeling applied to the inner surface of the cylinder has not been considered.
Further, it has not been considered to secure a stable machining state by outputting data for feeding back the machining state to the machining process based on the inspection result of the cylindrical inner surface.

  The first object of the present invention is to maintain constant detection conditions such as detection sensitivity even if processing oil or processing oil cleaning liquid adheres to the inner surface of a cylindrical object to be inspected such as an engine cylinder. Another object of the present invention is to provide a defect inspection apparatus and method capable of inspecting defects such as cast holes, scratches, and dirt at high speed.

  The second object of the present invention is to detect a defect around a hole caused by drilling and to detect detection conditions such as detection sensitivity even when the inner surface of a cylindrical object to be inspected such as an engine cylinder is drilled. It is an object of the present invention to provide a defect inspection apparatus and method capable of inspecting defects such as cast holes, scratches, and dirt at a high speed at a constant speed.

  In addition, a third object of the present invention is to be able to inspect defects such as scratches, dirt, plating peeling, etc. generated on the inner surface of a cylindrical object to be inspected at a high speed with a constant detection condition such as detection sensitivity. An object of the present invention is to provide a defect inspection apparatus and method therefor.

  In addition, the fourth object of the present invention is to provide stable processing by feeding back information on inspection results, such as castholes, scratches, dirt, plating peeling, etc., generated on the inner surface of a cylindrical object to be inspected. It is an object of the present invention to provide a defect inspection apparatus and method and a method for machining an inner surface of a cylindrical object that can ensure the state.

To achieve the above object, the present invention detects a cleaning unit that cleans residual liquid remaining on the inner surface of a cylindrical object to be inspected, and a two-dimensional image of the inner surface of the cylindrical object that is cleaned by the cleaning unit. An optical system, an image processing unit that acquires defect information based on a two-dimensional image detected by the optical system, and an output unit that outputs the defect information acquired by the image processing unit A defect inspection apparatus and method thereof.
Further, the present invention is characterized in that a control device for organically coupling and controlling the cleaning unit and the optical system is provided. Further, the present invention is characterized in that the cleaning unit is configured to eject a fluid (gas (gas) or liquid) radially onto the inner surface of the cylindrical object to be inspected.

  Further, the present invention is an optical system for detecting a two-dimensional image of the inner surface of a cylindrical object to be inspected whose inner surface is processed with a hole, and a two-dimensional image detected by the optical system. An image processing unit that acquires information on defects generated in the vicinity of the contour of the processed hole based on an image in the vicinity of the contour of the processed hole, and an output unit that outputs information on the defect acquired by the image processing unit A defect inspection apparatus and method thereof.

  The present invention also provides an optical system for detecting a two-dimensional image of the inner surface of a cylindrical object to be inspected, and a two-dimensional image detected by the optical system, which is divided into a plurality of regions, and for each of the divided regions. A defect inspection comprising: an image processing unit that acquires information on a processing flaw defect based on variation in brightness; and an output unit that outputs information on the processing flaw defect acquired by the image processing unit Apparatus and method thereof.

  Further, the present invention provides an optical system for detecting a two-dimensional image of the inner surface of a cylindrical object to be inspected with a hole processed on the inner surface, and a defect in a mold cavity and processing based on the two-dimensional image detected by the optical system. An image processing unit that classifies a defect near a hole and a processing flaw defect to acquire information on each defect, and an output unit that outputs information on each defect acquired by the image processing unit A defect inspection apparatus and method thereof.

  In the invention, it is preferable that the optical system illuminates a linear region facing the axial direction on the inner surface of the cylindrical object to be inspected, and a reflected light image from the linear region is the line. A reflective optical element that reflects a reflected light image from the linear region and directs it in the axial direction so as to detect from a direction inclined at a desired detection angle with respect to the normal direction of the linear region, and the reflective optical element An imaging optical element that linearly forms an image of reflected light from the linear region reflected in the axial direction and received by the imaging optical element and a linear reflected light image formed by the imaging optical element And a linear optical image sensor for outputting an image signal to provide a support member, and the linear image is obtained by rotating the support member around the axial center of the inner surface of the cylindrical object to be inspected. From the sensor to the inner surface of the cylindrical object to be examined Characterized by being configured to detect a two-dimensional image signal that.

  Further, the present invention is characterized in that the pre-processing unit of the image processing unit is configured to remove the cross-hatched mesh noise as the background noise by mask processing. According to the present invention, the image processing unit further extracts a defect candidate based on a two-dimensional digital image signal obtained from the preprocessing unit, calculates a feature amount of the extracted defect candidate, and calculates the calculation And a defect extraction processing unit that obtains defect information by making a determination based on the feature amount.

  The output unit may further include a display device that displays a defect map of a cylindrical object to be inspected as defect information acquired from the image processing unit. According to the present invention, in the output unit, any one of a projected view, a perspective view, a developed view, a three-dimensional solid view, and a plan view of a cylindrical object to be inspected as defect information acquired from the image processing unit. A display device that displays the above display screen and any one or more of the size, type, shape, and coordinates of the defect is provided.

  The present invention also provides a boring process for forming a bore of a cylindrical object by boring with a boring machine, a plating process for plating the bore inner surface of the cylindrical object formed in the boring process, and the plating A honing process for forming a honing eye by honing with a honing machine on the inner surface of the bore that has been plated in the processing process, and a first inner surface for cleaning the bore inner surface of the cylindrical object on which the honing eye has been formed in the honing process Cleaning step, a second cleaning step for cleaning residual liquid remaining on the bore inner surface of the cylindrical object cleaned in the first cleaning step, and a bore inner surface of the cylindrical object cleaned in the second cleaning step A two-dimensional image detected by an optical system, and an image process for acquiring defect information based on the two-dimensional image detected in the detection process. A defect inspection step having a process and an output process for outputting the defect information acquired in the image processing process, and the information on the defect output in the defect inspection process is the boring process Alternatively, the present invention provides a method for processing an inner surface of a cylindrical object, which is fed back to a honing process or the first cleaning process or the plating process.

As described above, according to the present invention, it is possible to automatically and stably detect defects such as a cast hole defect, a defect near the contour of a machining hole, and a machining scratch defect generated on a curved surface such as an inner wall of a cylindrical object. There is an effect that can.
In addition, according to the present invention, in the appearance inspection that automatically and stably detects defects such as a casting hole defect, a defect near the contour of a processing hole, a processing flaw defect, etc. generated on a curved surface such as an inner wall of a cylindrical object or the like. Because it is possible to obtain numerical information such as the coordinates (position), size, and area of the defect, it can be used for logistics management in the subsequent process and feedback of the process in the previous process, and manufacturing a high-quality product Can do.

  Embodiments of a defect inspection method and apparatus for an object surface such as a cylindrical inner surface and a processing method for a cylindrical inner surface according to the present invention will be described below with reference to the drawings.

  First, an embodiment of a defect inspection apparatus (appearance inspection apparatus) according to the present invention will be described. FIG. 1 is a diagram showing a schematic configuration of one embodiment, and the inspection object 1 is an inner surface of an engine cylinder as shown in FIG. In FIG. 2, an example of the appearance of an engine block having four cylinders is shown, but the number of cylinders per block is not limited as the inspection target 1. Further, a cylinder in which holes 13 are formed around the bottom dead center of the cylinder is also an inspection target.

  A cylinder 1 to be inspected (inspected cylinder object) 1 is mounted with a predetermined accuracy by a positioning pin (not shown) or the like on a work holder 2 that can reciprocate in the X1 and X2 directions based on control from the mechanism control unit 7. Placed. The cleaning unit 3 and the optical system 4 are held by a base 6 placed on a stage 5 and are organically connected and controlled by a mechanism control unit 7 so that they can reciprocate in the Z direction. .

  Then, the image detected by the optical system 4 is input to the image processing unit 8. The image processing unit 8 performs image processing on the input image, calculates a defect type, a defect size, coordinates, and the like and outputs them to the overall control unit 9. The overall control unit 9 determines pass / fail or classification based on the detection result of the defect type, defect size, coordinates, etc. output from the image processing unit 8 and displays the determination result on the display device 10 as well as the inspection. Data is input to the output device 11. The overall control unit 9 organically controls the mechanism control unit 7, the image processing unit 8, the display device 10, and the output device (including a network and a recording medium) 11. By the way, as an output part which outputs the information of the defect which concerns on this invention, the display apparatus 10 and the output device 11 shall be included.

  By the way, the defect inspection apparatus (appearance inspection apparatus) according to the present invention inspects defects on the inner surface of the bore of the cylinder to be inspected through the machining processes as shown in each of FIGS. become. In the case shown in FIG. 26, the inspected cylinder 1 to be inspected in the defect inspection process 202 boring the bore inner surface using a boring apparatus in the boring process 200, and then boring the inner surface in the honing process 201. The honing machine is processed using a honing machine. For this reason, in the defect inspection step 202, the inner surface of the bore of the cylinder 1 to be inspected, in which a small amount of processing fluid, chips, etc. remain in a number of islands due to surface tension, is inspected. In the case shown in FIGS. 27 and 28, the cylinder 1 to be inspected in the defect inspection process 202 is subjected to a honing process, and then a cleaning device for removing processing liquids, chips and the like in the cleaning process 203. It is what was washed using. Therefore, in the defect inspection step 202, the inner surface of the bore of the cylinder 1 to be inspected, in which a small amount of cleaning liquid or the like remains in a number of islands due to surface tension, is inspected. As described above, if a small amount of residual liquid such as a processing liquid or a cleaning liquid remains in an island shape due to surface tension, a shadow may be generated when obliquely irradiating light, which may cause erroneous detection. Therefore, the defect inspection apparatus (appearance inspection apparatus) according to the present invention is configured such that a trace amount of residual liquid is removed by the cleaning unit 3 before the optical system 4 detects an image of the inner surface of the bore.

  Next, the inspection operation in the defect inspection apparatus according to the present invention will be described in the case where the cylinder 1 to be inspected is composed of four bores No1 to No4. FIG. 1 shows a state in which the optical system 4 detects an image of the inner surface of the No. 1 bore and the cleaning unit 3 removes residual liquid from the inner surface of the No. 2 bore.

The bore inner surface is inspected according to the following procedure.
(1) The mechanism control unit 7 removes residual liquid on the inner surface of the No. 1 bore by the cleaning unit 3 while moving the Z stage 5 in the Z1 direction.
(2) Next, the mechanism control unit 7 moves the Z stage 5 in the Z2 direction to retract the cleaning unit 3.
(3) Next, the mechanism control unit 7 moves the work holder 2 in the X1 direction by the pitch distance P of the bore.
(4) Next, the mechanism control unit 7 removes residual liquid on the inner surface of the No. 2 bore by the cleaning unit 3 while moving the Z stage 5 in the Z1 direction.
(5) Next, the mechanism control unit 7 stops the Z stage 5 at a desired position in the Z1 direction, detects an image of the inner surface of the No. 1 bore with the optical system 4, and detects defects in the image processing unit 8 and the overall control unit 9. Perform an inspection.
(6) Next, the mechanism control unit 7 moves the Z stage 5 in the Z2 direction to retract the cleaning unit 3 and the optical system 4.
(7) Next, the mechanism control unit 7 moves the work holder 2 in the X1 direction by the pitch distance P of the bore.

Thereafter, the operations of (4) to (7) are sequentially repeated to inspect the four bores.
(8) After inspecting the four bores, the overall control unit 9 outputs the defect inspection result to the display device 10 and the inspection result data to the output device 11.

  Next, a specific embodiment of the cleaning unit 3 for cleaning the inner surface of the bore will be described with reference to FIG. The cleaning unit 3 includes a mechanism for injecting a fluid (a gas (for example, a gas such as air) or a liquid (for example, a volatile liquid)) onto the inner wall of the cylinder.

  FIGS. 3A and 3B show the implementation of a method of blowing a residual liquid by spraying a fluid (a gas (for example, a gas such as air) or a liquid (for example, a volatile liquid)) on the cylinder inner wall as the cleaning unit 3. An example is shown.

  In FIG. 3A, a fluid flow path 22 is formed in a shaft 21 that can reciprocate in the center of the inner wall of the cylinder in the direction of arrow Z, and fluid (for example, gas) is discharged from an ejection port 24 of a nozzle 23 formed in the shaft 21. A cleaning unit 3 is shown that ejects (injects) radially onto the inner wall of the cylinder and blows off the residual liquid on the inner wall of the cylinder. In addition, it is more preferable to incline the ejection direction of the fluid from the ejection port 24 of the nozzle 23 downward in the radial direction and in the gravitational direction because the residual liquid is blown off in the gravitational direction.

  FIG. 3B shows a nozzle 28 having a flow passage 27 formed in a shaft 26 that can reciprocate in the center of the inner wall of the cylinder in the direction of arrow Z and that is rotatably supported by a bearing 25, and has an ejection port 29. The cleaning unit 3 is configured by connecting the above to the flow path 27. Also in this embodiment, by rotating the shaft 26 and rotating the ejection port 29, the fluid is ejected (injected) radially onto the cylinder inner wall, and the residual liquid on the cylinder inner wall is blown away. Note that it is more preferable to incline the fluid ejection direction from the ejection port 29 of the nozzle 28 downward, which is the gravity direction, because the residual liquid is blown off in the gravity direction.

  FIG. 3C shows an embodiment of the cleaning unit 3 that removes residual liquid by wiping (sweeping) or sucking off. In this embodiment, a member 31 to be wiped or sucked is held by a holder 31 fixed to a shaft 30 capable of reciprocating in the center of the cylinder inner wall in the direction of arrow Z, and the member to be wiped or sucked member 32 is in contact with or close to the cylinder inner wall. The residual liquid on the inner wall of the cylinder is wiped off (sweeped) or sucked off. As the member (for example, brush) 32 to be wiped off, it is preferable to use a soft, water-absorbing cloth or resin so that the inner surface of the cylinder is not damaged.

  In addition, as another embodiment of the cleaning unit 3, a method of forming a transparent thin film made of processing oil with a uniform thickness on the entire inner wall of the cylinder is conceivable. In this embodiment, the chips are removed.

  Next, a specific embodiment of the defect inspection apparatus (excluding the cleaning unit 3) according to the present invention will be described with reference to FIG.

  Image detection is performed by a CCD linear image sensor 105 having, for example, 1024 pixels, and an imaging lens (imaging optical element) for forming an image of the linear image sensor 105 and the cylinder inner wall on the sensor surface of the linear image sensor 105. ) 104 generally cannot be accommodated in the cylinder. In this configuration example, a mirror (reflection optical system) 102 is provided in the cylinder 1 to be inspected, and the mirror 102 provides an optical path from the cylinder inner wall to the linear image sensor 105 by 90 degrees. It is configured to bend. An illumination optical system 103 including a light source illuminates the inner wall of the cylinder.

  The optical system 4 includes a support member 112 that integrally supports the illumination optical system 103 and the detection optical system 100, an XY fine movement stage 106 that attaches the support member 112 and finely moves it in the XY directions, and attaches the XY fine movement stage 106. The θ stage (rotary stage) 107 for rotating the support member 112, and the XYZ stage 108 that is attached to the θ stage 107 and whose position is adjusted with respect to the work holder 2.

  The detection optical system 100 includes a mirror (reflection optical element) 102, an imaging lens (imaging optical element) 104, and a CCD linear image sensor 105.

  With the above configuration, the detection optical system 100 including the illumination optical system 103, the mirror (reflection optical element) 102, the imaging lens (imaging optical element) 104, and the CCD linear image sensor 105 is integrally supported by the support member 112. The rotation is performed by a θ stage (rotary stage) 107 having the center of the cylinder circle as the center of rotation. The XY fine movement stage 106 between the θ stage 107 and the support member 112 adjusts the positional relationship between the θ stage 107 and the support member 112 as described later. The illumination optical system 103 and the detection optical system 100 are positioned with a structure (not shown) between the workpiece holder 2 on which the cylinder 1 to be inspected described above is placed, and adjust the positioning. There is an XYZ stage 108. In particular, the Z stage moves in the length direction of the cylinder 1 to be inspected (the vertical direction in FIG. 1), and the detection optical system 100 is taken in and out of the cylinder to be inspected, or the detection system as shown in FIG. When the visual field width is shorter than the cylinder length, it is sequentially moved in the Z direction to detect the entire inner wall of the cylinder 1 to be inspected or to position it at a necessary place.

  Although details will be described later, by rotating the illumination optical system 103 and the detection optical system 100 described above around the θ stage 107, the linear image sensor 105 can detect an image of one round of the cylinder inner wall. Note that the pixel size on the inner wall of the cylinder 1 in the alignment direction of the linear image sensors 105 is about 50 to 80 μm. In accordance with this, the imaging magnification of the imaging lens 104 is set.

  Further, the number of pixels for one round of the cylinder inner wall needs to be raised to a power of 2 (for example, 4096) by using a two-dimensional Fourier transform as described later. Therefore, the rotational speed of the θ stage 107 is determined.

  Thus, the detected image signal 131 for one round of the cylinder inner wall detected from the linear image sensor 105 is converted into the digital image signal 132 by the A / D converter 121, and the pre-processing unit 122 performs shading correction and brightness correction. Then, background noise removal processing is performed and the two-dimensional digital image 133 is stored in the memory 123. The defect extraction processing unit 124 performs binarization processing, for example, based on the two-dimensional image data 134 read from the memory 123 to extract defect candidates such as a cast hole, a flaw, and dirt, and the extracted cast hole, Defect feature values for defect candidates such as scratches and dirt are calculated, and the defect information (defect occurrence coordinates, defect feature values, defect count (inspected) is determined based on the calculated defect feature values. (Including the distribution of inner wall defects for each cylinder 1) and defect image signals, etc.) 135 can be obtained, input to the main control unit 125 (8, 9), stored in the storage device 143, etc., and output. . As described above, the A / D converter 121, the preprocessing unit 122, the memory 123, and the defect extraction processing unit 124 constitute the image processing unit 8.

  The main control unit 125 includes a display device 10, an input unit 142, a storage device 143, and an output device (including a recording medium and a network) 11. Based on the control signal 138, the linear image sensor 105, A / D converter The illumination optical system 103 attached to the support member 112 is controlled by controlling the θ stage (rotation stage) 107 based on the θ stage control signal 136 by controlling the 121 preprocessing unit 122, the memory 123, the defect extraction processing unit 124, and the like. In addition, the position of the detection optical system 100 in the cylinder inner circumferential direction is controlled, and the XYZ stage 108 is controlled based on the detection system positioning control signal 137 to control the position of the θ stage 107 in the XYZ direction with respect to the cylinder 1 to be inspected. It is.

  In particular, the main controller 125 (7, 9) controls the XY stage 108 to set the rotation center of the θ stage 107 to the cylinder 1 to be inspected in order to make the sensitivity uniform over the entire inner wall of the cylinder 1 to be inspected. It is necessary to match the center of the inner wall (center of the inner diameter). Further, the main controller 125 (7, 9) controls the XY fine movement stage 106 to change the distance L (that is, the detection angle θr) from the inner wall center (inner diameter center) of the cylinder 1 to be inspected to the detection point 151. Adjustments can be made.

  Further, the main control unit 125 can control the brightness (intensity) of the illumination light emitted from the illumination optical system 103 including the illumination light source by controlling the illumination power source 126 and the like based on the illumination control signal 139. Become. The illumination light emitted from the illumination optical system 103 may be a slit illumination light beam.

  Hereinafter, after describing the details and operations of each unit, the overall flow will be described.

  FIG. 5 shows the detection optical system 100 and the illumination optical system 103 mounted on the support member 112 supported on the θ stage 107 via the XY fine movement stage 106 in the present embodiment described above. It is. In this embodiment, since a CCD linear image sensor is used as the image detection sensor, when the detection optical system 100 is not rotated by the θ stage 107, a linear area 151 shown in FIG. 5B is a detection area. The image signal 131 detected from the linear image sensor 105 in synchronization with the rotation of the detection optical system 100 by the θ stage 107 is digitized by the A / D converter 121 and stored in the memory 122. An image of one round of the inner wall of the cylinder 1 to be inspected as shown in FIG. 5 (b) can be detected. Here, FIG. 6 shows an example of the positional relationship between the detection point 151 by the linear image sensor 105 and the illumination optical system 103 that illuminates the detection point 151. FIG. 6 is a top view of the mirror 102 and the cylinder 1 to be inspected of the detection optical system 100 of this embodiment.

  In this embodiment, the detection position 151 by the linear image sensor, that is, the linear area 151 shown in FIG. 5B is set to a position separated from the center line 152 of the cylinder by a distance L. The illumination optical system 103 emits light in a planar shape or a line shape, and is arranged so as to illuminate the detection position 151 of the linear image sensor. In general, the arrangement shown in FIG. 6 is desirable.

  Here, the reason why the linear detection area 151 of the linear image sensor 105 is set at a position separated from the center line 152 of the cylinder by a distance L will be described. Note that the optical axis of the imaging lens (imaging optical element) 104 exists on 159 shown in FIG. That is, the imaging lens 104 is configured to form on the linear image sensor 105 a light image of the linear region 151 shown in FIG. 5B obtained by being reflected by the mirror (reflection optical element) 102. The Therefore, in particular, the imaging lens 104 has a characteristic of focusing in a uniaxial direction like a cylindrical lens. By the way, the defect detected by this inspection device is honing when there is a hole such as a nest generated at the time of casting, when the surface is recessed or turned up like a scratch, or when the cylinder surface is plated. Plating peeling due to eccentricity during processing. In addition, in order to improve the combustion efficiency of the engine, there are burrs, processing flaws, etc. that occur around the hole processed in the side surface of the cylinder provided near the bottom dead center when the piston is driven. That is, the surface irregularities are basically detected.

  When detecting unevenness on the surface, according to oblique illumination / diagonal detection, the normal part is detected brightly, the uneven defect part is detected darkly, and as a result, the uneven defect part can be detected with high sensitivity. It becomes possible. That is, as shown in FIG. 7, when detecting an object with a hole 157 opened in the inner wall of the cylinder 1 to be inspected, such as a cast hole, the surface of the inner wall of the cylinder 1 to be inspected is inclined from its normal line 154. Illuminating from the direction tilted (θi) (obliquely illuminating), and similarly detecting from the direction tilted obliquely (θr) (detecting obliquely), the defect 157 in the cast hole is detected by the linear image sensor 105. As a result, a shadow (dark part) detection range 158 is detected as dark. This is because a shadow portion due to illumination and a shadow portion occur when detecting. Here, the defect detection sensitivity can be changed by changing the angles θi and θr from the normal lines 154 of the illumination optical system 103 and the linear image sensor 105. For example, the angles θi and θr from the normal line can be changed. By increasing the size, it is possible to detect even a hole having a shallow hole depth and a large hole diameter. Conversely, by decreasing the angles θi and θr, the hole depth is deep and the hole diameter is small. Only detection can be made.

  When the object to be inspected is a cylindrical inner wall as in this embodiment, the linear region 151 that is the detection region of the linear image sensor 105 is separated from the center line 152 of the cylinder by a distance L as shown in FIG. By setting the position, the linear detection region 151 is detected obliquely, and the angle can be obtained from the relationship between the distance L and the diameter R of the cylinder inner wall. That is, the angle θr from the normal line on the detection side can be obtained from the following equation (1).

θr = sin ⁻¹ (2L / R) (1)
Here, R is the diameter of the inner wall of the cylinder, and L is the distance from the center line 152 of the cylinder in the linear detection region 151 of the linear image sensor 105. From this, the main control unit 125 obtains and verifies the defect detection condition (especially the detection angle θr) with a certain inspected cylinder inner diameter, and even if the inner diameter of the inspected cylinder 1 changes, the same value can be obtained. Detection conditions (particularly the same detection angle θr) can be obtained. As a result, the main control unit 125 can control the XY fine movement stage 106 so that the distance L from the rotation center of the θ stage 107 is suitable for the inner diameter R of the cylinder 1 to be inspected. The main control unit 125 can also change the angle θr from the normal line on the detection side by controlling the XY fine movement stage 106 and changing the distance L from the rotation center of the θ stage 107. Of course, if a fine movement mechanism is provided between the detection optical system 100 and the support member 112, the detection optical system 100 can be finely adjusted with respect to the support member 112.

  Here, an embodiment of the optimum distance L is shown in FIG. FIG. 8 shows a case where an image is detected by the linear image sensor 105 by changing the distance L and background noise removal processing described later is performed. It is the experimental result calculated | required about contrast. As shown in FIG. 8, the brightness variation of the background becomes low at a detection angle θr of the detection optical system 100 of about 15 degrees (distance L / diameter R ratio is about 12%). On the other hand, depending on the shape of the defect, there are cases where the contrast of the defect portion is larger when the detection angle θr of the detection optical system 100 is smaller, and the contrast tends to be opposite. The detection angle θr is about 15 degrees. From the above results, it is desirable that the detection angle θr of the detection optical system 100 is about 15 degrees, that is, the distance L / diameter R ratio is about 12%. In addition, since optical parts such as illumination have errors in the spread angle, the detection angle θr of the detection optical system 100 is about 15 ° ± 5 ° (distance L / diameter R ratio is 8% to 17%). Configure to be able to adjust with. That is, if the inner diameter R of the cylinder 1 to be inspected is input using the input unit 142 or the like, the main control unit 125 causes the mirror (reflection optical element) 102, the imaging lens (imaging optical element) 104, and the linear image sensor 105 to operate. The XY fine movement stage 106 is driven and controlled such that the distance L from the rotation center of the θ stage 107 is controlled so that the detection angle θr of the detection optical system 100 is about 15 ° ± 5 °.

  As described above, by providing the XY fine movement stage 106 between the support member 112 that supports the detection optical system 100 and the θ stage 107, the distance L from the rotation center of the θ stage 107 is set to the cylinder 1 to be inspected. As a result, even if conditions such as the size of the cylindrical object to be inspected change, defects around the hole caused by drilling can be corrected as will be described later. In addition to detection, it is possible to inspect defects such as cast holes, scratches, and dirt at a high speed by making detection conditions such as detection sensitivity constant.

  At this time, since the illumination optical system 103 is also mounted on the support member 112, the illumination optical system 103 moves together with the detection optical system 100 when the XY fine movement stage 106 is moved. If an XY fine movement stage (not shown) is provided between the illumination optical system 103 and the support member 112, the irradiation angle θi by the illumination optical system 103 can be adjusted independently of the detection angle θr by the detection optical system 100. It becomes possible.

  When the detection optical system 100 arranged as described above is rotated by rotating the θ stage 107, it is detected by the linear image sensor 105 and is two-dimensionally shown in FIG. 9A from the A / D converter 121. An image 161 is obtained. If the center of the inner diameter of the cylinder 1 to be inspected and the rotation center of the θ stage 107 (the rotation center of the detection optical system 100) are not shifted, the brightness 162b of the obtained image is substantially uniform as shown in FIG. However, if it is deviated (eccentric), uneven brightness (brightness) in the circumferential direction of the inner wall occurs as indicated by 162a. Here, the projection waveform shown in FIG. 9B is one-dimensional data obtained by calculating the cumulative value of the brightness of each Y coordinate pixel, and is a global one in the Y direction (inner wall circumferential direction) of the image. You can see changes in brightness. If the center of the inner diameter of the cylinder does not match the rotation center of the detection optical system 100, (R / 2) in the previous equation (1) changes due to the rotation of the detection optical system 100. This is because the angle θr changes. This is because the illumination 171 by the illumination optical system 103 has directivity as shown in FIG. 10, and when θr changes, the angle component of the illumination directivity changes. Further, when θr changes, the defect detection sensitivity also changes as described above. Therefore, θr is preferably constant over the entire inner wall of the cylinder. That is, if the brightness of the entire image is constant, the projection waveform is constant as shown by 162b. If there is a certain level of brightness or darkness in the circumferential direction of the image as in 162a, the position of the cylinder circle can be determined from the coordinates in the light and dark image. It can be seen in which XY direction the θ stage 107 is corrected. Further, the correction amount can be set in advance by obtaining the deviation amount and the uneven brightness amount.

  Therefore, first, it is necessary to align the rotation center of the θ stage 107 with the inner diameter center of the cylinder 1 to be inspected positioned on the work holder 2. Therefore, when the cylinder 1 to be inspected is placed on the work holder 2, it is possible to align the center of the inner diameter of the cylinder 1 to be inspected with the rotation center of the θ stage 107 by mechanically positioning with a positioning pin or the like. . However, if this is insufficient, for example, a distance sensor (not shown) for measuring the distance from the inner wall of the cylinder is mounted on the support member 112, and the main control unit 125 rotates the θ stage 107 to From the distance detected by the distance sensor, the amount of eccentricity of the rotation center of the θ stage 107 with respect to the center of the inner diameter of the cylinder 1 to be inspected is obtained, and this eccentricity is fed back to the XY stage 108 to rotate the detection optical system 100. The center can be matched with the center of the inner diameter of the cylinder 1 to be inspected. Further, the main control unit 125 is pre-processed by the one-dimensional two-dimensional image 161 or the pre-processing unit 122 obtained by rotating the θ stage 107 and starting from the A / D converter 121 before starting the actual inspection. Based on the one-dimensional two-dimensional image 161 stored in the memory 123, as shown in FIG. 9B, the cumulative value of the brightness of the pixels across the linear image sensor at each Y coordinate is calculated. It is possible to check whether or not the calculated cumulative value is within an allowable range. If it is not within the allowable range, the XY stage 108 is controlled so that the calculated cumulative value falls within the allowable range.

  Thus, preparation before the start of inspection is completed.

  Next, the actual inspection will be started. By applying oblique illumination to the inner wall of the cylinder 1 to be inspected from the illumination optical system 103, the linear image sensor 105 detects the light image of the linear region 151 shown in FIG. 5B by a detection angle θr (= 15 degrees). The image is reflected by the mirror 102 and condensed by the imaging lens 104, particularly in the direction perpendicular to the longitudinal direction of the image sensor, and detected as a one-dimensional image 131. Then, by rotating the θ stage 107, the two-dimensional image signal 161 shown in FIG. 9A is detected from the linear image sensor 105, and the two-dimensional digital image signal 132 is obtained by the A / D converter 121. become. 163b is an enlarged view of the two-dimensional digital image signal in the region 163a. Reference numerals 165a and 165b denote defects having different sizes. Reference numeral 164 denotes a honing mark that has been honed for the purpose of lubricating the inner wall of a cylinder of an engine or the like.

  Processing for extracting defects from the two-dimensional digital image obtained as described above will be described with reference to FIG. FIG. 9A shows an example of the detected two-dimensional digital image 161. Here, a description will be given assuming that a defective casting hole is detected darkly. Of course, it is needless to say that even a defect that is detected brighter than the surroundings, such as a convex defect, can be easily adapted from the following description.

  FIG. 11 shows an example of the flow of image processing such as defect extraction processing. First, the two-dimensional digital image signal 161 (132) shown in FIG. 9A is detected from the A / D converter 121 (step S901). Next, shading correction is performed in the preprocessing unit 122 as necessary (S902). The shading correction is uneven brightness of an image caused by uneven brightness of the illumination optical system 103 and the detection optical system 100. This is corrected from reference data obtained in advance or the detected image signal 161 (132). For example, when correcting from the detected image signal 161 (132), the projection waveform shown in FIG. 9B described above is obtained in the X direction and the Y direction, and each pixel is determined for each direction from the projection waveform. By normalizing, global brightness unevenness can be corrected and eliminated.

  FIG. 9A shows an enlarged 163b of the detected image 163a, and regular diagonal lines are detected. This is a honing mark 164 and not a defect. In general, the inner wall of a cylinder of an engine or the like may form innumerable scratches 164 in a cross-hatched shape as shown in FIG. 9A by honing for the purpose of lubrication or the like. As a result, linear unevenness is generated, and light and dark are produced in the same manner as the defect portion, and are detected as an oblique line 164 as shown in FIG. For this reason, in order to extract the defective portions 165a and 165b from the detected image 163b, a dark portion, that is, a pixel having brightness below a certain threshold value is extracted (so-called binarization). 164 may also be extracted. Therefore, the pre-processing unit 122 performs a process of removing the diagonal line, so-called background noise (step S903). This background noise is characterized in that the diagonal lines 164 are periodic.

  In general, as a method for removing a periodic pattern in an image, for example, there is a method using a Fourier transform as shown in “Supervision by Tamura Hideyuki, Introduction to Computer Image Processing, pages 143 to 146, published by Soken Publishing”. . In this method, as shown in FIG. 9, first, a detection image 161 (132) is two-dimensionally Fourier-transformed to generate an image on the frequency plane (step S9031). The data of the corresponding place is masked (step S9032), and the image of the frequency plane is returned to the normal plane image again by the two-dimensional inverse Fourier transform (step S9033). By doing so, an image from which background noise is removed can be obtained.

  Here, as a method of creating a mask image, for example, an absolute value image of an amplitude is generated from a two-dimensional Fourier transform image (real part and imaginary part data), and the absolute value image of the amplitude is binarized with an appropriate threshold value. It can also be based on what has been done. Further, since the angle of the mesh of the cross hatch is determined, the mask shape in the Fourier transform image can be obtained from the angle. For example, as shown in FIG. 12, an original image is displayed, and two points of a scratch line on the screen are designated for each cross line (a pair of 180a and 180b and a pair of 181a and 181b), or a dotted line 182 As shown, specify an acceptable range of angles. Moreover, you may convert into the angle of the cross line on the detection image obtained from settings, such as a honing machine. At this time, the allowable range of the angle may be determined in consideration of an error of a processing machine or the like and an angle allowed when the inspection object is actually used.

  From these coordinate data, a mask in a two-dimensional Fourier transform image can be generated by calculation. In this way, instead of obtaining the mask shape from the obtained two-dimensional Fourier transform image, the mask shape is determined from the detected image or processing conditions, so that errors or mistakes during processing deviate from the condition angle during processing. Mask processing can be performed excluding the portion. As a result, the processing flaw deviating from the desired condition with respect to the original image shown in FIG. 13A can be detected by a subsequent defect detection process without being removed as shown in FIG. 13B. The result is that the quality of the wound can also be evaluated. FIG. 13B shows a case where the main controller 125 displays an image of a defect from which background noise has been removed on the screen of the display device 41.

  As described above, the two-dimensional image 133 from which the shading correction is performed and the background noise is removed is stored in the memory 123 from the preprocessing unit 122. Next, the defect extraction processing unit 124 can extract a defect candidate by binarizing the image signal obtained by removing background noise and the like obtained from the memory 123 with an appropriate threshold value (step S904). ). The defect extraction processing unit 124 further selects and extracts feature amounts such as coordinates, area, and size (maximum diameter, etc.) of each candidate point from the binarized image. A defect that is larger than a predetermined size standard, for example, the size of the defect is determined as a defect (step S905). Further, for example, as shown in FIG. 14, the defect extraction processing unit 124 selects an area of a defect candidate part or a diameter S of an ellipse approximation at a predetermined threshold Vth. The ratio (Vmean / S) of the average brightness Vmean of the defect candidate part is used, or the brightness V (CIR) mean around the defect candidate part and the darkest (or brightest) brightness of the defect candidate part The difference (V (CIR) mean−Vmin) or ratio (Vmin / V (CIR) mean) with respect to Vmin (or Vmax) is extracted to obtain the maximum brightness level B in the defect candidate portion, and this maximum brightness level B And the ratio (B / S) of the area of the defect candidate part or the diameter S of the elliptical approximation. Thus, by using the ratio (Vmean / S, B / S) between the area or diameter S of the defect candidate portion and the gray level (gradation value) level (Vmean, B) of the defect candidate portion as the evaluation value, the defect It is possible to select a deep hole-like cast hole defect having a small area and a small average value or minimum value of defect density values from defect candidate points. Then, the defect extraction processing unit 124 outputs defect information 135 (defect occurrence coordinates, defect feature amount, number of defects, defect map for each cylinder to be inspected, defect image signal, etc.) 135 for the determined defect. Are input to the main control unit 125 and stored in the storage device 143 or the like. The main control unit 125 can display the defect information 135 and the evaluation result (pass / fail judgment result) on the display device 141 (step S906). When the defect information 135 is displayed, the detection angle θr can be displayed at the same time.

  By the way, as a specific method for realizing the memory 123 and the defect extraction processing unit 124 shown in FIG. 1, a high-speed computer equipped with an image input board for inputting an image signal from the linear image sensor 105, such as a personal computer, is used. be able to. Of course, it is not deviated from the present invention to construct the whole or a part by a suitable dedicated or general-purpose hardware processing system.

  Next, an inspection method in the case where the hole processing 13 is performed on the side surface of the cylinder shown in FIG. 2 will be described with reference to FIGS. 15, 16, and 17.

  In order to improve the efficiency of driving the piston, a hole may be provided near the bottom dead center of the piston. When drilling holes from the side of the piston, defects such as burrs and processing flaws may occur around the holes, and these burrs need to be inspected. A method for detecting burrs and scratches around the hole caused by this processing will be described. The image is detected by the method shown in FIGS. Reference numeral 190 in FIG. 15 is a two-dimensional image of the cylinder 1 to be inspected when the hole is machined. Since reflected light is not detected from the machining hole, the machining hole image 191 is black and the cylinder side image 192 is a bright image. FIG. 15B shows the enlargement of the hole. A burr 194 is detected around the periphery 193 of the hole image 191. The burr 194 is also recessed like the case of FIG. 7, so that the defect image is detected darker (darker) than the surrounding 193. A process of extracting defects from the two-dimensional digital image obtained as described above in the defect extraction processing unit 124 will be described with reference to FIGS. 16 and 17.

  After image detection (step S910 (including image detection S901, shading / brightness correction S902, and background noise removal S903)), a hole image area is cut out based on pre-given processing hole position information (step (S911) A binarization process is performed (step S912), and brightness pixels below a certain threshold are extracted as processing hole candidates. From the detected binary image (FIG. 17A), a candidate point for the machining hole edge is extracted (step S913). The machining hole edge is searched for in the x and y directions of the image from the approximate center coordinates of the machining hole (position information is obtained in advance from the design data), and is first white (“1” (white), “ Points that are 1) pixels of the value 0 ″ (black) are sequentially searched and extracted as point coordinates 195 (FIG. 17B). Next, ellipse approximation is performed based on the point coordinates 195 to extract a hole shape (hole outline) 196 (step S914).

  The method of extracting the hole shape (hole outline) 196 by the ellipse approximation is, for example, “Automatic extraction of circular objects by the ellipse growth method, Mitsuo Okabe, D-II Vol. J85-D-II No. 12 PP. 1823-1831 December 2002 ", there is an apparatus that detects an edge and repeats ellipse fitting to extract an optimal ellipse (circular shape). By this method, the processed hole shape 196 can be accurately extracted without being affected by burrs or chips generated around the hole edge (FIG. 17C). In this example, the method of extracting the outer shape of the processed hole by ellipse approximation was described as the method of extracting the processed hole, but in this example, the processed hole is displaced from the design value or the processed hole is deformed. However, the feature is that the edge of the machined hole can be extracted accurately. When the displacement is small in the processed hole and the deformation is small enough not to affect the defect detection, the same effect can be obtained even if the processed hole outline is extracted based on the design value. Next, as shown in FIG. 17D, in order to make the inside 197 of the processed hole a non-inspection area (not to make the inside of the processed hole an inspection target area), an image (shading, An embedding process is performed in the processed hole (processed hole) with the brightness corrected image (step S915). For embedding, an embedding area of the image is determined based on the circle and the center coordinates obtained by the ellipse approximation in step S914. A defect candidate shown in FIG. 17E is set by setting a defect detection area composed of a region near the contour of the processed hole after image embedding (for example, an annular region having a width W) (step S916) and performing binarization processing. A point 199 is extracted (step S917). Features such as coordinates, area, size, and the like of the defect candidate points are selected and extracted to extract defects (burrs and scratches) 194 near the contour of the processed hole (step 918). Then, information on defects (burrs and scratches) in the vicinity of the extracted contour of the processed hole is output and displayed (S906). In addition, in order to distinguish between defects (burrs and scratches) in the vicinity of the contour of the casting hole and the machining hole, the size (defect length) based on the distance ΔK from the machining hole edge, which is one of the feature quantities of the defect candidate points. ) May be used to judge and identify defects (burrs and scratches). For example, as shown in FIG. 17F, it is possible to determine that the defect length ΔK is a non-defective product when ΔK ≦ T with respect to a certain distance T, and a defect when ΔK> T. Further, by extracting the inside of the contour vicinity region (for example, the ring-shaped region having the width W) away from the outer periphery of the processing hole as the defect near the contour of the processing hole, it is possible to identify the defect as a cast hole defect. In addition, although the round hole was demonstrated as a shape of the said process hole, a semicircle hole or a square hole may be sufficient. Naturally, when extracting a semi-circular or square hole shape (hole outline) based on the above point coordinates, there is a portion for performing linear approximation.

  Next, an embodiment in which the processing flaw defect shown in FIG. Types of processing flaw defects include cylinders such as rough surface processing (honed eyes are not processed correctly) and plating peeling (plating treatment to harden the cylinder surface made of aluminum alloy etc.) Occurs when boring or honing the inner wall. Processing for extracting defects such as rough processing and peeling of plating will be described with reference to FIGS. A process different from the defect extraction of the cast hole and the processing flaw shown in FIG. 11 will be described. After the shading correction and brightness correction (S902) are performed on the detected two-dimensional digital image (S901), the background noise of the honing trace is removed (S903). Next, in the defect extraction processing unit 124 and the like, the defect detection evaluation area 200 is determined by dividing the detection region (S920). Processing roughness and plating peeling occur due to eccentricity of the blade and cylinder during processing, wear of the blade, and the like, and thus occur in a relatively wide area. Therefore, it is preferable to set the defect detection evaluation area 200 to an area of about 10 mm or more. Variations in image brightness (tone value or gray value) are extracted from the divided image area 200 (S921). This is done by extracting the standard deviation (Dev) and the average brightness (Ave) of the image brightness in each region 200 and using the ratio (Dev) / (Ave) as an evaluation value as a feature amount. , Image variations can be extracted. In the case where the brightness variation value of the image is larger than a certain value, it is extracted as a defective area where there is a rough machining or plating stripped portion 201 that is a defect of a machining flaw (S921). Naturally, there is no defect in the processing flaw in the area where the brightness variation value of the image is detected smaller than a certain value. Then, the defect information on the extracted processing roughness or processing flaw as a plating peeling portion is output and displayed (S906). FIG. 19 shows a case where an image of a defect is displayed on the screen of the display device 41, and the detected image 200 can be divided to show a rough processing or plating stripped part 201 and a non-defective part 202.

  20 and 21 show a case where a defect is detected by identifying each of a defect in the vicinity of a cast hole, a processed hole, a rough processing, and a plating peeling defect. After the two-dimensional digital image detection (S930 (S901 to S903)) in FIG. 20, the inspection area is cut out (S931). The inspection area is determined by the detection pixel size and the diameter R and length of the cylinder inner surface. For example, when the diameter R = Φ70 mm and the length 50 mm, the area where the inner diameter is developed is 219.9 mm × 50 mm. If the detection pixel size is approximately 0.1 mm, almost the entire surface can be detected in the inspection area of 2048 pixels × 512 pixels, and the inspection process is performed in this area. FIG. 16 shows the first processing hole defect detection process (S932 (S911 to S917)) and feature amount extraction / selection (S933 (S918)) for defect candidate points detected by the processing hole defect detection process. Follow the procedure to extract defects (burrs and scratches) near the contour of the processed hole. Then, the defect extraction / selection (S935 (S905)) for the defect candidate point detected in the casting hole defect inspection process (S934 (S904)) and the defect defect inspection process is performed according to the procedure shown in FIG. Extract defects in the casting cavity. In this embodiment, if there is a machined hole, the machined hole is embedded in the machined hole defect detection process, so that the machined hole is not extracted as a defect during the hole defect inspection process, and high-precision inspection is realized. it can. After extracting the voids (S935 (S905)), the detection area is divided, and the standard deviation (Dev) of the image brightness and the average brightness (Ave) of the image are extracted for each divided area (processing) Roughness and plating peeling detection process) (S936 (S920 to S921)) is performed, and as a result, the brightness variation of the image ((Dev) / (Ave)) is extracted (S937 (S921)). Based on the result, information on the defect area where the processing flaw exists is output and displayed (S906).

  FIG. 21 shows a functional block diagram of the entire defect inspection apparatus. After the cylinder inner surface is cleaned by the cleaning unit 3, an image of the cylinder inner surface is detected by the optical system 4, and the functional image processing unit 8a in the image processing unit 8 is detected. In each of 8b and 8c, casting hole inspection, defect inspection near the processing hole, processing roughness, plating peeling defect are detected, defect position specification and distribution are extracted, and defect is determined, displayed, stored, and result Can be output as an attachment. Further, the detected results are input to the analysis device 9a of the overall control unit 9, and the defect frequencies and types are tabulated and analyzed for each day and each part, and the pre-processes 201, 203, and 204 shown in FIG. 26, FIG. 27, and FIG. It is possible to feed back to the processing machine and design information.

  Next, the overall operation of the optical system 4 will be described with reference to FIG. The case of inspecting an engine block having four inspected cylinders 1 as shown in FIG. 2 will be described. First, the engine block 12 to be inspected is placed on the work holder 2. At this time, the engine block to be inspected is placed with a predetermined accuracy by a positioning pin (not shown) on the work holder 2 (not shown), or the position thereof is measured by an image sensor (not shown), The amount of deviation from the normal position is fed back to the XY axes of the XYZ stage 108 as appropriate.

  Next, the Z axis of the XYZ stage 108 is moved, a part of the support member 112 to which the detection optical system 100 and the illumination optical system 103 are attached is inserted into the cylinder 1 to be inspected of the engine block to be inspected, and the θ stage 107 is set to 1 By rotating, an image for one round of the inner wall of the cylinder is detected by the linear image sensor 105, converted to a digital image signal by the A / D converter 121, preprocessed, and stored in the memory 123. At this time, since the θ stage 107 needs to be accelerated and decelerated, it may actually rotate one or more times. The image stored in the memory 123 is subjected to the processing described in detail above by the defect extraction processing unit 124 and the defect information 135 is extracted. When the length of the cylinder is longer than one image detection width, the Z axis of the XYZ stage 108 is moved, and image detection and defect extraction processing are repeated for the undetected portion. In this case, if image detection by the optical system 4 on the inner surface of each cylinder is slightly overlapped, a defect present at the boundary of image detection is not overlooked.

  When the inspection for one cylinder is completed as described above, the Z axis of the XYZ stage 108 is moved, the detection optical system 100 is extracted from the cylinder, and the XY stage of the XYZ stage 108 or the moving means of the work holder 2 is used. Then, the detection optical system 100 and the illumination optical system 103 or the engine block 1 to be inspected are moved to a position where the adjacent cylinder part can be inspected. By repeating this, the inner walls in the four cylinders can be inspected.

As shown in FIG. 22, the main control unit 125 (9) expands the map format and the defective part on the screen of the display device 41 as the inspection result of one engine block based on the defect information 135 detected as described above. It is possible to output and display an image (including a defective portion 172). As for this inspection result, in the example of four cylinders, four bores are displayed in a map format. This defect map is displayed by discriminating one that has been generated alone (×: single nest), one that has been generated at a certain distance (●: dense nest), processing flaw (Δ :), and the like. As a result, it is possible to grasp the tendency of defect position bias and the frequency of defect types. Although one block is displayed in FIG. 22, the inspection results of a plurality of engine blocks may be displayed in a superimposed manner. In that case, it is possible to easily grasp the lot variation, the time, and the tendency of defect occurrence every day. The square on the right side of FIG. 22 is an area for displaying an enlarged image (original image or processed image (defect image)) of the defective part, and is displayed when, for example, the defective part is specified in the defect map on the left side. A part extracted as a defect may be surrounded by a circle 172 on the enlarged image of the detected defect portion, or the maximum diameter (for example, 0.24 mm), area (for example, 0.03 mm ² ), or the like may be displayed on the side. Is possible. Since the defect is not necessarily circular, the maximum diameter and area are displayed. By performing such display, it becomes easy to set inspection conditions (for example, detection angle θr, intensity of illumination light, etc.), and confirmation of a site where a defect occurs is easy to understand and intuitive determination can be made. .

  FIG. 23 shows the defect type of the cast hole, rough machining, and hole machining abnormality, the number of defects and the determination result of the bore quality in the perspective map format that is the inspection result by the main control unit 125 (9). It is a figure which shows what is displayed. Since this defect map shows the defect occurrence status at the position corresponding to the actual shape, it is possible to more clearly grasp the tendency of the comparison with the actual product, the defect position bias, and the defect type. FIG. 24 is a diagram showing an example in which the main control unit 125 (9) outputs detailed information of the inspection result, and shows a case where the No. 2 bore in FIG. 23 is displayed. The size of the cast hole is indicated by a diameter (for example, Φ1.6 mm), the processing roughness is indicated by an area (for example, 25 mm × 80 mm), and an abnormality in a processing hole is indicated by width J × height ΔK (for example, as shown in FIG. 17F). 1.5 mm × 3 mm). Each coordinate (H, ω) is also represented by a height H (mm) obtained from the coordinates of the Z stage and the linear image sensor and a rotation angle ω (°) obtained from the rotation angle of the θ stage 106.

Further, the main control unit 125 (9), for example, shows a tabulation table of inspection results obtained by tabulating one day or one week based on the detected defect information 135 and the evaluation result (pass / fail judgment result), for example, as shown in FIG. Thus, it can be output on the screen of the display device 141. That is, the inspection date, inspection time, block No. Boa No. , Number of defects, type of defects such as nests and processing flaws, defect size (maximum diameter) (mm), defect area (mm ² ) and defect coordinates (X, Y), and evaluation results (for each bore) Pass / fail and total pass / fail for each block) are displayed on the screen of the display device 141. Thereby, the details of the inspection result can be known.

  FIGS. 26 to 28 are diagrams for explaining the engine cylinder manufacturing process and the flow of feeding back inspection results from the overall control unit 9 to the processing machine in the previous process. The engine block 12 shown in FIG. 2 is made of cast iron or aluminum casting. There are various manufacturing processes. FIG. 26 shows a case where the inner surface of the bore manufactured in the casting process (boring process 200 and honing process 201) is not cleaned, and FIG. 27 shows that the inner surface of the bore manufactured in the casting process is cleaned using a cleaning device. 28 is manufactured in the aluminum casting process, and after the boring process, the surface is plated in the plating process 204 to increase the hardness, and then the honing process 201 is performed to perform the honing process. 203 shows the case of inspecting the inner surface of the bore after cleaning.

  26, boring is performed on the bore inner surface using a boring machine (200), and then a honing eye is machined using a honing machine (201), and the bore inspection is performed using the defect inspection apparatus shown in FIG. The case where the inner surface is inspected (202) is shown. In the defect inspection process 202, when processing roughness has occurred, gradually increased, or the area has increased as a result of inspection by the defect inspection apparatus, information on processing abnormality is output to the honing machine, such as a network. It outputs using the apparatus 11 (feeds back). Possible causes of processing failure include misalignment of boring machine, non-uniform reciprocation stroke of honing machine, non-uniform expansion of grinding wheel, non-uniform clogging of grinding wheel, non-uniform cutting amount of grinding wheel, etc. It is done. Therefore, the defect inspection apparatus according to the present invention optimizes these processing conditions (such as boring machine misalignment, honing machine reciprocating stroke, grinding wheel expansion, grinding wheel clogging, grinding stone cutting amount, etc.). It is possible to provide information for performing this by feeding back defect inspection information to a honing machine or the like.

  FIG. 27 is a diagram in which the cleaning process 203 is added to the manufacturing process shown in FIG. 26 and the defect inspection apparatus in the defect inspection process 202 feeds back to the cleaning apparatus in the cleaning process 203. In the cleaning apparatus of the cleaning process 203, machined chips and the like are removed by ultrasonic vibration or the like by immersing the engine block in a cleaning solution. If this cleaning is applied to a large number of engine blocks for a long time, the cleaning liquid becomes dirty, and the dirt reattaches to the engine block. In this state where dirt is reattached, even if cleaning is performed by the cleaning unit 3 of the defect inspection apparatus shown in FIG. 1, water droplets and large foreign matter on the cylinder surface can be removed, but minute dust adhering to the cylinder surface can be removed. Chips cannot be removed and are detected as defects together with the defects in the defect inspection apparatus. Further, in the above-described cleaning apparatus, chips are attached when the cleaning liquid flows down, so that a defect is often generated on the lower surface side on which the engine block is placed in the defect inspection apparatus. When these characteristic inspection results are generated in the defect inspection apparatus, it can be determined that the cleaning performance of the cleaning apparatus in the cleaning step 203 has decreased, so that dirt has adhered to the cylinder inner surface, and the number of defects in the casting hole has increased significantly. . Therefore, the defect inspection apparatus according to the present invention feeds back information such as the number of defects obtained as a result of the defect inspection to the cleaning apparatus in the cleaning step 203, so that the cleaning apparatus can estimate the possibility of abnormality.

  In FIG. 28, the bore inner surface is bored using a boring machine in the boring process 200, and then the surface is plated using a plating apparatus in the plating process 204, and the honing is processed in the honing process 201. Then, the case where the inner surface of the bore is inspected using the defect inspection apparatus shown in FIG. The defect inspection apparatus according to the present invention shows a case in which information related to processing roughness is also provided to a plating processing apparatus that performs plating processing on an aluminum casting or the like in the plating processing step 204. If the plating thickness is insufficient, the plating is not applied partially, or the adhesion of the plating is reduced, the inner surface of the cylinder inspected after processing is abnormally rough and occurs in a large area. Or the evaluation value of brightness variation increases. Process management can be performed by feeding back the information on the processing roughness to the plating process conditions of the plating apparatus.

  As described above, the defect inspection apparatus in the defect inspection process 202 according to the present invention feeds back a defect inspection result to a honing machine, a cleaning apparatus, a plating processing apparatus, etc., thereby processing or processing conditions in the engine cylinder manufacturing process. Can be maintained appropriately, and the product yield can be improved.

  In addition, the defect inspection apparatus according to the present invention transfers the obtained defect information and the evaluation result (pass / fail judgment result) to the upper management system online or offline so that the upper management system It can be used for logistics management and feedback of processes in the previous process.

  As described above, according to the present embodiment, it is possible to stably and automatically detect defects such as irregularities and scratches generated on a curved surface such as an inner wall of a cylindrical object.

  Further, according to the present embodiment, the coordinates (position) and size of the detected defect in the appearance inspection for automatically and stably detecting defects such as irregularities and scratches generated on a curved surface such as an inner wall of a cylindrical object or the like. In addition, since the information such as the area is digitized, it can be used for physical distribution management in the subsequent process and feedback of the process in the previous process, and it becomes possible to manufacture a product with good quality.

It is a figure which shows schematic structure of one Example of the defect inspection apparatus which test | inspects the defect which generate | occur | produced in the inner surface of the to-be-inspected cylindrical object which concerns on this invention. It is external appearance explanatory drawing of a to-be-inspected cylindrical object. It is a block diagram which shows the Example of the washing | cleaning part which wash | cleans the inner surface of the to-be-inspected cylindrical object with which the defect inspection apparatus shown in FIG. It is a concrete block diagram which shows one Example of the defect inspection apparatus except the cleaning part which concerns on this invention. To describe the relationship between a cylindrical object to be inspected in a defect inspection apparatus according to the present invention and an illumination optical system and a detection optical system mounted on a support member supported by a θ stage (rotation stage) via an XY fine movement mechanism. FIG. It is a figure for demonstrating the relationship of the image detection position by the mirror which is a detection optical system in the inner surface of a to-be-inspected cylindrical object, and a linear image sensor. It is a figure for demonstrating the manifestation principle of a defect part by making a concave defect into an example. It is a figure which shows the relationship between ratio of detection angle (theta) r and distance L / diameter R, the contrast of a defective part, and background dispersion | variation (normalization value). It is a figure for demonstrating the two-dimensional image detected from the inner surface of a to-be-inspected cylindrical object with a linear image sensor, and the fluctuation | variation of the brightness detected with the linear image sensor by center shift. It is a figure for demonstrating the directional characteristic of the illumination by an illumination optical system. It is a figure which shows the flow of the image process including the defect extraction process in the image process part shown in FIG. It is a figure explaining one Example of the data designation | designated for the mask image generation for removing the cross-hatch mesh noise as background noise. It is the figure which showed the detected original image and the defect image displayed as a test result. It is explanatory drawing of the method of extracting defects, such as a deep cast hole, based on the feature-value of the image signal extracted from a defect candidate. It is a figure for demonstrating the two-dimensional image detected from the inner surface of the to-be-inspected cylindrical object by which the hole was machined in the cylinder side surface with the linear image sensor. It is a figure which shows the flow of the image process including the extraction process of the outline vicinity defect of the processed hole in the image shown in FIG. It is a figure which shows the image processing procedure of extraction of the outline vicinity defect of the processed hole in the image shown in FIG. It is a figure which shows the flow of an image process including the extraction process of a process flaw defect, such as process roughening and plating peeling. It is a figure which shows the example of extraction of the process flaw defect, such as process roughening and plating peeling. It is a figure which shows the flow of an image process in the case of identifying and inspecting a casting hole defect, a defect near the outline of a processing hole, and a processing flaw defect. It is a functional block diagram which shows one Example of a defect inspection apparatus in the case of identifying and inspecting a casting hole defect, a defect near the outline of a processing hole, and a processing flaw defect, respectively. It is a figure explaining one Example (defect map of four bores) of the display of the defect inspection result in 1 engine block. It is a figure explaining other one Example (defect map of four bores) of the display of the defect inspection result in 1 engine block. FIG. 24 is a diagram for explaining an example (defect map of one bore) of detailed display of defect inspection results in one engine block shown in FIG. 23. It is a figure which shows one Example of the total display of a defect inspection result. It is a figure which shows the 1st Example of the manufacturing method (processing method) which feeds back the test result of the defect inspection apparatus which concerns on this invention to a processing process. It is a figure which shows the 2nd Example of the manufacturing method (processing method) which feeds back the test result of the external appearance inspection apparatus which concerns on this invention to a process process and a washing process. It is a figure which shows the 3rd Example of the manufacturing method (processing method) which feeds back the test result of the external appearance inspection apparatus which concerns on this invention to a process process, a washing process, and a plating process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Cylinder to be inspected (cylindrical object to be inspected), 2 ... Work holder, 3 ... Cleaning part, 4 ... Optical system, 5 ... Stage, 6 ... Base, 7 ... Mechanism control part, 8 ... Image processing part, 9 ... Whole Control unit, 10 ... display device, 11 ... output device, 21, 26, 30 ... shaft, 22, 27 ... fluid flow path, 23, 28 ... nozzle, 24, 29 ... outlet, 31 ... holder, 32 ... Wiping member or wiping member, 102 ... mirror (reflection optical element), 103 ... illumination optical system, 104 ... imaging lens (imaging optical element), 105 ... linear image sensor, 106 ... XY fine movement stage, 107 ... θ stage ( Rotating stage), 108 ... XYZ stage, 100 ... detection optical system, 112 ... support member, 121 ... A / D converter, 122 ... pre-processing unit, 123 ... memory, 124 ... defect extraction processing unit, 125 ... main control unit , 26 ... Illumination power supply, 142 ... Input means, 143 ... Storage device, 151 ... Linear detection area (detection point) by the line sensor, 200 ... Boring process, 201 ... Honing process, 202 ... Defect inspection process, 203 ... Washing process, 204 ... plating process.

Claims (23)

  A cleaning unit for cleaning the residual liquid remaining on the inner surface of the cylindrical object to be inspected, an optical system for detecting a two-dimensional image of the inner surface of the cylindrical object to be inspected cleaned by the cleaning unit, and 2 detected by the optical system A defect inspection apparatus comprising: an image processing unit that acquires defect information based on a three-dimensional image; and an output unit that outputs the defect information acquired by the image processing unit.   The defect inspection apparatus according to claim 1, wherein the cleaning unit is configured to inject and clean the residual liquid remaining on the inner surface of the cylindrical object to be inspected.   The defect inspection apparatus according to claim 1, wherein the cleaning unit is configured to perform cleaning by wiping or sucking off residual liquid remaining on an inner surface of the cylindrical object to be inspected.   An optical system for detecting a two-dimensional image of the inner surface of a cylindrical object to be inspected with a hole processed on the inner surface, and the vicinity of the contour of the processed hole extracted or set for the two-dimensional image detected by the optical system An image processing unit that acquires information on defects generated in the vicinity of the contour of the processing hole based on the image of the image, and an output unit that outputs information on the defects acquired by the image processing unit. Defect inspection equipment.   An optical system that detects a two-dimensional image of the inner surface of a cylindrical object to be inspected, and a two-dimensional image detected by the optical system is divided into a plurality of regions, and brightness variations among the divided regions are based on A defect inspection apparatus comprising: an image processing unit that acquires information on a processing flaw defect; and an output unit that outputs information on the processing flaw defect acquired by the image processing unit.   An optical system for detecting a two-dimensional image of the inner surface of a cylindrical object to be inspected with a hole processed on the inner surface, and a hole defect, a defect in the vicinity of a processed hole, and processing based on the two-dimensional image detected by the optical system A defect inspection apparatus comprising: an image processing unit that classifies a defect as a defect and acquires information about each defect; and an output unit that outputs information about each defect acquired by the image processing unit. The optical system is
An illumination optical system that illuminates a linear region facing the axial direction on the inner surface of the cylindrical object to be inspected, and a reflected light image from the linear region is desired with respect to the normal direction of the linear region The reflection optical element that reflects the reflected light image from the linear region so as to be detected from the direction inclined at the detection angle of the light and directs it in the axial direction, and is reflected by the reflective optical element and directed in the axial direction. An imaging optical element that linearly forms a reflected light image from a linear region, and a linear image sensor that receives the linear reflected light image formed by the imaging optical element and outputs an image signal And a support member attached to the detection optical system, and the support member is rotated about the axial center of the inner surface of the cylindrical object to be inspected, so that 2 on the inner surface of the cylindrical object to be inspected from the linear image sensor. Dimensional image signal is detected Defect inspection apparatus according to claim 1 or 4 or 5 or 6, characterized in that the configuration.   The optical system further includes an XYZ stage, a rotary stage provided on the XYZ stage for rotating the support member, and provided between the rotary stage and the support member. 8. The defect inspection apparatus according to claim 7, further comprising an XY fine movement stage for finely moving the rotary stage in the XY direction.   In the image processing unit, an A / D converter that converts a two-dimensional image signal detected from the optical system into a two-dimensional digital image signal, and a two-dimensional digital image signal obtained from the A / D converter 7. The defect inspection apparatus according to claim 1, further comprising a preprocessing unit that performs shading correction and further removes background noise.   The defect inspection apparatus according to claim 9, wherein the pre-processing unit is configured to remove cross-hatch mesh noise as the background noise by mask processing.   The image processing unit further extracts a defect candidate based on the two-dimensional digital image signal obtained from the preprocessing unit, calculates a feature amount of the extracted defect candidate, and based on the calculated feature amount. The defect inspection apparatus according to claim 9, further comprising a defect extraction processing unit that determines and obtains defect information.   The defect inspection according to claim 1, 4, 5, or 6, wherein the output unit includes a display device that displays a defect map of a cylindrical object to be inspected as defect information acquired from the image processing unit. apparatus.   In the output unit, one or more display screens of a projection view, a perspective view, a development view, a three-dimensional view, and a plan view of a cylindrical object to be inspected as defect information acquired from the image processing unit; 7. The defect inspection apparatus according to claim 1, further comprising a display device that displays any one or more of the size, type, shape, and coordinates of the defect. A cleaning process for cleaning residual liquid remaining on the inner surface of the cylindrical object to be inspected;
A detection process in which an optical system detects a two-dimensional image of the inner surface of the cylindrical object cleaned in the cleaning process, and an image processing process in which defect information is acquired based on the two-dimensional image detected in the detection process; And a defect inspection step having an output step of outputting information on the defect acquired in the image processing step.   A detection process in which an optical system detects a two-dimensional image of the inner surface of a cylindrical object to be inspected whose inner surface has been processed, and the processed hole extracted or set for the two-dimensional image detected in the detection process Defect inspection process comprising: an image processing process for acquiring information on defects generated in the vicinity of the contour of the processed hole based on an image in the vicinity of the contour of the machine; and an output process for outputting information on the defects acquired in the image processing process A defect inspection method characterized by comprising:   A detection process in which an optical system detects a two-dimensional image of the inner surface of the cylindrical object to be inspected, and the two-dimensional image detected in the detection process is divided into a plurality of areas, and the brightness of each divided area is determined. A defect having a defect inspection step having an image processing process for acquiring information on a processing flaw defect based on variation and an output process for outputting information on the processing flaw defect acquired in the image processing process Inspection method.   Detection process in which an optical system detects a two-dimensional image of the inner surface of a cylindrical object to be inspected with a hole processed on the inner surface, and the vicinity of a casting hole defect and a processed hole based on the two-dimensional image detected in the detection process A defect inspection step comprising: an image processing process for acquiring information on each defect by classifying into defects and processed scratch defects; and an output process for outputting information on each defect acquired in the image processing process. Defect inspection method. The optical system in the detection process of the defect detection process,
An illumination optical system that illuminates a linear region facing the axial direction on the inner surface of the cylindrical object to be inspected, and a reflected light image from the linear region is desired with respect to the normal direction of the linear region The reflection optical element that reflects the reflected light image from the linear region so as to be detected from the direction inclined at the detection angle of the light and directs it in the axial direction, and is reflected by the reflective optical element and directed in the axial direction. An imaging optical element that linearly forms a reflected light image from a linear region, and a linear image sensor that receives the linear reflected light image formed by the imaging optical element and outputs an image signal And a support member attached to the detection optical system, and the support member is rotated about the axial center of the inner surface of the cylindrical object to be inspected, so that 2 on the inner surface of the cylindrical object to be inspected from the linear image sensor. Dimensional image signal is detected Defect inspection method according to any one of claims 14 to 17, characterized by being configured to.   In the image processing process of the defect inspection process, an A / D conversion process for converting a two-dimensional image signal detected from the detection process into a two-dimensional digital image signal, and a two-dimensional digital signal converted by the A / D conversion process The defect inspection method according to any one of claims 14 to 17, further comprising a pre-processing step of performing shading correction on the image signal and further removing background noise.   20. The defect inspection method according to claim 19, wherein the preprocessing step includes a removal step of removing the cross-hatch mesh noise as the background noise by mask processing. A boring process for forming a bore of a cylindrical object by boring with a boring machine;
A honing process for forming a honing eye on the inner surface of the bore formed in the boring process by honing with a honing machine;
A cleaning step of cleaning residual liquid remaining on the bore inner surface of the cylindrical object in which the honing eyes are formed in the honing processing step;
A detection process for detecting a two-dimensional image of the bore inner surface of the cylindrical object cleaned in the cleaning process with an optical system; an image processing process for acquiring defect information based on the two-dimensional image detected in the detection process; A defect inspection step having an output process for outputting information of the defect acquired in the image processing process,
A method of processing an inner surface of a cylindrical object, comprising feeding back information on defects output in an output process of the defect inspection process to the boring process or honing process. A boring process for forming a bore of a cylindrical object by boring with a boring machine;
A honing process for forming a honing eye on a bore inner surface of the cylindrical object formed in the boring process by honing by a honing machine;
A first cleaning step of cleaning the bore inner surface of the cylindrical object in which the honing eyes are formed in the honing processing step;
A second cleaning step of cleaning residual liquid remaining on the inner surface of the bore of the cylindrical object cleaned in the first cleaning step;
A detection process for detecting a two-dimensional image of the bore inner surface of the cylindrical object cleaned in the second cleaning process by an optical system, and an image processing process for acquiring defect information based on the two-dimensional image detected in the detection process And a defect inspection step having an output process for outputting information on the defect acquired in the image processing process,
A method of processing an inner surface of a cylindrical object, comprising feeding back information on defects output in the output process of the defect inspection process to the boring process, honing process, or first cleaning process. A boring process for forming a bore of a cylindrical object by boring with a boring machine;
A plating process for plating the inner surface of the bore of the cylindrical object formed in the boring process;
A honing process for forming honing eyes on the inner surface of the bore subjected to the plating process in the plating process by honing by a honing machine;
A first cleaning step of cleaning the bore inner surface of the cylindrical object in which the honing eyes are formed in the honing processing step;
A second cleaning step of cleaning residual liquid remaining on the inner surface of the bore of the cylindrical object cleaned in the first cleaning step;
A detection process for detecting a two-dimensional image of the bore inner surface of the cylindrical object cleaned in the second cleaning process by an optical system, and an image processing process for acquiring defect information based on the two-dimensional image detected in the detection process And a defect inspection step having an output process for outputting information on the defect acquired in the image processing process,
A method of processing an inner surface of a cylindrical object, comprising feeding back information on defects output in the output process of the defect inspection process to the boring process, honing process, first cleaning process, or plating process.

Images