JP2024121242A - Method and device for inspecting machining tools - Google Patents

Abstract

[Problem] To provide an effective inspection method and inspection device for machining tools. [Solution] The machining tool inspection method includes an image acquisition step 10 for photographing the tip of a machining tool that has been subjected to a specified machining process and acquiring an image of the processed part, an image extraction step 12 for extracting worn parts from the acquired image of the processed part by filtering using a Fourier transform and extracting the image of the worn part as a processed part wear image, a binarization step 14 for binarizing each pixel of the extracted image of the worn part based on a set predetermined threshold value to form a binary image in which the worn parts are composed of white pixels, a noise removal step 16 for removing noise from the binary image to extract a noise-processed image that more clearly displays the worn parts, and an evaluation step 16 for evaluating the wear state of the tip of the machining tool from the white pixels that compose the worn parts in the noise-processed image. [Selected Figure] Figure 1

Description

The present invention relates to a method and device for inspecting machining tools, and in particular to a method and device that can effectively inspect the wear condition of machining tools used in a specified machining process.

Conventionally, drilling and cutting of workpieces have been performed using processing machines equipped with processing tools such as drills and end mills. However, as such processing progresses, the tips of the processing tools such as drills and end mills gradually wear out and chip, causing problems such as a decrease in processing accuracy. For this reason, it is important to inspect and manage the cutting surface of the tip of the processing tool used for the intended processing, but it is extremely difficult to judge this by visual inspection.

For example, alumina parts used as materials for semiconductor manufacturing equipment parts require many processes of drilling holes with diamond-coated small-diameter drills. However, since small-diameter holes are drilled in alumina parts, which are difficult to process, there is a problem that the drill tip (cutting edge) wears as the processing progresses, and the drill breaks frequently during processing, which causes product defects and reduced workability (productivity). For this reason, the drill tip processing part is inspected. In this inspection, the drill is removed from the processing machine every time a predetermined amount of workpiece is processed, and a skilled technician uses a magnifying glass or the like to visually check the wear condition of the drill tip and decide whether to replace it. However, such drill inspection is a personal and time-consuming and complicated task, which creates inherent problems in stabilizing the inspection accuracy and efficiency.

Since there are many types of machining tools to be inspected, each with a different outer diameter and cutting edge shape, various methods have been proposed to use image processing technology to replace visual inspection with image processing. For example, Japanese Patent Application Laid-Open No. 2001-129711 discloses a drill inspection method in which the drill is imaged not only in the direction of the drill's rotation axis but also in a direction perpendicular to the drill cutting edge, so that the entire cutting edge is captured within the depth of field of the lens, the captured image is then binarized and converted into shape data, and the shape data is compared with a preset value to determine whether the drill is good or bad. Furthermore, Japanese Patent No. 4128613 proposes a drill inspection device that includes an imaging unit that images the cutting edge of a drill in a direction along the rotation axis, an image processing unit that binarizes the pixel values of the image input from the imaging unit and aligns the resulting binary image, a measurement unit that measures shape information of the cutting edge area in the image processed by the image processing unit, and a judgment unit that judges the condition of the cutting edge of the drill from the shape information measured by the measurement unit.

However, in these conventional drill inspection methods and inspection devices, shape data is obtained from image information obtained by photographing, through simple binarization processing. Therefore, if the surface of the drill bit, which is the processing tool to be photographed, is rough, for example when it has an uneven shape due to a coating, there is a large variation in the amount of reflected light from the illumination light during photography, and there is an inherent problem of misdetecting parts of the drill other than the worn parts. In addition, as the processing with the drill progresses, dirt begins to adhere to the surface of the drill bit, and there is an inherent risk that such adhered dirt will be detected at the same time and mistakenly recognized as a worn part.

JP 2001-129711 A Patent No. 4128613

The present invention was made against the backdrop of these circumstances, and the problem to be solved is to provide an effective inspection method and inspection device for machining tools, and another problem is to provide a technology that can more effectively grasp the wear condition of the tip machining surface of a machining tool that has been used for a specified machining process.

The present invention can be suitably implemented in various aspects as listed below in order to solve the above problems, and each aspect described below can be adopted in any combination. It should be understood that the aspects and technical features of the present invention are not limited to those described below, but can be recognized based on the description of the entire specification and the inventive concept disclosed in the drawings.

First, the first aspect of the present invention for solving the above-mentioned problems is an inspection method for a machining tool, which includes an image acquisition step of photographing the tip of a machining tool subjected to a predetermined machining and acquiring an image of the machining part to be inspected; an image extraction step of extracting the wear part caused by the predetermined machining of the tip of the machining part from the acquired image of the machining part by filtering using a Fourier transform, and extracting it as a machining part wear image; a binarization step of binarizing each pixel from the extracted machining part wear image based on a set predetermined threshold value to form a binary image in which the wear part is composed of white pixels; a noise removal step of removing minute white pixel parts that do not correspond to the wear part with a high density of white pixels as noise from the formed binary image, thereby extracting a noise-processed image that displays the wear part more clearly; and an evaluation step of evaluating the wear state of the tip of the machining tool from the white pixels that constitute the wear part in the noise-processed image.

The second aspect of the present invention is characterized in that the image acquisition step includes an imaging step of acquiring a plurality of tip images by photographing the tip machining portion of the machining tool in such a way that the focal points are located at a predetermined interval from the tool tip toward the tool contour portion in the tool axial direction; and an image synthesis step of using the plurality of tip images to form a single synthetic image in focus over the entire tip machining portion by a depth of field synthesis method as the machining portion image.

Furthermore, a third aspect of the present invention is characterized in that in the image synthesis process, each tip image obtained in the imaging process is divided vertically and horizontally into multiple parts to form a large number of divided images, and then the divided images of the same location in each tip image are compared to select the clearest divided image, thereby forming the single synthesized image.

In addition, the fourth aspect of the present invention is characterized in that in the image extraction step, the processed part image is subjected to a fast Fourier transform to form a spectral image, and then the spectral image is subjected to a masking process to leave high-frequency parts, and then an inverse fast Fourier transform is performed to extract the processed part wear image.

The fifth aspect of the present invention is characterized in that in the noise removal process, at least an expansion/reduction process and a filling process of a small area of white pixels are performed on each pixel constituting the binarized image.

The sixth aspect of the present invention is characterized in that it further includes a tool center detection step in which the vertical and horizontal sections in the processed part image in which the processing tool is shown are determined by plotting the sums of the amplitudes obtained by fast Fourier transform of each column and each row of pixels located in the vertical and horizontal directions of the processed part image, respectively, and then determining the point where the pixel columns and pixel rows located in the center of the vertical and horizontal sections intersect as the tool center of the processing tool.

Furthermore, a seventh aspect of the present invention is characterized in that it further includes a phase correction step in which the noise-processed image is rotated around the tool center of the machining tool determined in the tool center detection step by a predetermined angle, the image at each rotation angle is divided into two equal parts, and the two divided images are compared to correct the phase of the noise-processed image so that the wear portion extends in the horizontal direction.

In addition, the eighth aspect of the present invention is characterized in that the evaluation step is carried out after removing white pixel areas that are located on a circumference at a certain distance from the center of rotation and at a specified angle from the wear part in the horizontal position, using the noise-processed image that has been phase-corrected by the phase correction step.

A ninth aspect of the present invention is characterized in that the machining tool is a drill, and a tenth aspect is characterized in that the machining tool is a diamond-coated drill.

The eleventh aspect of the present invention is an inspection device for a machining tool, which includes: an image acquisition means for acquiring an image of the machining part to be inspected by photographing the tip machining part of a machining tool that has been subjected to a predetermined machining; an image extraction means for extracting the wear part caused by the predetermined machining of the tip machining part from the acquired image of the machining part by filtering using a Fourier transform, and extracting it as a machining part wear image; a binarization means for binarizing each pixel from the extracted machining part wear image based on a set predetermined threshold value, and forming a binary image in which the wear part is composed of white pixels; a noise removal means for removing minute white pixel parts that do not correspond to the wear part with a high density of white pixels as noise from the formed binary image, thereby extracting a noise-processed image that displays the wear part more clearly; and an evaluation means for evaluating the wear state of the tip machining part of the machining tool from the white pixels that constitute the wear part in the noise-processed image.

In a twelfth aspect of the present invention, the image collection means includes a ring-shaped illumination means located in a circular shape around the tool axis below the machining tool and irradiating light onto the machining tip of the machining tool; a right-angle prism that refracts the light reflected downward from the machining tip of the machining tool in a horizontal direction by the light emitted from the ring-shaped illumination means; and a camera that receives the light refracted by the right-angle prism and captures an image of the machining tip.

Furthermore, a thirteenth aspect of the present invention is characterized in that the image collection means is disposed within a processing machine that performs a predetermined processing operation using the processing tool.

In addition, a fourteenth aspect of the present invention is characterized in that the image collection means includes an imaging device that captures images of the tip machining portion of the machining tool so that the focal points are located at predetermined intervals from the tool tip toward the tool contour in the tool axial direction, thereby acquiring a plurality of tip images; and an image synthesis means that uses the plurality of tip images to form, as the machining portion image, a single synthetic image that is in focus over the entire tip machining portion by a depth of field synthesis method.

In a fifteenth aspect of the present invention, the imaging device has a ring-shaped illumination means that is located below the machining tool in a circular shape centered on the tool axis and that irradiates light onto the machining tip of the machining tool; a right-angle prism that refracts in a horizontal direction the light that is reflected downward from the machining tip of the machining tool by the irradiation of light from the ring-shaped illumination means; and a camera that receives the light refracted by the right-angle prism and captures an image of the machining tip.

Furthermore, a sixteenth aspect of the present invention is characterized in that the imaging device is disposed in a processing machine that performs a predetermined processing with the processing tool, and is configured so that the focal position of the camera is moved at a predetermined interval from the tip of the tool toward the tool contour by the axial movement of the processing tool by the processing machine.

According to the present invention, it is possible to more effectively detect the wear state of the tip of a machining tool used for a specific machining operation, particularly a drill with a large surface unevenness such as a diamond-coated drill, and to accurately identify the worn part and advantageously evaluate it.

1A to 1C are process explanatory diagrams illustrating an overview of a machining tool inspection method according to the present invention. 1 is a schematic explanatory diagram showing an example of an inspection device for a machining tool according to the present invention; 3 is an explanatory diagram showing a schematic diagram of a plurality of tip images obtained by photographing while moving the focus in the drill axial direction in the device shown in FIG. 2 . FIG. 1 shows an example of a tip image obtained by photographing the tip of the drill with the focus set on the tip. 1 shows an example of a tip image obtained by photographing the drill tip while focusing on its contour. FIG. 13 is an explanatory diagram showing a form of division of one tip image shown typically. FIG. 13 is an explanatory diagram showing a form in which a plurality of tip portion images are divided and the divided images are compared. 1A and 1B are diagrams illustrating an example of a composite image obtained by depth of field stacking, including an enlarged image of a portion of the composite image. FIG. 13 is an explanatory diagram showing an example of extracting a processed portion wear image (IFFT image) from a composite image by filtering using a Fourier transform. FIG. 11 is an explanatory diagram showing an example of the result of a fast Fourier transform (FFT) used to measure the center of a drill, where (a) and (b) are explanatory diagrams showing an example of the result of a fast Fourier transform (FFT) for one row and one column, respectively, and the sum of the amplitudes of each row and the sum of the amplitudes of each column. 1 shows an example of a binary image, where (a) is a binary image obtained from an image obtained by inverse fast Fourier transform (IFFT image) without setting a threshold, and (b) is a binary image obtained from the IFFT image by setting a threshold so as to suppress detection of areas other than worn areas. The following shows an example of an image obtained by noise processing, including a partially enlarged image, where (a) shows an image obtained by performing expansion and reduction processing on a binary image, and (b) shows an image obtained by performing a small area filling process. FIG. 13 is a diagram for explaining a method of phase correction. 13A and 13B are images showing the state before and after a second noise processing step that removes noise present at a position away from the worn portion, where (a) is the image before the processing and (b) is the image after the processing. 1 is a schematic explanatory diagram showing a preferred embodiment of an inspection device for a machining tool according to the present invention;

In order to clarify the configuration of the present invention more specifically, a representative embodiment of the present invention will be described in detail below with reference to the drawings.

First, FIG. 1 shows a schematic diagram of the steps of the inspection method for machining tools according to the present invention. In the image acquisition process, 10 is an image acquisition process in which the tip of a machining tool used for a specified machining process is photographed to acquire an image of the machining part to be inspected. The machining tools that are targeted include drills and end mills for drilling and cutting, and among these, the features of the present invention can be advantageously exhibited in drills, particularly drills that are diamond coated and have a large variation in the amount of reflected light due to surface irregularities.

In the case where the depth of field of the processed part image captured in the image capture process 10 is insufficient, it is desirable to configure the image capture process to include an imaging process for acquiring multiple tip images by photographing the tip of the processing tool so that the focal points are located at a predetermined interval from the tool tip toward the tool contour in the tool axial direction, and an image synthesis process for forming a single synthetic image in focus over the entire tip of the processing part as the processed part image by a depth of field synthesis method using the multiple obtained tip images.

In the image extraction step 12, the processed part image collected above is filtered using a Fourier transform to extract the wear part caused by the specified processing of the tip processed part, and the processed part wear image is extracted. In particular, in this image extraction step 12, it is advantageous to perform a fast Fourier transform on the processed part image to form a spectral image, and then perform a mask process on the spectral image so that high-frequency parts remain, and then perform an inverse fast Fourier transform to extract the processed part wear image.

In addition, in the binarization step 14 following the image extraction step 12, each pixel is binarized based on a predetermined threshold value set from the processed part wear image extracted as described above, and a binary image in which the wear part is composed of white pixels is formed.

Next, in the noise removal process 16, minute white pixel areas that do not correspond to the worn areas with a high density of white pixels are removed as noise from the binary image thus formed, and a noise-processed image that more clearly displays the worn areas is extracted. Advantageously, in this noise removal process 16, at least expansion/contraction processing and a filling-in processing of minute areas of white pixels are performed on each pixel that constitutes the binary image.

Finally, in evaluation step 18, the noise-processed image obtained in the noise removal step 16 is used to evaluate the wear state of the tip processing part of the tool from the white pixels that make up the worn part in the image. In this evaluation, a method is generally advantageously adopted in which the wear area of the tip processing part of the tool is calculated from the white pixels that make up the worn part, and the wear state of the tip processing part is evaluated.

Figure 2 shows a schematic diagram of an example of an inspection device capable of evaluating the wear state of a drill as a machining tool. In this example, 20 is a drill, and inspection light is irradiated onto its tip processing portion (blade portion) from ring illumination 22, a ring-shaped illumination means located below the drill. The ring illumination 22 is configured by positioning multiple light sources in a ring shape around the drill axis below the drill 20, and irradiates the tip processing portion of the drill 20 with inspection light. In addition, a right-angle prism 24 is positioned so that the light reflected downward from the wear area (wear surface) of the drill tip processing part is refracted horizontally by the light emitted from the ring illumination 22. The reflected light refracted horizontally by the right-angle prism 24 is received by a camera 28 such as a CCD camera or CMOS camera via a lens 26 and captured as an image of the processing part of the drill 20, and the image information is output to an image processing device 30.

In addition, if the depth of field of the processed part image captured by camera 28 in this manner is sufficiently satisfactory, an image extraction operation is carried out in which the image information is used as is to extract the processed part wear image. However, if the depth of field of the image captured by camera 28 is insufficient, the imaging device (unit) including ring illumination 22, right-angle prism 24 and camera 28 is made to be displaced relative to the drill 20 in the drill axial direction, and the tip processed part of the drill 20 is photographed so that the focal points are positioned at predetermined intervals, for example 20 μm intervals, from the drill tip to the drill contour in the drill axial direction, thereby enabling multiple tip images to be obtained. In addition, an image synthesis means is provided that uses these multiple tip images to form a single synthesized image that is in focus over the entire tip processing part as a processing part image using a depth of field synthesis method. In the inspection device shown in Figure 2, such an image synthesis means is shown as image synthesis unit 32 in image processing device 30.

In particular, when the tip machining portion (machining surface) of the machining tool to be inspected is not a horizontal flat surface but is configured as an inclined curved surface like a drill (20), or when there is a retraction of the cutting edge due to wear or a positioning error of the drill (20), a configuration including the above-mentioned imaging process (device) and image synthesis process (means) is advantageously adopted as the image acquisition process (means).

The image processing device 30 in the inspection device shown in FIG. 2 has an image extraction unit (means) 34 that uses the processed part image made up of the composite image formed in the image synthesis unit 32 to extract the wear parts caused by the specified processing of the nozzle tip processing part from the processed part image by filter processing using Fourier transform, and extracts it as a processed part wear image. It also includes a binarization unit (means) 36 that binarizes each pixel from the extracted processed part wear image based on a set predetermined threshold value to form a binary image in which the wear part is composed of white pixels, a noise removal unit (means) 38 that removes minute white pixel parts that do not correspond to the wear parts with high white pixel density as noise from the formed binary image, and extracts a noise-processed image that displays the wear part more clearly, and an evaluation unit (means) 40 that evaluates the wear state of the tip processing part of the drill 20 from the white pixels that constitute the wear part in the noise-processed image.

Specifically, in the inspection device shown in FIG. 2, the camera 28 constituting the imaging device captures images with the focus positioned at a predetermined interval in the axial direction of the drill from the tip to the contour of the drill 20, thereby acquiring multiple (six in this case) images of the tip portion with the focus position changing sequentially. Note that FIG. 3 shows a schematic representation of these six tip portion images. Also, FIGS. 4 and 5 show actual captured images, where the image in FIG. 4 was captured by focusing on the tip of the drill, and the image in FIG. 5 was captured by focusing on the contour of the drill.

Next, when the image synthesis unit 32 of the image processing device 30 uses the multiple tip images obtained in this way to form a single synthesized image (processed part image) in focus over the entire tip processing part of the drill 20 using the depth of field synthesis method, the multiple tip images are generally divided vertically and horizontally into multiple parts with a dozen or so pixel corners each to form multiple divided images, and then the divided images of the same location in each tip image are compared to select the clearest divided image, thereby forming the desired detailed synthetic image. An example of this is shown in Figures 6 and 7.

That is, the multiple tip images, which are images made up of a large number of pixels, are each composed of 2048 pixels vertically and 2448 pixels horizontally, as shown in FIG. 6, and are divided vertically and horizontally to provide 19,584 divided images, each composed of 16 pixels vertically and 16 pixels horizontally. Then, the divided images in the same locations in each tip image are compared, as shown in FIG. 7, and the sharpest divided image is selected to compose a single image, which is extracted as a single composite image that is in focus across the entire field of view.

The comparison and extraction (selection) of the divided images of each tip image is performed by using a Laplacian filter to find the pixel value of each pixel in each divided image, and then determining and selecting the divided images with the largest variance (Laplacian value) in the filtered image as the image that is in focus. The selected divided images are then used to form a single overall image, resulting in the construction of a composite image such as that shown in Figure 8. In this composite image, the mallet-shaped white area located in the center and extending horizontally so as to gradually widen toward both ends (see the enlarged image at the bottom of Figure 8) indicates the reflected light from the worn part (surface), while the thin cloud areas of a certain width in the shape of arcs located above and below it (see the enlarged image at the top of Figure 8) indicate the reflected light caused by dirt.

In addition, in the image extraction unit 34 that executes the image extraction process of extracting the processed part wear image from the composite image (processed part image) collected as described above, a filter process using a Fourier transform is performed on the composite image, thereby extracting the wear part caused by the specified processing at the tip processing part of the drill 20. Specifically, as shown in FIG. 9, the composite image is subjected to a fast Fourier transform (FFT) to form a spectral image, and then the spectral image is subjected to a mask process that leaves the high frequency parts, and then an inverse fast Fourier transform (IFFT) is performed. By suppressing the detection of dirt on the drill surface while extracting the wear part (high frequency region), the desired processed part wear image (IFFT image) can be advantageously extracted.

In short, the drill 20, which is the machining tool to be photographed, has a coating surface with large irregularities, which results in large variations in the amount of reflected light, and simple binarization would result in erroneous detection of areas other than the worn parts. Therefore, in this invention, a fast Fourier transform (FFT) is used to detect only the worn parts, which are in the high frequency range, and the use of such FFT makes it possible to advantageously suppress the detection of dirt adhering to the drill surface.

Furthermore, in the present invention, the above-mentioned drill center detection process using the composite image is advantageously adopted when carrying out the phase correction process described below and the process of removing unnecessary white pixel portions from the phase-corrected noise-processed image. That is, the sum of the amplitudes obtained by fast Fourier transform (FFT) of each column and row of pixels located in the vertical and horizontal directions of the composite image, which is the processed part image, is plotted to determine the vertical and horizontal sections in the composite image in which the drill 20 is captured, and then the drill center detection process is carried out by determining the point where the pixel columns and pixel rows located in the center of these vertical and horizontal sections intersect as the drill center of the drill 20.

Specifically, by performing FFT on each pixel row (2048 rows) and each pixel column (2448 columns) constituting the composite image, the amplitude versus frequency for each row and column is obtained, an example of which is shown in the upper part of Figures 10(a) and (b), respectively. Then, by plotting the sum of the amplitudes of each pixel row and each pixel column obtained by such FFT, the sum of the amplitudes of each row and each column is obtained as a graph versus the number of rows and columns, as shown in the lower part of Figures 10(a) and (b), where the difference in the sum of the amplitudes is large between the background part and the drill part, as is clear from these figures, the drill (20) part and the background part can be clearly recognized. In the diagram on the left showing the sum of the amplitudes of each row in FIG. 10, the boundaries between the drill (20) and the background are found to be located at rows 340 and 1558, while in the diagram on the right showing the sum of the amplitudes of each column, the boundaries between the background and the drill (20) can be found at columns 560 and 1978. Then, after the center of the sum of the amplitudes obtained by FFT of each row in the section in which the drill 20 is shown (here, row 949) and the center of the sum of the amplitudes obtained by FFT of each column (here, column 1269) are found, the point where the pixel column (column 1269) and pixel row (row 949) located at the center intersect is found as the drill center of the drill 20.

In the present invention, the binarization unit 36 of the image processing device 30 uses the IFFT image (grayscale) in which pixel values are displayed in 256 gradations from 0 to 255, which is the processed part wear image extracted from the composite image as described above, to perform a binarization process on each pixel in which pixel values are displayed in two gradations, 0 and 255, to form a binarized image in which the worn part is composed of white pixels, and thus it is possible to measure the area, etc., using the number of 255 (white) pixels.

In this binarization process, a threshold value is selected to suppress detection of areas other than the worn areas and to more accurately identify the worn areas. Specifically, a histogram of pixel values in the IFFT image is obtained, and a predetermined threshold value is selected so that pixels in the background and pixels of the drill outline are removed (blackened). The binarization process described above is performed based on the selected threshold value to obtain a binarized image, an example of which is shown in FIG. 11(b). FIG. 11(a) shows a binarized image obtained without setting such a threshold value, and it can be seen that there are many white pixels in areas other than the worn areas.

Next, the binary image thus obtained is subjected to noise removal processing in accordance with the present invention in the noise removal section 38, and minute white pixel areas that do not correspond to the worn areas with a high density of white pixels are removed as noise from the binary image, thereby extracting a noise-processed image that more clearly displays the desired worn areas.

Specifically, in such noise removal processing, at least expansion/reduction processing and a process of filling in a small area of white pixels are performed on each pixel that constitutes the binary image. The expansion/reduction processing has the role of eliminating noise in the background and clarifying the worn parts, and generally, image processing using an opening filter and a closing filter is performed. The opening filter is a process of performing an expansion process after performing an equal number of contraction processes, and a closing filter is a process of performing an expansion process after performing an equal number of expansion processes. Note that the closing filter process fills in the thin lines inside the white parts, while the opening filter process fills in the thin white lines in the black background. In addition, when extracting such worn parts, these filters are used to make the worn parts clearer using a closing filter, and then to remove minute noise and the like other than the worn parts using an opening filter. Figure 12 (a) shows an example of an image obtained by performing an expansion/reduction process on such a binary image.

In addition, in the process of filling in tiny areas of white pixels, tiny areas of white pixels in the background of the binarized image are filled in black, just like the majority of the background (black). The tiny areas of white pixels to be filled in are judged to be noise or not, with a threshold set as the limit of the area that can be recognized, and the area below that threshold is filled in black, resulting in an image in which most of the noise has been removed, as shown in Figure 12 (b).

When filling in such tiny white pixels, if a large amount of tiny areas exists in the binarized image, the filling process takes a long time. Therefore, prior to the filling process, some noise is removed using a median filter, which advantageously reduces the program execution time required for noise processing. Here, median filter processing involves sorting the pixel values of the pixels of a kernel size set around the target pixel in ascending order and using the central pixel value as the pixel value of the target pixel to determine the pixel values of each pixel in the entire image, and a median filter processed image is obtained, which provides an image with reduced noise.

In addition, in the present invention, a phase correction process (operation) for correcting the phase of the noise-processed image obtained by the above noise processing is advantageously adopted. In this process, the center of the drill 20 obtained in the above-mentioned drill center detection process is used to rotate the noise-processed image around the drill center by a predetermined angle (for example, 1°), as shown in FIG. 13, and then the image at each rotation angle is divided into two equal parts, left and right, at the drill center, and the two divided images are compared, for example, by flipping the images left and right, and comparing the positions of the white pixels (comparing the number of positions that match), thereby correcting the phase of the noise-processed image so that the worn part extends horizontally. In this type of phase correction operation, even if the worn part is located in a direction perpendicular to the horizontal direction, there will be a large number of matching white pixels when comparing the images divided into two halves, so a pixel area of a specified width including the center of the phase-corrected image (for example, an area of 200 pixels including the horizontal center) is cropped, and the white pixels of the cropped image are counted. If the number of white pixels is equal to or greater than a threshold value, the image is rotated by 90 degrees, but if it is less than the threshold value, it is determined that the worn part is located horizontally.

Furthermore, since the noise-processed image obtained by phase correction as described above contains noise that is located on a circumference at a certain distance from the center and at a large angle from the horizontal position, as shown in Figure 14(a), threshold values are set for the distance and angle, respectively, and the white pixel areas that become noise are removed, resulting in an extracted image of the worn area, as shown in Figure 14(b).

In the present invention, the noise-processed image (wear-part extraction image) obtained as described above is used in the evaluation unit 40 of the image processing device 30, and the wear state of the tip processing part of the drill 20, which is the processing tool, can be quantitatively evaluated in terms of the wear-part area, wear-part width, etc., from the white pixels that constitute the wear parts in such an image. Furthermore, such evaluation results are displayed on the display unit or printed out by a printing unit (not shown).

For example, when measuring the area of worn parts, the number of white pixels in the noise-processed image (image extracted from worn parts) is counted, and the area of one pixel is multiplied by the number of white pixels, making it possible to measure the area of the white parts that become worn parts (wear area). This makes it possible to quantitatively evaluate the wear condition of the drill (machining tool).

In the above embodiment, the image processing device 30 that processes the image data output from the camera 28 is generally configured as a general-purpose computer, and the image synthesis unit 32, image extraction unit 34, binarization unit 36, noise removal unit 38, evaluation unit 40, etc. can be configured using the CPU, memory, and other LSIs of such a computer, and image processing is realized in software terms by image processing programs and numerical analysis programs loaded into the memory.

In addition, it is desirable to install the image acquisition means of the present invention, for example, an imaging unit including a ring light 22, a right-angle prism 24, a lens 26, and a camera 28 as shown in Fig. 2, in a processing machine that performs processing with a specified processing tool, an example of which is shown in Fig. 15. In this case, the drilling machine 50 performs a specified processing by rotating a drill 60 held by a drill chuck 58 with a drive mechanism of the processing machine body and moving it three-dimensionally (vertical movement and horizontal movement) with respect to a processing object 56 set on a processing table 54 arranged on the base 52, while an imaging unit 62 according to the present invention is arranged on the base 52, and when inspecting the tip processing part of the drill 60 as the processing progresses, the drill 60 is moved horizontally by the drive mechanism of the processing machine body to be positioned above the imaging unit 62, and inspection according to the present invention is performed. Therefore, by using the imaging unit 62 disposed within such a drilling machine 50 to capture images of the processed part to be inspected, the effort of attaching and detaching the drill 60 from the machine 50 can be eliminated, eliminating the time loss caused by such attachment and detachment, and advantageously avoiding a decrease in productivity.

By disposing the imaging unit 62 inside such a processing machine (50), the tip processing portion of a processing tool such as a drill 60 can be easily moved and photographed using the drive mechanism of the main body of the processing machine 50 so that the focal points are located at predetermined intervals in the tool axial direction from the tool tip toward the tool contour, thereby making it possible to easily and quickly obtain multiple tip images.

Although typical embodiments of the present invention have been described above in detail, it should be understood that these are merely examples, and that the present invention should not be interpreted in any way as being limited by the specific descriptions of such embodiments.

For example, in the illustrated embodiment, a drill has been described as the processing tool, but other rod-shaped rotating tools for cutting, such as taps and various end mills, can also be used.

In addition, when the camera (28) that photographs the tip of the machining tool (drill) used for a specific process such as drilling has a sufficient depth of field, in other words, when an image of the machining part that is in focus over the entire tip of the machining tool can be obtained, the image data from the camera (28) can be input directly to the image extraction unit 34 in the image processing device 30 shown in Figure 2, and the image synthesis unit 32 can be omitted.

In addition, in the embodiment shown in FIG. 15, the imaging unit 62 is arranged on the base 52, but it is also possible to adopt a structure in which the imaging unit 62 is housed in a form in which it is embedded within the base 52.

Furthermore, in terms of the structure of the imaging device (unit), in addition to the exemplified structure, it goes without saying that various structures that can be easily adopted by those skilled in the art can be used, such as capturing reflected light from the worn part (surface) of the tip machining part of a machining tool such as a drill and collecting an image of the machining part.

Although we will not list them all here, the present invention can be implemented in various forms with changes, modifications, improvements, etc. based on the knowledge of those skilled in the art, and it should be understood that all such embodiments fall within the scope of the present invention as long as they do not deviate from the spirit of the present invention.

REFERENCE SIGNS LIST 10 Image acquisition process 12 Image extraction process 14 Binarization process 16 Noise removal process 18 Evaluation process 20 Drill 22 Ring illumination 24 Right-angle prism 26 Lens 28 Camera 30 Image processing device 32 Image synthesis section 34 Image extraction section 36 Binarization section 38 Noise removal section 40 Evaluation section 50 Hole drilling machine 52 Base 54 Processing table 56 Workpiece 58 Drill chuck 60 Drill 62 Imaging unit

Claims (16)

an image taking step of taking an image of a tip processed portion of a processing tool used for a predetermined processing to take an image of the processed portion to be inspected;
An image extraction process in which a worn portion caused by the predetermined processing of the tip processed portion is extracted from the collected processed portion image by filtering using a Fourier transform, and taken out as a processed portion wear image;
a binarization step of binarizing each pixel of the extracted processed portion wear image based on a set predetermined threshold value to form a binary image in which the wear portion is composed of white pixels;
a noise removal process for removing minute white pixel areas that do not correspond to the worn areas having a high density of white pixels from the formed binary image as noise, thereby extracting a noise-processed image that more clearly displays the worn areas;
an evaluation step of evaluating the wear state of the tip processing portion of the processing tool from white pixels constituting the worn portion in the noise-processed image;
13. A method for inspecting a machining tool, comprising: The image acquisition step includes:
an imaging step of acquiring a plurality of tip images by photographing the tip processing portion of the processing tool such that focal points are positioned at predetermined intervals from the tool tip toward the tool contour portion in the tool axial direction;
An image synthesis process in which a single synthetic image in focus over the entire tip processing portion is formed as the processing portion image by a depth of field synthesis method using the multiple tip processing portion images;
2. The method for inspecting a machining tool according to claim 1, further comprising: The machining tool inspection method according to claim 2, characterized in that in the image synthesis process, each tip image obtained in the imaging process is divided vertically and horizontally into multiple parts to form a large number of divided images, and then the divided images of the same location in each tip image are compared to select the clearest divided image, thereby forming the single synthesized image. The machining tool inspection method according to claim 1, characterized in that in the image extraction process, the processed part image is subjected to a fast Fourier transform to form a spectral image, and then the spectral image is subjected to a masking process to leave high-frequency parts, and then an inverse fast Fourier transform is performed to extract the processed part wear image. The machining tool inspection method according to claim 1, characterized in that in the noise removal process, at least an expansion/reduction process and a small area white pixel filling process are performed on each pixel constituting the binary image. The method for inspecting a machining tool according to claim 1, further comprising a tool center detection step of determining the vertical and horizontal sections in the machining part image in which the machining tool is captured by plotting the sums of amplitudes obtained by fast Fourier transform of each column and row of pixels located in the vertical and horizontal directions of the machining part image, respectively, and then determining the point where the pixel columns and pixel rows located in the center of the vertical and horizontal sections intersect as the tool center of the machining tool. The method for inspecting a machining tool according to claim 6, further comprising a phase correction step of: rotating the noise-processed image by a predetermined angle around the tool center of the machining tool determined in the tool center detection step, dividing the image at each rotation angle into two equal parts, and comparing the two divided images to correct the phase of the noise-processed image so that the worn portion extends in the horizontal direction. The inspection method for machining tools according to claim 7, characterized in that the evaluation step is carried out after removing white pixel areas located on a circumference at a certain distance from the center of rotation and at a specified angle from the wear part in the horizontal position, using the noise-processed image that has been phase-corrected by the phase correction step. The method for inspecting a machining tool according to any one of claims 1 to 8, characterized in that the machining tool is a drill. The method for inspecting a machining tool according to any one of claims 1 to 8, characterized in that the machining tool is a diamond-coated drill. an image acquisition means for acquiring an image of a tip end of a processing tool used for a predetermined processing operation, and
An image extraction means for extracting a wear portion caused by the predetermined processing of the tip processing portion from the collected processed portion image by filtering using a Fourier transform, and taking it out as a processed portion wear image;
a binarization means for binarizing each pixel of the extracted processed portion wear image based on a set predetermined threshold value, and forming a binary image in which the wear portion is composed of white pixels;
a noise removing means for removing minute white pixel areas that do not correspond to the worn areas having a high density of white pixels from the formed binary image as noise, thereby extracting a noise-processed image that more clearly shows the worn areas; and
an evaluation means for evaluating a wear state of a tip processing portion of the processing tool from white pixels constituting a worn portion in the noise-processed image;
An inspection device for a machining tool comprising: The image acquisition means
a ring-shaped illumination means located in a circular shape around a tool axis below the processing tool and irradiating light onto a tip processing portion of the processing tool;
a right-angle prism for refracting, in a horizontal direction, light reflected downward from the tip processing portion of the processing tool by irradiation with light from the ring-shaped illumination means;
a camera that receives the light refracted by the right-angle prism and captures an image of the processed portion;
12. The inspection device for a machining tool according to claim 11, further comprising: The inspection device for a machining tool according to claim 11 or 12, characterized in that the image collection means is disposed in a machining machine that performs a predetermined machining process using the machining tool. The image acquisition means
an imaging device that captures a plurality of tip images by photographing the tip processing portion of the processing tool such that focal points are positioned at predetermined intervals from the tool tip toward the tool contour portion in the tool axial direction;
An image synthesis means for forming a single synthetic image in focus over the entire tip processing portion as the processing portion image by a depth of field synthesis method using the plurality of tip processing portion images;
12. The inspection device for a machining tool according to claim 11, further comprising: The imaging device,
a ring-shaped illumination means located in a circular shape around a tool axis below the processing tool and irradiating light onto a tip processing portion of the processing tool;
a right-angle prism for refracting, in a horizontal direction, light reflected downward from the tip processing portion of the processing tool by irradiation with light from the ring-shaped illumination means;
a camera that receives the light refracted by the right-angle prism and captures an image of the processed portion;
15. The inspection device for a machining tool according to claim 14, further comprising: The inspection device for machining tools according to claim 14 or 15, characterized in that the imaging device is disposed in a machining machine that performs a predetermined machining operation with the machining tool, and the focal position of the camera is moved at a predetermined interval from the tip of the tool toward the contour of the tool by the axial movement of the machining tool by the machining machine.

Images