JP2009270984A - Position detecting device, position detecting program, and position detecting method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a position detecting device capable of detecting the position of a transparent member by recognizing the shape of the transparent member even if there exists noises or the edge is imperfect. <P>SOLUTION: The position detecting device comprises: an image input part 2 acquiring the image of a plurality of transparent members, which are endlessly arranged so that respective faces confront to each other, photographed from the vertical direction of the surface; a ROI setting part 10 acquiring the area which involves at least each contour of the photographed plurality of transparent members in the image acquired by the image input part 2; an edge extracting part 11 extracting the edge of the plurality of the transparent members from the area acquired by the ROI setting part 10; an optimum edge selecting part 12 selecting the edge of the predetermined transparent member among the plurality of transparent members from the edges selected by the edge selecting part 11; and a position computing part 13 computing the position of the predetermined transparent member based on the edge selected by the optimum edge selecting part 12. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a technique for detecting the position of a transparent member in a non-contact manner. More specifically, a position detection device and a position detection program that detect a parameter related to the center position and size of a measurement target by photographing a transparent member with an image input unit such as a camera and performing image processing on the acquired image. The present invention relates to a position detection method.

  In the magnetic film forming apparatus, when the substrate is set from the atmosphere to the vacuum apparatus, it is necessary to accurately hold the substrate by a holder disposed in the vacuum apparatus in order to prevent the substrate from falling. An image sensor is effective for measuring the positional deviation between the substrate and the holder in a non-contact manner.

  Aluminum or glass is mainly used as a substrate material for forming a magnetic film of an HDD (Hard Disk Drive). When the position of the substrate is detected by the image sensor, pattern matching based on a silhouette shape is possible in the case of an aluminum substrate, but since a glass substrate transmits light, a silhouette is not formed and pattern matching cannot be applied. Particularly in recent years, glass substrates that are stronger and smoother than aluminum have become the mainstream, and the fact that pattern matching cannot be applied causes a decline in productivity, which is more serious. It has become a problem.

  In addition, as a conventional technique, in order to detect a transparent member by image processing, the edge of the member is detected by illuminating from the back of the transparent member or using light reflected from the front and detecting the position. A technique is disclosed. Also disclosed are a method for detecting the edge position by rotating a transparent wafer, and a method for detecting burrs and defects on the outer periphery and inner periphery of an O-ring shaped object with a double circular window.

JP 2002-158275 A JP-A-4-102050

  However, when the background changes, when there are a plurality of substrates and the edge of the back substrate is also extracted, and when the edge of the contour is further interrupted, the edge detection fails.

  In order to solve the above-mentioned problems, noise occurs due to changes in the background of the member, the edges of the substrate on the back side are extracted because there are multiple substrates, and the edge of the member outline is interrupted, resulting in incomplete edges Even if it is, it aims at providing the position detection apparatus which can recognize the edge shape of a transparent member, and can detect the position of a transparent member, a position detection program, and a position detection method.

  The position detection device captures an image acquired from the vertical direction of the surface with respect to a plurality of transparent members arranged so that the surfaces face each other, and an image acquired by the image acquisition unit. An area acquisition unit that acquires an area including at least the contours of each of the plurality of transparent members, an edge extraction unit that extracts edges of the plurality of transparent members from the area acquired by the area acquisition unit, An edge selection unit that selects an edge of a predetermined transparent member among the plurality of transparent members from edges extracted by the edge extraction unit, and the edge of the predetermined transparent member based on the edge selected by the edge selection unit A position calculation unit that calculates a position.

  In addition, the position detection program includes an image acquisition step for acquiring an image captured from a direction perpendicular to the surface with respect to a plurality of transparent members arranged so that the surfaces face each other, and an image acquired by the image acquisition step. An area acquisition step of acquiring an area including at least the contours of each of the plurality of transparent members imaged; an edge extraction step of extracting edges of the plurality of transparent members from the area acquired by the area acquisition step; , An edge selection step for selecting an edge of a predetermined transparent member among the plurality of transparent members from the edges extracted by the edge extraction step, and the predetermined transparent based on the edge selected by the edge selection step And causing the computer to execute a position calculating step of calculating the position of the member.

  Further, the position detection method includes an image acquisition step of acquiring an image captured from a direction perpendicular to the surface with respect to a plurality of transparent members arranged so that the surfaces face each other, and an image acquired by the image acquisition step. An area acquisition step of acquiring an area including at least the contours of each of the plurality of transparent members imaged; an edge extraction step of extracting edges of the plurality of transparent members from the area acquired by the area acquisition step; , An edge selection step for selecting an edge of a predetermined transparent member among the plurality of transparent members from the edges extracted by the edge extraction step, and the predetermined transparent based on the edge selected by the edge selection step And a position calculating step for calculating the position of the member.

  Even if there is noise or an incomplete edge on the image of the transparent member, the shape of the transparent member can be recognized and the position of the transparent member can be detected.

(Embodiment 1)
A position detection apparatus according to Embodiment 1 is shown in FIG. In the first embodiment, it is assumed that a detection target is a circular donut-shaped transparent substrate (transparent member) used for a CD, a DVD, or the like, for example, a glass substrate used for an HDD. In the first embodiment, the position (center position) of the frontmost transparent substrate with respect to the camera to be imaged is detected.

  The position detection apparatus 1 includes an image input unit 2 (image acquisition unit), an image processing unit 3, and a display 4 that displays an input image and the like. The position detection device 1 also includes a keyboard 5 that is means for receiving input from the user such as numerical input, and a camera 6 that captures an image of the transparent substrate.

  The image input unit 2 is connected to the keyboard 5, the display 4, and the camera 6, and includes an input / output control unit 14 that controls data to be input / output. Further, the image input unit 2 includes a control unit 15 that controls input / output of all data in the position detection apparatus 1 including input / output of the input / output control unit 14. The image input unit 2 uses the input / output control unit 14 and the control unit 15 to acquire an image sent from the camera 6 at a constant sampling time interval (for example, 33 ms), and sends the acquired image to the image processing unit 3. . In the first embodiment, the image input unit 2 acquires images from the camera 6. However, the image input unit 2 may acquire images from media such as videos and DVDs, and a moving image file stored on the hard disk. You may get it.

  An image input by the image input unit 2 is, for example, a rectangle of horizontal 640 pixels × vertical 480 pixels, and in the case of a color image, a red component (R), a green component (G), and a blue component (B ) Gradation value. In the case of a monochrome image, a gradation value (luminance) of brightness is given to one pixel. For example, the gradation values of the red component (R), green component (G), blue component (B), and luminance (I) of the pixel at coordinates (i, j) indicated by integers i and j are respectively digital values R ( i, j), G (i, j), B (i, j), I (i, j), and the like. In this embodiment, the luminance (black and white gray value) is used as the pixel value used in the calculation. However, only R, G only, B only, or a combination of RGB and other RGB values obtained by converting RGB values are used. A color system (HSV color system, YUV color system, etc.) may be used.

  Here, FIG. 2 shows an arrangement state of imaging of the transparent substrate and an image example of the transparent substrate to be imaged. As shown in FIG. 2A, when two transparent substrates arranged so that their surfaces face each other are imaged from the direction perpendicular to the front surface of the substrate, the circumference of the front substrate is large, and the circle of the back substrate is large. The circumference is taken small. That is, an image as shown in FIG. 2B is obtained. In the first embodiment, the distance between the transparent substrates is about 10 mm. However, it is only necessary to ensure a distance that does not allow the edges to intersect at the time of edge extraction described later.

  The description returns to the configuration diagram of FIG. The image processing unit 3 includes a ROI (Region of Interest) setting unit 10 (region acquisition unit), an edge extraction unit 11, an optimum edge selection unit 12 (edge selection unit), and a position calculation unit 13. Each unit in the image processing unit 3 is realized by cooperation between a program stored in advance in the memory 102 as a storage medium and hardware resources such as a CPU (Central Processing Unit) 101 and the memory 102. .

The ROI setting unit 10 acquires a donut-shaped ROI (region) including at least the outlines of the two transparent substrates on the image, either manually or automatically, and sets the ROI on the image. FIG. 3 shows an example of setting ROI. Since the ROI is set and only the inside of the ROI is a processing target, the processing time is shortened, and the influence of noise such as a change in the background of the transparent substrate is reduced. When the ROI is manually set, the ROI setting unit 10 acquires the ROI range by inputting the next parameter from the keyboard 5 by the user.
(1) Inner and outer circle radii r1 and r2
(2) Center coordinates of inner circle and outer circle (i _ROI , j _ROI )

  Alternatively, when the position detection device 1 is provided with an input device such as a mouse, the ROI setting unit 10 draws an inner circle and an outer circle on the image displayed on the display 4 so that the ROI range is drawn. May be obtained. Note that the size of the ROI to be set is set to such an extent that the outline of the substrate does not deviate from the ROI region even when the position of the substrate is shifted.

その二乗和の平方を画素値として用いる。

The description returns to the configuration diagram of FIG. The edge extraction unit 11 performs an edge extraction process on the two transparent substrates on the image for which the ROI is set. There are various methods for extracting an edge using a difference in brightness in an image. In the first embodiment, a Sobel filter that is a typical edge extraction method is employed. In the Sobel filter, products f _A and f _B of a matrix composed of 3 × 3 pixels in the vertical and horizontal directions of the pixel (i, j) of interest and the following matrix are calculated,

The square of the square sum is used as the pixel value.

  FIG. 4 shows an example of an edge image extracted by the edge extraction unit 11. In the first embodiment, the edge extraction processing is performed based on 3 × 3 pixels on the top, bottom, left, and right, but the mode is not limited, and may be, for example, 5 × 5 pixels or 7 × 7 pixels.

  The description returns to the configuration diagram of FIG. The optimum edge selection unit 12 selects the edge (optimum edge) of the frontmost transparent substrate for the camera 6 from the edges extracted by the edge extraction unit 11. In FIG. 4, the edge of the front substrate and the edge of the back substrate are not separated, and it is unclear which substrate contains the extracted edge. Therefore, the optimum edge selection unit 12 extracts only the edge of the front substrate by gradually narrowing down the ROI.

上式を次のように変形し、点（Ｘｎ，Ｙｎ）を代入し、２乗の総和を求める。

ただし、α＝−２ｉ_c、β＝−２ｊ_c、γ＝ｉ_c ²＋ｊ_c ²−γ_c ²

次に、α、β、γに関して偏微分する。

上式をα，β，γに関して解くと、以下の式を得る。

上式に、点列（Ｘｎ，Ｙｎ）の値を代入して逆行列を求め、α、β、γを算出すると、中心座標と半径が求まる。

図５に円フィッティングの結果の例を示す。 Here, the details of the processing of the optimum edge selection unit 12 will be described. First, using the extracted point sequence (Xn, Yn) on all edges, an approximation process by the least square method (hereinafter referred to as a circle fitting if necessary) is performed (n = 1,..., N). . Assuming that the center and radius of the circle to be obtained are (i _C , j _C ) and r _C , the circle equation is as follows.

The above equation is modified as follows, and the point (Xn, Yn) is substituted to find the sum of squares.

However, α = −2i _c , β = −2j _c , γ = i _c ² + j _c ² −γ _c ²

Next, partial differentiation is performed with respect to α, β, and γ.

Solving the above equation with respect to α, β, γ, the following equation is obtained.

By substituting the value of the point sequence (Xn, Yn) into the above equation to obtain an inverse matrix and calculating α, β, and γ, the center coordinates and the radius are obtained.

FIG. 5 shows an example of the result of circle fitting.

Next, the optimum edge selection unit 12 resets the donut-shaped ROI with reference to the circle obtained by the above-described circle fitting process. That is, the following ROIs are set as inner and outer circles.
Inner circumference circle: center coordinates: (i _C , j _C ), radius: r _C -L ₁
Peripheral circle: center coordinates: (i _C , j _C ), radius: r _C + L ₂
Here, L ₂ > L ₁ > 0

The position detection apparatus 1 performs the edge extraction process and the circle fitting process again in the same manner as described above using the reset ROI. In this way, by setting L ₂ larger than L _1, it becomes possible to perform edge extraction that excludes the inner edge of the reference circle (radius r _C ) and extracts the outer edge (FIG. 6). . Conversely, by setting L ₂ <L ₁ , it is possible to exclude the outer edge and extract the inner edge.

図７に最適エッジの抽出例を示す。 The position detection apparatus 1 performs the above-described processing k times a predetermined number of times k, and calculates a radius r _C ^{(K) as} a result of the circle fitting and a k−1th radius r _C ^(K-1). Calculate the difference between When the difference is equal to or less than the threshold value r _TH , the extracted edge is regarded as the optimum edge (the edge of the substrate on the forefront from the imaging position).

FIG. 7 shows an example of optimum edge extraction.

The description returns to the configuration diagram of FIG. The position calculation unit 13 regards the center position of the circle fitting obtained when extracting the optimum edge as the center position of the front substrate.
(I _C ^(K) , j _C ^(K) )

When a series of processes by each unit described above is completed and the next sampling time is reached, the camera 6 captures the next frame image. When the substrate position does not change, the above-described value can be used as the ROI setting value for the next frame image. That is, the ROI is set as follows for the next frame image.
Inner circumference: center coordinates: (i _C ^(K) , j _C ^(K) ), radius: r _C ^(K) −L ₁
Peripheral circle: center coordinates: (i _C ^(K) , j _C ^(K) ), radius: r _C ^(K) + L ₂
L ₂ > L ₁ > 0
When the substrate position changes, the ROI is set to be large so that the substrate image does not deviate from the ROI.
Inner circumference: center coordinates: (i _C ^(K) , j _C ^(K) ), radius: r _C ^(K) −L ₃
Peripheral circle: center coordinates: (i _C ^(K) , j _C ^(K) ), radius: r _C ^(K) + L ₄
L ₃ > L ₁ , L ₄ > L ₂
Normally, it is better to set a larger ROI in consideration of detection errors and the like. When the ROI is automatically set using the substrate position (i _C ^(K) , j _C ^(K) ) detected in this way, manual ROI setting only needs to be performed for the first frame image. The load can be reduced.

位置検出装置１は、エッジを形成する点の総数Ｎが円周長さの一定の割合p_THに到達しないときは、判定基準を満足しないものとみなす。
Ｎ＜ｌ×p_TH ・・・ 判定基準を満たさない。
Ｎ≧ｌ×p_TH ・・・ 判定基準を満たす。 In addition, when the number of edge pixels is small, such as when the edge of the substrate extracted by the edge extraction unit 11 is interrupted, the reliability of the detected center position and radius may be low. Therefore, the position detection device 1 performs processing so that the detection result is not adopted when the following determination criterion is not satisfied. At the time of pre-sampling, the circumference length l when the criterion is satisfied is calculated from the radius r _C ^(K) .

The position detection device 1 considers that the criterion is not satisfied when the total number N of points forming the edge does not reach a constant ratio _pTH of the circumferential length.
N <l × p _TH: Does not meet the criteria.
N ≧ l × p _TH・ ・ ・ Meets the criteria.

  Next, the process of the position detection apparatus 1 in Embodiment 1 is demonstrated, referring the flowchart of FIG.

  The camera 6 starts photographing two continuous transparent substrates (step S1). The camera 6 captures an image from a direction perpendicular to the surface of the transparent substrate at a predetermined sampling period interval.

  The image input unit 2 acquires a frame image with a sampling period captured by the camera 6 (step S2).

  The ROI setting unit 3 acquires the center positions and radii of the inner and outer circles of the ROI so as to be an area including at least the outline of each transparent substrate (step S3). Note that the center position and radius of the inner and outer circles of the ROI are input by the user via the keyboard 5 only for the first processing immediately after the position detection device 1 is activated.

  The ROI setting unit 3 sets the ROI acquired in step S3 from the frame image acquired by the image input unit 2 to the frame image (step S4).

  The edge extraction unit 11 applies the above-described edge extraction processing to the set ROI (step S5).

  The optimum edge selection unit 12 acquires all the point sequences (Xn, Yn) of the extracted edges (step S6), and performs the above-described circle fitting using the point sequence (Xn, Yn) to obtain the reference circle. Calculate (step S7).

The optimum edge selection unit 12 calculates the center coordinates (i _C ^(K) , j _C ^(K) ) and the radius r _C ^(K) of the reference circle (step S8).

The optimum edge selector 12 calculates the difference between the radius r _C ^(K) and the k−1th radius r _C ^(K−1) (step S9). If the difference is greater than or equal to the threshold value r _TH (step S9, N), the inner and outer circles of ROI are reset as follows, and the process returns to step S3 (step S10).
Inner circumference: center coordinates: (i _C ^(K) , j _C ^(K) ), radius: r _C ^(K) −L ₁
Peripheral circle: center coordinates: (i _C ^(K) , j _C ^(K) ), radius: r _C ^(K) + L ₂
In step S3, the ROI setting unit 10 acquires the center coordinates and the radius calculated in step S10 as the ROI range.

The optimum edge selection unit 12 can further narrow down the ROI range by gradually reducing the values of L ₁ and L ₂ every time the loop from step S10 to step S3 is performed.

On the other hand, when the difference is smaller than the threshold value r _TH (step S9, Y), the optimum edge selection unit 12 determines whether or not the total number N of points forming the edge has reached a constant ratio p _TH of the circumferential length. Is determined (step S11). When the determination criterion is not satisfied (step S11, N), the optimum edge selection unit 12 resets the inner and outer circles of the ROI based on the calculation result at the time of the previous sampling (step S12), and the processing is performed on the next frame. Transition to image acquisition (step S2).

On the other hand, when the determination criterion is satisfied (step S11, Y), the position calculation unit 13 detects the center position (i _C ^(K) , j _C ^(K) ) of the substrate (step S13). Thereafter, the position calculation unit 13 sets the inner and outer circles of the ROI as follows, and the process transitions to acquisition of the next frame image (step S2).
Inner circumference: center coordinates: (i _C ^(K) , j _C ^(K) ), radius: r _C ^(K) −L ₃
Peripheral circle: center coordinates: (i _C ^(K) , j _C ^(K) ), radius: r _C ^(K) + L ₄
In step S3 for the next frame image, the ROI setting unit 10 acquires the center coordinates and radius of the inner and outer circles calculated in step S14 as the ROI range.

  The above steps are executed for the frame image captured by the camera 6.

In the above flowchart, after obtaining the result of the circle fitting and resetting the ROI using L ₁ and L ₂ , the process loops to the ROI setting unit 10 and the edge extraction unit 11, but the ROI is reset. After that, the optimum edge selection unit 12 may perform the process before the processing by the position calculation unit 13.

(Embodiment 2)
The optimum edge selection unit 12 according to the first embodiment repeatedly performs ROI resetting, edge extraction, and approximation by the least square method until the determination criterion is satisfied, thereby obtaining the optimum edge (the edge of the transparent substrate that is the earliest with respect to the imaging position). ) Has been calculated, the optimum edge selection unit 12 of the second embodiment selects the optimum edge by line scanning. In the second embodiment, the configuration of the position detection device 1 is the same as that of the first embodiment, and the description thereof is omitted (see FIG. 1). In addition, since the conditions such as the detection target and the imaging arrangement are the same as those in the first embodiment, the description thereof is omitted (see FIG. 2).

  Here, a method of selecting the optimum edge by line scanning will be described with reference to FIG. For the edges extracted by the edge extraction unit 11, a point sequence of continuous edges is expressed as (X (m) n, Y (m) n). Here, m indicates that it is included in the mth edge, and n indicates the nth point in the mth edge. Further, in the following description, a processing method when scanning with a vertical line is shown, but when scanning with a horizontal line, the same procedure can be used.

A vertical line is scanned from the left edge (i = 0) to the right edge (i = imax) of the image. When the vertical line is i, the coordinates of the intersection with the edge m intersecting the vertical line are
(I, Y (m) n)
It is. The optimum edge selection unit 12 calculates the maximum and minimum values of the y coordinate among the intersection coordinates with all edges, and registers the corresponding edge numbers in the upper outermost edge candidate and lower outermost edge candidate (memory) 102). For example, in FIG. 9, when the vertical line is i = i ₁ , the maximum value and the minimum value are max {Y (1) n,..., Y (m) n} = Y (1) n.
min {Y (1) n,..., Y (m) n} = Y (1) n
It becomes. Note that max {} and min {} are operators for calculating the maximum value and the minimum value, respectively. Therefore, m = 1 is registered in the upper outermost peripheral edge candidate and the lower outermost peripheral edge candidate.
Upper outermost edge candidate [1]
Lower outermost edge candidate [1]

Similarly, when the vertical line in FIG. 9 is i = i ₂ ,
max {Y (1) n,..., Y (m) n} = Y (3) n
min {Y (1) n,..., Y (m) n} = Y (3) n
Thus, the optimum edge selection unit 12 registers edge number 3 corresponding to the newly obtained maximum value and minimum value.
Upper outermost edge candidate [1,3]
Lower outermost edge candidate [1,3]

Furthermore, in FIG. 9, when the vertical line is i = i ₃ ,
max {Y (1) n,..., Y (m) n} = Y (2) n
min {Y (1) n,..., Y (m) n} = Y (4) n
It becomes. At this time, the third edge (m = 3 edge) also intersects the vertical line i ₃ , but does not correspond to either the maximum value or the minimum value. In this case, the optimum edge selection unit 12 deletes 3 from the upper outermost peripheral edge candidate and the lower outermost peripheral edge candidate, and registers the edge numbers corresponding to the newly calculated maximum value and minimum value.
Upper outermost edge candidate [1,2]
Lower outermost peripheral edge candidate [1, 4]

By repeating the above processing, only the outermost peripheral edge can be extracted. The extraction result of FIG. 9 is as follows.
Upper outermost edge candidate [1, 2, 5]
Lower outermost peripheral edge candidate [1, 4, 9]
The edge finally registered in the upper outermost peripheral edge candidate and the lower outermost peripheral edge candidate is the optimum edge. The optimum edge extracted from the edges in FIG. 9 is as shown in FIG.

Next, processing of the position detection apparatus 1 according to Embodiment 2 will be described with reference to the flowchart of FIG. In the flowchart of FIG. 11, in addition to the method of selecting the optimum edge by line scanning, it is assumed that a circle fitting process is also performed. In this way, more accurate center coordinates (i _C , j _C ) and radius r _C of the substrate are obtained.

  Steps S21 to S25 are the same as those in the first embodiment, and thus description thereof is omitted (see steps S1 to S5 in FIG. 8).

  The optimum edge selection unit 12 acquires a point sequence (X (m) n, Y (m) n) of each continuous edge (step S26).

  The optimum edge selection unit 12 sets an initial value 0 to the coordinate value i of the vertical line to be scanned (step S27).

  The optimum edge selection unit 12 calculates the maximum value and the minimum value of the y coordinate from the intersection coordinates of all edges that intersect the vertical line. (Step S28) Further, the optimum edge selection unit 12 registers the edge numbers corresponding to the calculated maximum value and minimum value in the upper outermost edge candidate and the lower outermost edge candidate, respectively (Step S29).

  When the edge number corresponding to the y coordinate other than the maximum value and the minimum value is registered in the upper outermost edge candidate and the lower outermost edge candidate, the optimum edge selecting unit 12 deletes the registration of the edge number. (Step S30).

  The optimum edge selection unit 12 adds 1 to the coordinate value i of the vertical line (step S31). Here, if the coordinate value i is less than imax (rightmost end) (step S32, N), the process proceeds to step S28, and the processes from step S28 to step S31 are repeated until the coordinate value i reaches imax.

  On the other hand, when the coordinate value i has reached imax (step S32, Y), the optimum edge selection unit 12 uses a point sequence (X (m) n) registered in the upper outermost edge candidate and the lower outermost edge candidate. , Y (m) n), circular fitting is performed in the same manner as in the first embodiment (step S33).

The optimum edge selection unit 12 calculates the center coordinates (i _C , j _C ) and the radius r _C by the same process as step S8 in the first embodiment (step S34).

  Since the subsequent processing from step S35 to step S38 is the same as that of the first embodiment, description thereof will be omitted (see steps S11 to S14).

  In both Embodiment 1 and Embodiment 2, the detection target is a circular donut-shaped transparent substrate, but the shape is not limited. For example, a rectangular substrate can be applied.

  A program that causes a computer constituting the position detection apparatus to execute the above-described steps can be provided as a position detection program. By storing the above-described program in a computer-readable recording medium, the computer constituting the position detection device can be executed. Here, examples of the recording medium readable by the computer include an internal storage device such as a ROM and a RAM, a portable storage such as a CD-ROM, a flexible disk, a DVD disk, a magneto-optical disk, and an IC card. It includes a medium, a database holding a computer program, another computer and its database, and a transmission medium on a line.

(Additional remark 1) The image acquisition part which acquires the image imaged from the perpendicular | vertical direction of the said surface with respect to the several transparent member arrange | positioned so that a surface may face each other,
An area acquisition unit for acquiring an area including at least the contours of each of the plurality of transparent members captured in the image acquired by the image acquisition unit;
An edge extraction unit that extracts edges of the plurality of transparent members from the region acquired by the region acquisition unit;
An edge selection unit that selects an edge of a predetermined transparent member among the plurality of transparent members from the edges extracted by the edge extraction unit;
A position calculating unit that calculates the position of the predetermined transparent member based on the edge selected by the edge selecting unit;
A position detection device comprising:
(Supplementary note 2) In the position detection device according to supplementary note 1,
The edge selection unit calculates a reference shape based on the edge extracted by the edge extraction unit, and further narrows down the range of the region acquired by the region acquisition unit based on the shape, thereby the predetermined transparent A position detecting device for selecting an edge of a member.
(Supplementary note 3) In the position detection device according to supplementary note 2,
The position detection device, wherein the edge selection unit calculates the reference shape by approximating all edges extracted by the edge extraction unit by a least square method.
(Additional remark 4) In the position detection apparatus of Additional remark 1,
When the predetermined transparent member is the frontmost transparent member with respect to the imaging position, the edge selection unit scans a horizontal or vertical line of the image, and a scan line and an edge extracted by the edge extraction unit are detected. A position detection apparatus characterized by obtaining an edge having an intersection point coordinate having a maximum value or a minimum value, and selecting the edge as an edge of the predetermined transparent member.
(Additional remark 5) In the position detection apparatus of Additional remark 4,
The edge selector is
Scan a horizontal or vertical line of the image, register an edge where the intersection coordinate between the scanning line and the edge extracted by the edge extraction unit is the maximum value as a first edge candidate,
Register an edge at which the intersection coordinate between the scanning line and the edge extracted by the edge extraction unit is the minimum value as a second edge candidate,
When an edge whose intersection coordinate between a scanning line and the edge extracted by the edge extraction unit is not the maximum value or the minimum value is registered as the first edge candidate or the second edge candidate, the registration of the edge is deleted. ,
Edges registered as the first edge candidate and the second edge candidate when the first edge candidate registration, the second edge candidate registration, and the edge deregistration process are performed on all the scanning lines of the image. The position detecting device is selected as an edge of the predetermined transparent member.
(Supplementary note 6) In the position detection device according to supplementary note 5,
The edge selection unit further performs the first edge candidate registration, the second edge candidate registration, and the edge deregistration processing on the scanning lines of all the images when the first edge candidate and the second edge candidate are registered. A position detection apparatus, wherein a shape obtained by approximating all edges registered as edge candidates by a least square method is selected as an edge of the predetermined transparent member.
(Supplementary Note 7) An image acquisition step of acquiring images captured from a direction perpendicular to the surface with respect to a plurality of transparent members arranged so that the surfaces face each other;
An area acquisition step of acquiring an area including at least the outline of each of the plurality of transparent members captured in the image acquired by the image acquisition step;
An edge extraction step of extracting edges of the plurality of transparent members from within the region acquired by the region acquisition step;
An edge selection step of selecting an edge of a predetermined transparent member among the plurality of transparent members from the edges extracted by the edge extraction step;
A position calculating step for calculating a position of the predetermined transparent member based on the edge selected by the edge selecting step;
Position detection program that causes a computer to execute.
(Supplementary note 8) In the position detection program according to supplementary note 7,
The edge selection step calculates a reference shape based on the edge extracted by the edge extraction step, and further narrows down the range of the region acquired by the region acquisition unit based on the shape, thereby the predetermined transparent A position detection program for selecting an edge of a member.
(Supplementary note 9) In the position detection program according to supplementary note 8,
The position selecting program calculates the reference shape by approximating all the edges extracted by the edge extracting step by a least square method.
(Supplementary note 10) In the position detection program according to supplementary note 7,
When the predetermined transparent member is the transparent member that is the frontmost with respect to the imaging position, the edge selection step scans a horizontal or vertical line of the image, and a scan line and an edge extracted by the edge extraction step are A position detection program characterized by obtaining an edge having an intersection point coordinate having a maximum value or a minimum value and selecting the edge as an edge of the predetermined transparent member.
(Supplementary note 11) In the position detection program according to supplementary note 10,
The edge selection step includes:
Scan a horizontal or vertical line of the image, register an edge where the intersection coordinate between the scanning line and the edge extracted by the edge extraction step is a maximum value as a first edge candidate,
Register an edge having a minimum intersection coordinate between the scanning line and the edge extracted by the edge extraction step as a second edge candidate;
If an edge whose intersection coordinates between the scanning line and the edge extracted by the edge extraction step are not the maximum value or the minimum value is registered as the first edge candidate or the second edge candidate, the registration of the edge is deleted. ,
Edges registered as the first edge candidate and the second edge candidate when the first edge candidate registration, the second edge candidate registration, and the edge deregistration process are performed on all the scanning lines of the image. A position detection program that selects an edge of the predetermined transparent member.
(Supplementary note 12) In the position detection program according to supplementary note 11,
In the edge selection step, the first edge candidate and the second edge candidate are registered when the first edge candidate registration, the second edge candidate registration, and the edge deregistration process are performed on all scanning lines of the image. A position detection program that selects, as an edge of the predetermined transparent member, a shape obtained by approximating all edges registered as edge candidates by a least square method.
(Appendix 13)
An image acquisition step of acquiring an image captured from a direction perpendicular to the surface with respect to a plurality of transparent members arranged so that the surfaces face each other;
An area acquisition step of acquiring an area including at least the outline of each of the plurality of transparent members captured in the image acquired by the image acquisition step;
An edge extraction step of extracting edges of the plurality of transparent members from within the region acquired by the region acquisition step;
An edge selection step of selecting an edge of a predetermined transparent member among the plurality of transparent members from the edges extracted by the edge extraction step;
A position calculating step for calculating a position of the predetermined transparent member based on the edge selected by the edge selecting step;
Position detection method to execute.
(Supplementary note 14) In the position detection method according to supplementary note 13,
The edge selection step calculates a reference shape based on the edge extracted by the edge extraction step, and further narrows down the range of the region acquired by the region acquisition unit based on the shape, thereby the predetermined transparent A position detection method comprising selecting an edge of a member.
(Supplementary note 15) In the position detection method according to supplementary note 14,
The position detection method characterized in that the edge selection step calculates the reference shape by approximating all edges extracted by the edge extraction step by a least square method.
(Supplementary note 16) In the position detection method according to supplementary note 13,
When the predetermined transparent member is the transparent member that is the frontmost with respect to the imaging position, the edge selection step scans a horizontal or vertical line of the image, and a scan line and an edge extracted by the edge extraction step A position detection method characterized in that an edge having a maximum or minimum intersection coordinate is obtained and the edge is selected as an edge of the predetermined transparent member.
(Supplementary note 17) In the position detection method according to supplementary note 16,
The edge selection step includes:
Scan a horizontal or vertical line of the image, register an edge where the intersection coordinate between the scanning line and the edge extracted by the edge extraction step is a maximum value as a first edge candidate,
Register an edge having a minimum intersection coordinate between the scanning line and the edge extracted by the edge extraction step as a second edge candidate;
If an edge whose intersection coordinate between the scanning line and the edge extracted by the edge extraction step is not the maximum value or the minimum value is registered as the first edge candidate or the second edge candidate, the registration of the edge is deleted. ,
Edges registered as the first edge candidate and the second edge candidate when the first edge candidate registration, the second edge candidate registration, and the edge deregistration process are performed on all the scanning lines of the image. A position detecting method comprising selecting the edge of the predetermined transparent member.
(Supplementary note 18) In the position detection method according to supplementary note 17,
In the edge selection step, the first edge candidate and the second edge candidate are registered when the first edge candidate registration, the second edge candidate registration, and the edge deregistration process are performed on all scanning lines of the image. A position detection method, wherein a shape obtained by approximating all edges registered as edge candidates by the least square method is selected as an edge of the predetermined transparent member.

It is a figure which shows an example of a structure of the position detection apparatus based on embodiment. It is a figure which shows an example of the arrangement state of the imaging of a transparent substrate based on embodiment, and the image of the transparent substrate imaged. 6 is a diagram illustrating an example of ROI setting according to Embodiment 1. FIG. It is a figure which shows an example of the edge image extracted by the edge extraction part based on Embodiment 1. FIG. It is a figure which shows an example of the result of a circle fitting based on Embodiment 1. FIG. It is a figure which shows an example of the inner periphery circle of ROI based on the result of circle fitting based on Embodiment 1, and an outer periphery circle. 6 is a diagram illustrating an example of an optimum edge according to Embodiment 1. FIG. 6 is a flowchart illustrating an example of processing of the position detection device according to the first embodiment. FIG. 10 is a diagram for explaining an example of a method for selecting an optimum edge by line scanning according to the second embodiment. FIG. 10 is a diagram illustrating an example of an optimum edge according to the second embodiment. 10 is a flowchart illustrating an example of processing of the position detection device according to the second embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Position detection apparatus, 2 Image input part, 3 Image processing part, 4 Display, 5 Keyboard, 6 Camera, 10 ROI setting part, 11 Edge extraction part, 12 Optimal edge selection part, 13 Position calculation part, 14 Input / output control part , 15 control unit, 101 CPU, 102 memory.

Claims (6)

An image acquisition unit for acquiring an image captured from a direction perpendicular to the surface with respect to a plurality of transparent members arranged so that the surfaces face each other;
An area acquisition unit for acquiring an area including at least the contours of each of the plurality of transparent members captured in the image acquired by the image acquisition unit;
An edge extraction unit that extracts edges of the plurality of transparent members from the region acquired by the region acquisition unit;
An edge selection unit that selects an edge of a predetermined transparent member among the plurality of transparent members from the edges extracted by the edge extraction unit;
A position calculating unit that calculates the position of the predetermined transparent member based on the edge selected by the edge selecting unit;
A position detection device comprising: The position detection device according to claim 1,
The edge selection unit calculates a reference shape based on the edge extracted by the edge extraction unit, and further narrows down the range of the region acquired by the region acquisition unit based on the shape, thereby the predetermined transparent A position detecting device for selecting an edge of a member. The position detection device according to claim 2,
The position detection device, wherein the edge selection unit calculates the reference shape by approximating all edges extracted by the edge extraction unit by a least square method. The position detection device according to claim 1,
When the predetermined transparent member is the frontmost transparent member with respect to the imaging position, the edge selection unit scans a horizontal or vertical line of the image, and a scan line and an edge extracted by the edge extraction unit are detected. A position detection apparatus characterized by obtaining an edge having an intersection point coordinate having a maximum value or a minimum value, and selecting the edge as an edge of the predetermined transparent member. An image acquisition step of acquiring an image captured from a direction perpendicular to the surface with respect to a plurality of transparent members arranged so that the surfaces face each other;
An area acquisition step of acquiring an area including at least the outline of each of the plurality of transparent members captured in the image acquired by the image acquisition step;
An edge extraction step of extracting edges of the plurality of transparent members from within the region acquired by the region acquisition step;
An edge selection step of selecting an edge of a predetermined transparent member among the plurality of transparent members from the edges extracted by the edge extraction step;
A position calculating step for calculating a position of the predetermined transparent member based on the edge selected by the edge selecting step;
Position detection program that causes a computer to execute. An image acquisition step of acquiring an image captured from a direction perpendicular to the surface with respect to a plurality of transparent members arranged so that the surfaces face each other;
An area acquisition step of acquiring an area including at least the outline of each of the plurality of transparent members captured in the image acquired by the image acquisition step;
An edge extraction step of extracting edges of the plurality of transparent members from within the region acquired by the region acquisition step;
An edge selection step of selecting an edge of a predetermined transparent member among the plurality of transparent members from the edges extracted by the edge extraction step;
A position calculating step for calculating a position of the predetermined transparent member based on the edge selected by the edge selecting step;
Position detection method to execute.

Images