JP2007285869A - Surface inspection system and surface inspection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a surface inspection system and a surface inspection method which can detect flaws on the surface of an object to be inspected, such as a pipe assembly with high accuracy and decide the quality of the object to be inspected, on the basis of the detection result with high accuracy. <P>SOLUTION: The surface inspection system (1) comprises a two-dimensional detector (24) for acquiring an inspection image, obtained by photographing the area, including an inspection area of the object to be inspected, a flaw detection section (33) for detecting information on the flaws of the object to be inspected, on the basis of the inspection image, and a quality decision section (34) for deciding the quality of the object to be inspected, on the basis of the information. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a surface inspection apparatus and a surface inspection method, and more specifically, the quality of an inspection object is checked by examining the presence or absence of scratches on the surface of the inspection object based on an image obtained by photographing the inspection object. The present invention relates to a surface inspection apparatus and a surface inspection method for performing determination.

  In the manufacturing process, parts may be damaged in processes such as parts processing and assembly. Therefore, the manufacturing process is generally provided with an inspection process to eliminate damaged parts. For example, a manufacturing process of a pipe ASSY used for an injector will be described as an example. In the manufacturing process of the pipe assembly, the pipe assembly may be held by a chuck when the pipe assembly is transported to another process such as grinding of the inner diameter of the pipe and grinding of the end face. However, if the chuck grasps the pipe ASSY at an inappropriate position, the surface of the O-ring contact portion of the connector connection portion of the pipe ASSY may be damaged. Here, when a pipe assembly having a scratch on the surface of the contact portion of the O-ring is used, the O-ring is cured due to deterioration with time and temperature change, and the sealing performance is deteriorated. Therefore, the inspection process must surely eliminate defective products. Therefore, the inspection apparatus used in the inspection process is very important, and it must be able to discriminate between good and defective products with high accuracy.

  In view of this, an inspection apparatus has been developed that acquires an image while rotating a cylindrical inspection object and determines the presence or absence of scratches on the surface of the inspection object based on the image (see Patent Document 1). In such an inspection apparatus, a line sensor is used to acquire an image. Therefore, it is necessary to synchronize the rotation of the inspection object and the image acquisition by the line sensor.

Japanese Patent Application Laid-Open No. 2000-121569

As described above, for example, in a pipe assembly in which an O-ring is used for the connector portion, if the O-ring contact portion is damaged, the O-ring deteriorates and the sealing performance is deteriorated.
Accordingly, an object of the present invention is to provide a surface inspection apparatus and a surface inspection method capable of accurately detecting, for example, a scratch on the surface of an inspection object such as a pipe ASSY and accurately determining the quality of the inspection object based on the detection result. Is to provide.

  According to the first or fourth aspect of the present invention, in order to solve the above-described problem, the surface inspection apparatus (1) according to the present invention performs an inspection in which an area including an inspection area of an inspection object is imaged. A two-dimensional detector (24) for acquiring an image, a flaw detection unit (33) for detecting information on a flaw of the inspection object based on the inspection image, and determining the quality of the inspection object based on the information A pass / fail judgment unit (34).

  In addition, according to the second or sixth aspect, the scratch detection unit (33) extracts the feature amount related to the scratch on the inspection target based on the difference image between the inspection image and the unsharp image. By including the part (44) and the determination part (45) for determining the presence or absence of a flaw based on the feature amount, even a thin and shallow flaw can be accurately detected.

  In addition, as described in claim 3 or 7, it is preferable that the flaw detection unit (33) further includes a region extraction unit (41) for extracting an inspection region from the inspection image. With such a configuration, even when the position of the inspection object varies at the time of acquiring the inspection image, it is possible to detect the scratch after excluding the area unnecessary for the detection of the scratch, so that it is possible to detect the scratch more accurately. .

  Further, the two-dimensional detector is preferably arranged so as not to receive the illumination light emitted from the illumination light source and specularly reflected in the inspection region. With such a configuration, it is possible to increase the difference between the brightness value of the pixel corresponding to the scratch on the inspection image and the brightness value of the pixel corresponding to other than the scratch.

  According to claim 8, in order to solve the above-mentioned problem, the surface inspection method according to the present invention includes a first 2 image obtained by photographing a region including the inspection region of the inspection object with a two-dimensional detector. A step of acquiring a two-dimensional image (S201), a step of detecting information related to a scratch on the inspection object based on the first two-dimensional image (S203), and determining the quality of the inspection object based on the information. Step (S204).

  According to the ninth aspect of the present invention, the step (S203) of detecting information related to the flaw includes a step (S102) of generating an unsharp image of the inspection area based on the first two-dimensional image, Generating a difference image between the two-dimensional image and the unsharp image (S103), extracting a feature amount related to the scratch on the inspection object based on the difference image (S104), and scratching based on the feature amount It is preferable to include the step of determining the presence or absence (S105).

  Furthermore, as described in claim 10, in the surface inspection method according to the present invention, a part of the inspection object included in the inspection region at the time of acquiring the first two-dimensional image is included in a different position of the inspection region. A step (S206) of moving the inspection object so as to be acquired, a step (S201) of acquiring a second two-dimensional image obtained by photographing a region including the inspection region with a two-dimensional detector, and a second two-dimensional image. It is preferable to include a step (S203) of detecting information related to the flaw of the inspection object based on and a step (S204) of determining pass / fail of the inspection object based on the information related to the flaw. When sequentially imaging different areas of the inspection object while moving the inspection object, there is some variation in the amount of movement of the inspection object between each image by overlapping a part of the imaging area. Even in this case, it is possible to reliably photograph the portion for inspecting the inspection object.

  According to the eleventh aspect of the present invention, there is provided an injector inspection method for determining the presence or absence of a flaw on the surface of a pipe constituting the injector using the surface inspection method according to the present invention.

  In addition, the code | symbol in the parenthesis attached | subjected to each said means is an example which shows a corresponding relationship with the specific means as described in embodiment mentioned later.

Hereinafter, an inspection apparatus to which the present invention is applied will be described in detail with reference to the drawings.
In this embodiment, the inspection object is a pipe assembly used for the injector, and whether or not the surface of the O-ring contact portion for preventing fuel leakage provided in the connector portion is flawed or not is a good product. To be judged. The O-ring contact portion of the pipe ASSY has a cylindrical shape and has a smooth surface enough to specularly reflect light incident on the surface.

FIG. 1 shows a configuration block diagram of an inspection apparatus 1 to which the present invention is applied.
An inspection apparatus 1 to which the present invention is applied includes an imaging apparatus 2 and a processing apparatus 3. And in the imaging device 2, the imaging | photography area | region 12 is set to a part of surface of the O-ring contact part 11 of pipe ASSY10 which is a test object. Then, the imaging device 2 rotates the pipe ASSY 10 by a predetermined angle around the rotation axis α so that different portions of the O-ring contact portion 11 are included in the imaging region 12, and each time the imaging region 12 is detected two-dimensionally. An image photographed by the instrument (hereinafter referred to as an inspection image) is acquired and transmitted to the processing device 3. On the other hand, the processing device 3 detects information about a flaw such as the presence or absence of a flaw on the surface of the O-ring contact portion 11 based on each received inspection image, and determines whether the pipe ASSY 10 is a non-defective product based on the information.
The inspection apparatus to which the present invention is applied can acquire the inspection image by using a two-dimensional detector, thereby checking the presence / absence of scratches on the inspection object at high speed and providing an inexpensive configuration. In addition, the inspection device creates an unsharp image from the inspection image and determines the presence or absence of scratches based on the difference image between the inspection image and the unsharp image, so that fine and shallow scratches can be detected with high accuracy. It is an apparatus capable of making a determination.

Hereinafter, each part will be described in detail.
FIG. 2 is a schematic plan view of the imaging device 2. In FIGS. 1 and 2, portions unnecessary for the description are omitted for the sake of clarity. As shown in FIG.1 and FIG.2, in the imaging device 2, the holding | maintenance part 21 hold | maintains the pipe ASSY10 which is a test object. For this purpose, the holding unit 21 has a chuck for fixing the pipe ASSY 10 and a stage on which the pipe ASSY 10 is placed. The stage can rotate about the center of the stage as the rotation axis α, and the chuck fixes the pipe ASSY 10 so that the rotation axis α of the stage and the center axis of the pipe ASSY 10 substantially coincide. The stage is connected to the drive unit 22 through a gear. The stage is rotated by the power transmitted from the drive unit 22 through the gears. When the stage rotates, the pipe ASSY 10 held by the holding unit 21 also rotates about the center thereof as a rotation axis. Note that the transmission of power from the drive unit 22 to the stage of the holding unit 21 can be performed by a pulley and a belt.

  The drive unit 22 includes a stepping motor or a gear attached to a rotation shaft of the stepping motor. Then, the stepping motor of the drive unit 22 rotates the stage (that is, the pipe ASSY 10) by a predetermined angle by rotating a predetermined number of rotation steps at a time based on a control signal from the processing device 3. Further, the drive unit 22 may use a servo motor instead of the stepping motor.

  On the other hand, in the imaging device 2, when there is a scratch on the O-ring contact portion 11 of the pipe assembly 10 in the imaging region 12, the illumination light emitted from the illumination light source 23 irregularly reflected by the scratch is acquired by the detector 24. Each part is arranged. Therefore, as shown in FIG. 2, the illumination light source 23 has rod-shaped light sources 23 a and 23 b configured by LED arrays. Each of the rod-like light sources 23a and 23b is disposed substantially parallel to the central axis of the pipe assembly 10 and symmetric with respect to the normal n at the center of the imaging region 12 in order to uniformly illuminate the imaging region 12. . Each of the rod-like light sources 23a and 23b forms an angle of 60 ° with the normal line n of the imaging region 12, and is disposed at a distance of 28 mm from the imaging region 12. Further, for each rod-like light source 23a, 23b, when the surface of the O-ring contact portion 11 of the pipe ASSY 10 is scratched in the imaging region 12, a short wavelength light source is used so that the scratch easily diffuses the illumination light. Is done. In the present embodiment, each rod-shaped light source 23a, 23b is configured as a line array of blue LEDs (wavelength 470 nm) as an example.

  The detector 24 is arranged so as not to receive the illumination light that is specularly reflected at the substantially central portion of the imaging region 12. In the present embodiment, the detector 24 is arranged at a position 110 mm away from the imaging region 12 on the normal line n of the imaging region 12. The detector 24 includes a two-dimensional detector composed of a photoelectric converter such as a CCD or C-MOS sensor, and an imaging optical system that forms an image of the imaging region 12 on the two-dimensional detector. Of course, the resolution of the two-dimensional detector, the imaging magnification of the imaging optical system, and the like are set so that the scratch to be detected can be resolved on the image. In this embodiment, a 1/3 inch CCD having 652 × 492 pixels is used as the two-dimensional detector. Further, as the imaging optical system, a camera photographing lens having a focal length of 50 mm and a close-up ring having a cylinder length of 25 mm were used. Further, in the present embodiment, since the detector 24 is a two-dimensional detector, an image having a certain width in the circumferential direction can be acquired by one imaging. Therefore, when photographing the entire periphery of the O-ring contact portion 11 of the pipe ASSY 10, it is not necessary to synchronize the rotation of the pipe ASSY 10 and the photographing. In addition, in two inspection images obtained by photographing the adjacent sections of the O-ring contact portion 11, a part of the adjacent section is overlapped so that the same portion of the O-ring contact section 11 is captured, so that the pipe ASSY 10 is taken between the respective photographings. There may be some variation in the angle of rotation of the shaft, or some deviation between the rotation axis of the stage and the rotation axis of the pipe ASSY 10. Therefore, the adjustment of the imaging device 2 is easy. Further, in the case of using a one-dimensional detector, if such a variation or a rotational axis shift occurs, an inspection image that accurately represents the O-ring contact portion 11 cannot be obtained, and the quality determination accuracy decreases. . However, in the present embodiment, even if the above-described variation or rotation axis deviation occurs, each inspection image itself can accurately represent the O-ring contact portion 11 included in the imaging region 12, and the pass / fail judgment accuracy decreases. do not do.

  FIG. 3 shows an example of the inspection image 300 of the imaging region 12 acquired by the detector 24. The inspection image 300 is a multi-value gray image with 256 gradations (0-255), and is expressed brighter as the pixel value is larger. In the inspection image 300, an area indicated by a solid line corresponds to the imaging area 12. As described above, the imaging region 12 may be smaller than the region actually acquired by the detector 24. Therefore, the detector 24 acquires only the image corresponding to the imaging region 12 and transfers it to the processing device 3 in order to speed up the image transfer to the processing device 3 and the processing in the processing device 3. Also good. It should be noted that the image acquisition area can be limited by designating an area for acquiring an image to the detector 24 in advance.

As another embodiment of the imaging device 2, the drive unit 22 can be configured by a continuously rotating motor instead of the stepping motor. In this case, the imaging device 2 can acquire an inspection image of the entire periphery of the O-ring contact portion 11 by repeating imaging at a constant time interval. However, in this case, it is necessary to provide the imaging device 2 or the processing device 3 with a mechanism for detecting that the pipe ASSY 10 has rotated 360 °. As an example of such a mechanism, for example, it is possible to provide a timer that measures the time required for the pipe ASSY 10 to rotate 360 ° from the start of acquisition of the inspection image. The time required for the pipe assembly 10 to rotate 360 ° can be measured in advance.
Furthermore, the illumination light source 23 may be only one light source. However, when the illumination light source 23 is composed of a single light source, it is preferable to perform unevenness correction as preprocessing for the inspection image. Moreover, you may comprise the illumination light source 23 with another light source which has sufficient brightness | luminance, such as white LED.
Note that the imaging device 2 is provided with a housing (not shown) made of an opaque member in order to prevent light from other than the illumination light source 23 from being applied to the imaging region 12, and each part is arranged in the housing. You may comprise.

  Based on each inspection image acquired from the imaging device 2, the processing device 3 detects information about a flaw such as the presence or absence of a flaw on the surface of the O-ring contact portion 11 of the pipe ASSY 10, and the pipe ASSY 10 based on the information. Whether or not is a good product is determined. For this purpose, the processing device 3 includes a communication unit 30, a storage unit 31, a control unit 32, a flaw detection unit 33, and a pass / fail determination unit 34, and includes, for example, a personal computer (PC) and its peripheral devices.

  Here, the communication unit 30 is an input / output interface that transmits and receives control signals and images between the processing device 3 and the imaging device 2, and various I / O ports such as USB, SCSI, RS232C, and Ethernet (registered trademark). And their drivers. Then, the processing device 3 receives the inspection image from the imaging device 2 through the communication unit 30. On the other hand, the control signal generated by the control unit 32 is transmitted to the drive unit 22, the illumination light source 23, or the detector 24 of the imaging device 2 through the communication unit 30. Further, the communication unit 30 outputs the pass / fail judgment result of the pipe ASSY 10 to an external device.

  The processing device 3 temporarily stores the inspection image received from the imaging device 2 in a storage unit 31 including a random access memory (RAM) or a nonvolatile memory such as a magnetic disk, an optical disk, or a flash memory. On the other hand, the control unit 32, which includes a semiconductor memory such as a central processing unit (CPU), a read only memory (ROM), and a random access memory (RAM) of a PC and operates according to a program read into the CPU, is a processing device. 3 and each part of the imaging device 2 are controlled. The flaw detection unit 33 examines information related to flaws such as the presence or absence of flaws on the surface of the O-ring contact portion 11 of the pipe ASSY 10 and the length of the flaws based on the inspection image. And the quality determination part 34 determines whether the pipe ASSY10 which is a test object is a quality product based on the information regarding the damage | wound obtained by the damage | wound detection part 33. FIG. Note that the flaw detection unit 33 and the pass / fail determination unit 34 are implemented as program modules executed on the CPU, for example. Alternatively, the flaw detection unit 33 and the quality determination unit 34 may be mounted as an image processing board including an image processing processor provided separately from the CPU.

  Here, with reference to FIG. 3 again, the characteristics of the inspection image 300 acquired by the imaging device 2 and used for the processing of the wound detection unit 33 will be described. As shown in FIG. 3, the dark area at the center of the imaging area 12 is an inspection area 310 to be inspected for scratches attached to the O-ring contact portion 11 of the pipe ASSY 10. In the inspection area 310, the pixel value is relatively small except for the area corresponding to the scratch 320 formed on the surface. This is because, as described above, the illumination light source 23 and the detector 24 are arranged so that the illumination light emitted from the illumination light source 23 and specularly reflected at the substantially central portion of the imaging region 12 does not enter the detector 24. is there. On the other hand, the flaw 320 diffusely reflects the illumination light emitted from the illumination light source 23 and diffuses the illumination light over a wide range, so that the flaw 320 is detected as a bright region on the inspection image 300. In addition, since the pipe assembly 10 has a cylindrical shape, the specularly reflected illumination light enters the detector 24 in the vicinity of the left end or the right end of the imaging region 12. Therefore, the left end portion 330 and the right end portion 340 of the imaging region 12 on the inspection image 300 are bright regions.

Further, the O-ring contact portion 11 of the pipe assembly 10 acquires an inspection image for the imaging region 12 because the rotation axis of the stage and the rotation axis of the pipe assembly 10 are shifted when the chuck of the holding unit 21 holds the pipe assembly 10. It may be slightly different each time. Therefore, the imaging region 12 is set wider than the width of the O-ring contact portion 11 in the left-right direction so that the O-ring contact portion 11 is necessarily included.
Further, when the scratch 320 is shallow and thin, in the image, the difference between the pixel value corresponding to the scratch 320 and the pixel value included in the other part of the inspection region 310 is small, and the difference is the signal in the inspection region 310. It may be the same as unevenness. Therefore, even if simple binarization processing is performed on the inspection region 310, it is difficult to accurately detect the scratch 320.
The flaw detection unit 33 will be described in detail with reference to FIGS. 4 and 5. FIG. 4 is a functional block diagram of the flaw detection unit 33. As shown in FIG. 4, the scratch detection unit 33 includes a region extraction unit 41, an unsharp image generation unit 42, a difference image generation unit 43, a feature amount extraction unit 44, and a determination unit 45. FIG. 5 is a flowchart of an operation for examining the presence or absence of a flaw in the flaw detection unit 33.

As described above, there are bright areas (areas 330 and 340 shown in FIG. 3) that are not desirable for detecting flaws in the imaging area 12. Therefore, first, the region extraction unit 41 of the scratch detection unit 33 extracts an inspection region corresponding to the O-ring contact unit 11 from the inspection image (step S101).
Next, in order to obtain a difference image in which the difference between the pixel value corresponding to the scratch and the pixel value corresponding to another region is emphasized, the unsharp image generation unit 42 of the scratch detection unit 33 is based on the inspection image. Thus, an unsharp image with a blurred inspection area is generated (step S102). And the difference signal generation part 43 of the damage | wound detection part 33 performs the difference of an inspection image and an unsharp image, and produces | generates a difference image (step S103). When the difference image is generated, the feature amount extraction unit 44 of the scratch detection unit 33 extracts feature amounts such as the length and area of the region corresponding to the scratch (step S104). Finally, the determination unit 45 of the scratch detection unit 33 compares the extracted feature amount with a predetermined determination criterion to determine whether there is a scratch (step S105). And the information regarding a wound, such as the presence or absence of a damage | wound, length, is output (step S106). Hereinafter, each part of the damage | wound detection part 33 is demonstrated in detail.

  The region extraction unit 41 extracts the inspection region 310 corresponding to the O-ring contact unit 11 from the inspection image 300. Therefore, as an example, the area extraction unit 41 generates a binarized image obtained by dividing the inspection image 300 into an inspection area 310 and other areas. Therefore, the region extraction unit 41 calculates an average pixel value of the entire inspection image 300 and sets the average pixel value as a binarization threshold value. The region extraction unit 41 sequentially sets each pixel of the inspection image 300 as a target pixel, and compares the value of the target pixel with the calculated binarization threshold value. If the pixel value of the target pixel of the inspection image 300 is equal to or greater than the binarization threshold, the corresponding pixel value of the binarized image is set to “1”, and if the pixel value is less than the binarization threshold, the binarized image is handled. The pixel value to be set is set to “0”. The inspection area 310 corresponds to the area of the pixel value “0” of the binarized image. Further, the region extraction unit 41 prevents the relatively bright pixels in the inspection region 310 from being included in the inspection region 310, such as the pixels corresponding to the scratch 320, and the pixel value “0” of the binarized image. For these pixels, a morphological calculation closing process may be performed. The obtained binarized image is stored in a memory constituting the processing device 3 so that it can be used by the unsharp image generation unit 42 or the like.

  Another example of extracting the inspection area 310 will be described below. When the pipe assembly 10 that is the inspection object is one type, the size of the inspection region 310 corresponding to the O-ring contact portion 11 in the inspection image 300 is also substantially constant. Therefore, the area extraction unit 41 sets a filter area having approximately the same size as the assumed inspection area 310 so that the left end of the filter area coincides with the left end of the imaging area 12, and the average pixel value in the filter area Is calculated. Thereafter, the region extraction unit 41 sets the filter region by shifting it to the right by one pixel, and calculates the average pixel value in the filter region again. Thereafter, similarly, the region extraction unit 41 repeats the calculation of the average pixel value in the filter region while shifting the filter region to the right by one pixel until the right end of the filter region matches the right end of the imaging region 12. Then, the filter area when the average pixel value is minimized is set as the inspection area 310.

すなわち、アンシャープ画像生成部４２は、検査画像３００の画素Ｉ_ｐ（ｘ_ｋ，ｙ_ｋ）を中心とした横（２ｍ＋１）×縦（２ｎ＋１）画素（ただし、ｍ、ｎは０以上の整数）の領域（以下、マスク領域という）に含まれる画素の平均値を算出し、その平均値をアンシャープ画像における対応画素Ｕ_ｐ（ｘ_ｋ，ｙ_ｋ）の値とする。ただし、アンシャープ画像では、後述する差分画像において傷３２０に対応する画素の値が大きくなるように、傷３２０に対応する画素の値がその周囲の画素の値とともに平均化されることが好ましい。そこで、マスク領域は、想定される傷に対して、傷に対応する画素以外の画素も含むように設定する。 The unsharp image generation unit 42 generates an unsharp image from the inspection image 300. For this purpose, the unsharp image generation unit 42 includes arbitrary pixels I _p (x _k , y _k ) included in the inspection area 310 of the inspection image 300 ( _where x _k represents horizontal coordinates and y _k represents vertical coordinates). , The value of the corresponding pixel U _p (x _k , y _k ) in the unsharp image is calculated based on the following equation.

That is, the unsharp image generation unit 42 has horizontal (2m + 1) × vertical (2n + 1) pixels centered on the pixel I _p (x _k , y _k ) of the inspection image 300 (where m and n are integers of 0 or more). The average value of the pixels included in the area (hereinafter referred to as the mask area) is calculated, and the average value is set as the value of the corresponding pixel _Up (x _k , y _k ) in the unsharp image. However, in the unsharp image, it is preferable that the value of the pixel corresponding to the flaw 320 is averaged together with the values of the surrounding pixels so that the value of the pixel corresponding to the flaw 320 is increased in the differential image described later. Therefore, the mask region is set so as to include pixels other than the pixels corresponding to the scratches with respect to the assumed scratches.

  Here, as shown in FIG. 3, when the assumed scratch occurs along the axial direction of the pipe ASSY 10, that is, when the assumed scratch occurs in the substantially vertical direction in the inspection image 300, for example, m = 10, n = 1 can be set. However, when a pixel that is not included in the inspection region 310 exists in the mask region, the pixel is excluded from the calculation of the average value. Whether or not the pixel of interest is included in the inspection region 310 can be determined by the value of the corresponding pixel in the binarized image. When the value of the corresponding pixel is “0”, the pixel of interest Is included in the inspection area 310. An unsharp image is generated by performing the above processing on each pixel included in the inspection area 310. The generated unsharp image is an image in which a structure smaller than the size of the mask area is blurred.

The shape of the mask area used is not limited to the above. For example, in the area corresponding to the inspection area 310 of the unsharp image, the vertical length of the mask area is twice the vertical length of the inspection area 310 so that all the pixels in the same column have the same value. You may set as follows. Moreover, when the assumed damage | wound arises in the circumferential direction of pipe ASSY10 (namely, when it arises in the horizontal direction with respect to the test | inspection image 300), you may set to m = 1 and n = 10, for example. Further, the mask region may have a two-dimensional shape. For example, m = 5 and n = 5 may be set. However, in the unsharp image, it is preferable that the value of the pixel corresponding to the flaw 320 is averaged with the values of the surrounding pixels so that the value of the pixel corresponding to the flaw 320 is increased in the differential image described later. Therefore, the mask region is set so as to include pixels other than the pixels corresponding to the scratches with respect to the assumed scratches.
The obtained unsharp image is stored in a memory constituting the processing device 3 so that it can be used by the difference image generation unit 43 or the like.

アンシャープ画像では、傷３２０に対応する画素の値がその周辺の画素の値とともに平均化されているため、差分画像では、例えば照明のムラなどに起因する全体的な画素値の変化はキャンセルされ、傷３２０のように局所的に画素値が変化する部分だけが相対的に大きな画素値となる。そのため、傷３２０に対応する画素を他の画素と容易に識別することができる。
得られた差分画像も、特徴量抽出部４４などで利用できるように、処理装置３を構成するメモリに記憶される。 The difference image generation unit 43 generates a difference image from the inspection image 300 and the unsharp image. For this purpose, the difference image generation unit 43 calculates the value and the corresponding pixel U _p (x _k ,) of the unsharp image for any pixel I _p (x _k , y _k ) included in the inspection region 310 of the inspection image 300. The value of the corresponding pixel S _p (x _k , y _k ) in the difference image is calculated using the value of y _k ) and the following expression.

In the unsharp image, the value of the pixel corresponding to the scratch 320 is averaged together with the values of the surrounding pixels. Therefore, in the difference image, changes in the overall pixel value due to, for example, uneven illumination are canceled. Only the part where the pixel value locally changes like the scratch 320 has a relatively large pixel value. Therefore, the pixel corresponding to the scratch 320 can be easily distinguished from other pixels.
The obtained difference image is also stored in the memory constituting the processing device 3 so that it can be used by the feature amount extraction unit 44 or the like.

  The feature amount extraction unit 44 extracts a feature amount due to the scratch 320 based on the difference image. One example will be described below. The feature amount extraction unit 44 first binarizes the difference image. The feature amount extraction unit 44 calculates an average pixel value of the entire difference image S, and sets a value obtained by adding a predetermined offset value to the average pixel value as a binarization threshold value. The offset value is set to, for example, 10 so that a part of the inspection area 310 is not mistakenly extracted as the damage 320 when the inspection area 310 does not have a scratch of the O-ring contact portion 11. Then, the feature amount extraction unit 44 sequentially sets each pixel of the difference image as a target pixel, and compares the value of the target pixel with the calculated binarization threshold value. If the value of the target pixel of the difference image is equal to or greater than the binarization threshold, the value of the target pixel is set to “1”, and if the value is less than the binarization threshold, the value of the target pixel is set to “0”. . Since the pixel corresponding to the flaw 320 has a relatively high value, there is a high possibility of having a pixel value “1” in the binarized difference image. On the other hand, in the difference image, pixels other than the scratch 320 have a relatively low value, and thus there is a high possibility that the binarized difference image has the pixel value ‘0’. That is, in the binarized difference image, the presence or absence of the flaw 320 can be examined based on the distribution of the pixels having the pixel value “1”.

Therefore, the feature amount extraction unit 44 obtains the longest run length for the pixel having the pixel value “1” in the binarized difference image, and sets the value as the first feature amount. Further, the feature quantity extraction unit 44 obtains the number of pixels having the pixel value “1” (corresponding to the area occupied by the pixels having the pixel value “1”) in the binarized difference image, and sets it as the second feature quantity.
Note that the feature amount extraction unit 44 performs the morphological calculation closing process on the pixel having the pixel value “1” of the binarized difference image in order to obtain the feature amount more accurately, and then performs the above run. The length, area, etc. may be obtained. Furthermore, when there is little possibility of the presence of scratches, it is preferable to speed up the processing by omitting useless calculations. Therefore, the feature quantity extraction unit 44 checks the number of pixels having the pixel value “1” in the binarized difference image. If the number of pixels is equal to or less than the predetermined number, the closing process is not performed, and the first feature quantity is also obtained. You may make it output as '0'.

Further, the extracted feature amount is not limited to these. For example, the position of the top and bottom pixels of the set of pixels having the pixel value “1” having the longest run length, and the average of the pixels of the difference image corresponding to the pixels having the pixel value “1” in the binarized difference image A value (when obtaining this average value, a binarized difference image needs to be created separately from the original difference image) or the like may be extracted as a feature amount.
Each extracted feature amount is sent to the determination unit 45.

The determination unit 45 determines whether or not there is a scratch on the surface of the O-ring contact unit 11 by comparing the acquired feature amount with a predetermined determination criterion. The predetermined determination standard can be determined based on the standard of the pipe assembly 10 that is the inspection object. For example, when a scratch having a length of 1 mm and an average width of 0.2 mm or more is detected in the O-ring contact portion 11, the first determination criterion for the first feature amount (the pixel run length of the pixel value “1”) Can be set to the number of pixels corresponding to a length of 1 mm. Similarly, the second criterion for the second feature amount (the area occupied by the pixel having the pixel value “1”) can be set to the number of pixels corresponding to 1 mm in length × 0.2 mm in width. .
The determination unit 45 compares the first feature amount and the second feature amount with the first determination criterion and the second determination criterion, respectively. If any one of these feature quantities exceeds the determination criterion, the determination unit 45 determines that the O-ring contact portion 11 of the pipe ASSY 10 is damaged. If the determination unit 45 determines that there is a flaw, the determination unit 45 sends information about the flaw such as a feature amount such as the length and area of the flaw and information indicating that there is a flaw to the pass / fail determination unit 34. On the other hand, when the determination unit 45 determines that there is no flaw, the determination unit 45 sends information indicating that there is no flaw to the pass / fail determination unit 34.

  The quality determination unit 34 determines whether or not the pipe ASSY 10 that is the inspection object is a non-defective product based on the information about the scratch acquired from the scratch detection unit 33. Therefore, the pass / fail determination unit 34 checks whether or not the information about the scratch acquired from the scratch detection unit 33 satisfies a predetermined pass / fail determination criterion. And when the quality determination criteria are satisfy | filled, the quality determination part 34 determines the pipe ASSY10 as inferior goods. The pass / fail judgment criteria can be determined based on the standard of the pipe assembly 10 that is the inspection object, and may be determined based on the number of detected scratches, or based on a combination of the number of scratches and a feature amount related to the scratches. May be determined. In the present embodiment, as the simplest example, the pass / fail judgment criterion is that there are one or more scratches. That is, if it is determined that any inspection image 300 has a scratch, the pass / fail determination unit 34 determines that the pipe ASSY 10 is defective.

  In another example, the pass / fail criterion may be one or more scratches having a length of 2 mm or more, or two or more scratches regardless of the length of the scratches. When the pass / fail criterion is set to a value of 2 or more with respect to the number of scratches, the pass / fail determination unit 34 adds the number of inspection images determined to have scratches to calculate the cumulative number of scratches, The cumulative total is used for pass / fail judgment. Thus, the pass / fail criterion can be suitably set according to the specification of the inspection object.

  The operation | movement about the test | inspection implemented with the test | inspection apparatus 1 to which this invention is applied is demonstrated using the flowchart shown in FIG. In the inspection carried out by the inspection apparatus 1 to which the present invention is applied, the pipe assembly 10 as an inspection object is rotated by a predetermined angle until it makes one turn, and an inspection image is acquired each time it is rotated to detect the presence or absence of scratches. . If a scratch is found, the pipe assembly 10 is determined as a defective product.

  In the inspection apparatus 1, the pipe ASSY 10 as an inspection object is held in the holding unit 21 of the imaging apparatus 2, and predetermined setting information such as a threshold parameter is received from the storage unit 32 of the processing apparatus 3 in the control unit 31 of the processing apparatus 3. When it is read and the imaging device 2 and the processing device 3 are initialized, the inspection is started. When the inspection is started, the imaging device 2 acquires an inspection image of the imaging region 12 set in a part of the O-ring contact portion 11 of the pipe ASSY 10 (Step S201). The control unit 31 stores the acquired inspection image in the storage unit 32 so that it can be used by the scratch detection unit 33 of the processing device 3 (step S202).

  Next, the flaw detection unit 33 reads an inspection image from the storage unit 32 and detects information about the flaw such as the presence / absence of a flaw on the surface of the O-ring contact portion 11 of the pipe ASSY 10 based on the image (step). S203). Note that the details of the processing for detecting information related to scratches are as shown in FIG. Next, the pass / fail determination unit 34 determines whether or not the pipe ASSY 10 is a non-defective product based on the information regarding the scratch (step S204). If the pass / fail determination unit 34 determines that the pipe ASSY 10 is defective, the pipe ASSY 10 outputs a signal indicating that the pipe ASSY 10 is defective through the communication unit 30 (step S205), and ends the inspection.

  In step S204, when the pass / fail determination unit 34 does not determine that the pipe assembly 10 is defective, the control unit 31 determines that the pipe assembly 10 fixed to the holding unit 21 is at a predetermined angle (with respect to the drive unit 22 of the imaging device 2). For example, a command to rotate by a predetermined number of rotation steps is sent so as to rotate. The stepping motor of the drive unit 22 receives the command and rotates by a predetermined number of rotation steps (step S206). Note that the above-mentioned predetermined angle is the O-ring contact in the two inspection images obtained by photographing the adjacent sections of the O-ring contact portion 11 so that the angle at which the pipe assembly 10 actually rotates may vary. It is set so that a part of the adjacent section overlaps so that the same part of the part 11 can be seen.

  The control unit 31 determines whether or not the stepping motor of the drive unit 22 has finished rotating, that is, whether or not the pipe ASSY 10 has rotated by a predetermined angle (step S207). If the rotation of the stepping motor has not ended, the control is returned to before step S207, and after the predetermined time has elapsed, the determination at step S207 is performed again. On the other hand, when it is determined in step S207 that the rotation of the stepping motor has been completed, the control unit 31 determines whether or not the number of rotation steps of the stepping motor has reached a predetermined number indicating that the pipe ASSY 10 has rotated 360 °. Is determined (step S208). When the control unit 31 determines that the number of rotation steps of the stepping motor has not reached the predetermined number, the control unit 31 returns the control to before step S201, and repeats the processing after step S201 again.

On the other hand, when the control unit 31 determines that the number of rotation steps of the stepping motor has reached a predetermined number, the control unit 31 outputs a signal indicating that the pipe ASSY 10 is a non-defective product through the communication unit 30 (step S209), and ends the inspection. .
Note that the inspection apparatus 1 may be configured such that the steps (steps S203 and S204) related to the detection of the scratch and the pass / fail judgment described above and the steps (steps S206 and S207) related to the rotation of the pipe ASSY 10 can be executed in parallel. Good.

  As described above, the inspection apparatus 1 to which the present invention is applied does not synchronize the rotation of the inspection object and the acquisition of the image by using the detector 24 of the imaging apparatus 2 as a two-dimensional detector. Because it is good, it does not require complex adjustments and does not require a mechanism for synchronizing it. Furthermore, since the inspection apparatus 1 detects a flaw based on a difference image between the inspection image and an unsharp image obtained by blurring the inspection image, it can accurately detect even a thin and shallow flaw. Therefore, the inspection apparatus 1 can determine the quality of the pipe assembly 10 with high accuracy.

  In addition, embodiment mentioned above is for demonstrating this invention, and this invention is not limited to these embodiment. The present invention can also be applied to an inspection apparatus for the surface of, for example, a flat inspection object other than a cylindrical object such as a pipe assembly. When the present invention is applied to such an inspection object inspection apparatus, it is preferable that the holding unit in the above embodiment includes a stage movable in a horizontal plane, such as an XY stage.

  In the above embodiment, the detector is disposed so as not to receive the illumination light emitted from the illumination light source and mirror-reflected in the inspection area, but conversely, the illumination light specularly reflected in the inspection area is received. A detector may be arranged as described above. In this case, since the part with a flaw becomes darker than the other part on the inspection image, it is necessary to handle light and dark in the above flaw detection process.

Further, in the operation procedure of the inspection apparatus described above, after the inspection apparatus acquires all the inspection images and performs scratch detection on each inspection image, instead of performing pass / fail determination each time an inspection image is obtained. The pass / fail judgment may be performed. In this case, the number of detected scratches may be output together with the pipe ASSY quality determination result.
As described above, various modifications can be made within the scope of the present invention according to the embodiment to be implemented.

It is a block diagram of the configuration of an inspection apparatus to which the present invention is applied. It is a plane schematic diagram of an imaging device. It is an example of the image acquired with an imaging device. It is a functional block diagram of a flaw detection part. It is a flowchart which shows the operation | movement procedure of a flaw detection process. It is a flowchart which shows the operation | movement procedure of the test | inspection apparatus to which this invention is applied.

Explanation of symbols

1 Inspection device 10 Pipe ASSY
DESCRIPTION OF SYMBOLS 11 O ring contact part 12 Imaging region 2 Imaging device 21 Holding part 22 Drive part 23, 23a, 23b Illumination light source 24 Detector 3 Processing apparatus 30 Communication part 31 Control part 32 Memory | storage part 33 Scratch detection part 34 Quality determination part 41 Area extraction Unit 42 unsharp image generation unit 43 difference image generation unit 44 feature quantity extraction unit 45 determination unit 300 inspection image 310 inspection region 320 scratch 330, 340 light region

Claims (11)

A two-dimensional detector (24) for acquiring an inspection image obtained by imaging an area including the inspection area of the inspection object;
A wound detection unit (33) for detecting information on the wound of the inspection object based on the inspection image;
A pass / fail judgment unit (34) for judging pass / fail of the inspection object based on the information;
A surface inspection apparatus characterized by comprising: The flaw detection unit (33)
An unsharp image generator (42) for generating an unsharp image of the inspection region based on the inspection image;
A difference image generation unit (43) for generating a difference image between the inspection image and the unsharp image;
A feature amount extraction unit (44) that extracts a feature amount related to a scratch on the inspection object based on the difference image;
A determination unit (45) for determining the presence or absence of scratches based on the feature amount,
The surface inspection apparatus according to claim 1, wherein the feature amount or the presence / absence of a scratch is output as the information. The flaw detection unit (33)
An area extracting unit (41) for extracting the inspection area from the inspection image;
The surface inspection apparatus according to claim 2. A surface inspection apparatus (1) for determining the quality of an inspection object by detecting scratches on the surface of the inspection object,
An illumination light source (23) for illuminating an inspection area of the inspection object;
A two-dimensional detector (24) for obtaining an inspection image obtained by imaging an area including the inspection area;
A holding unit (21) that holds the inspection object and is movable to include different parts of the inspection object in the inspection region;
A drive unit (22) for driving the holding unit;
An imaging device (2) having:
A communication unit (30) for receiving the inspection image from the imaging device (2);
A wound detection unit (33) for detecting information on the wound of the inspection object based on the inspection image;
A pass / fail judgment unit (34) for judging pass / fail of the inspection object based on the information;
A processing device (3) having:
A surface inspection apparatus characterized by comprising: The two-dimensional detector (24)
The surface inspection apparatus according to claim 4, wherein the surface inspection apparatus is arranged so as not to receive the illumination light emitted from the illumination light source and specularly reflected in the inspection area. The flaw detection unit (33)
An unsharp image generator (42) for creating an unsharp image of the inspection area based on the inspection image;
A difference image generation unit (43) for generating a difference image between the inspection image and the unsharp image;
A feature amount extraction unit (44) that extracts a feature amount related to a scratch on the inspection object based on the difference image;
A determination unit (45) for determining the presence or absence of scratches based on the feature amount,
The surface inspection apparatus according to claim 4, wherein the feature amount or the presence / absence of a scratch is output as the information. The flaw detection unit (33)
An area extracting unit (41) for extracting the inspection area from the inspection image;
The surface inspection apparatus according to claim 6. A surface inspection method for determining the quality of an inspection object by detecting scratches on the surface of the inspection object,
Obtaining a first two-dimensional image obtained by photographing a region including the inspection region of the inspection object with a two-dimensional detector (S201);
Detecting information relating to the scratch on the inspection object based on the first two-dimensional image (S203);
A step of determining pass / fail of the inspection object based on the information about the scratch (S204);
A surface inspection method comprising: The step (S203) of detecting information about the scratch includes
Generating an unsharp image of the inspection region based on the first two-dimensional image (S102);
Generating a difference image between the first two-dimensional image and the unsharp image (S103);
Extracting a feature amount related to the scratch of the inspection object based on the difference image (S104);
Determining whether or not there is a scratch based on the feature amount (S105);
The surface inspection method according to claim 8 comprising: Furthermore, the step of moving the inspection object so that a part of the inspection object included in the inspection area is included in a different position of the inspection area when acquiring the first two-dimensional image (S206). When,
Obtaining a second two-dimensional image obtained by photographing a region including the inspection region with a two-dimensional detector (S201);
Detecting information related to the scratch on the inspection object based on the second two-dimensional image (S203);
A step of determining pass / fail of the inspection object based on the information (S204);
The surface inspection method according to claim 8 or 9.   An injector inspection method for determining the presence or absence of scratches on the surface of a pipe constituting the injector using the surface inspection method according to claim 8.

Images