JP2018146401A - Controller for inspection device, inspection device, method for controlling inspection device, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a controller for an inspection device which can accurately detect cracks generated in an inspection target while suppressing an increase in scan time.SOLUTION: A controller for an inspection device controls a body of the inspection device, the inspection device including: a sensor unit for acquiring image data and shape data from a surface of an inspection target; a moving mechanism for moving and rotating the sensor unit around the inspection target; and a control unit for controlling the sensor unit and the moving mechanism. The controller for the inspection device includes: a crack candidate operation processor; and a detection result output device, the crack candidate operation processor detecting candidates of cracks from the shape data and correcting the image data on the basis of the shape data, detecting the candidates of the cracks from the corrected image data, integrating the candidates of the cracks detected from the shape data and the candidates of the cracks detected from the image data with each other and outputting the integrated candidates as a result of inspection, and the detection result output device outputting the result of the inspection output from the crack candidate operation processor to a user.SELECTED DRAWING: Figure 1

Description

  Embodiments described herein relate generally to an inspection apparatus control apparatus, an inspection apparatus, an inspection apparatus control method, and a program.

  Non-destructive inspection techniques such as visual inspection and ultrasonic flaw inspection are used for quality confirmation by non-destructive inspection after use for inspection objects such as stationary blades of gas turbines. For example, ultrasonic flaw inspection and ECT flaw inspection are mainly used for the internal defects of a stationary blade, but defects such as damage on the surface such as cracks and thinning are recorded manually by visual inspection. Yes. For this reason, a long inspection time is required, including crack repair when a crack is generated, resulting in high inspection costs. Therefore, an automatic inspection that can improve the yield and shorten the construction period in the inspection process is desired.

  The main methods for detecting cracks include a method based on shape data comparison and a method based on image processing of image data. When using the method based on shape data comparison, it is necessary to take a wide range of images at a distance from the object to be inspected in order to shorten the inspection time. There is a risk of detection omission. Furthermore, in order to acquire shape data, the principle of triangulation is basically used, so that a blind spot may occur. For example, there is a problem that one side or an end of a crack outline is missing. On the other hand, when using a method by image processing of image data, a crack is detected by comparing the luminance characteristics of the standard state and the luminance state of the current state, but along with the shape change of the inspection object, There is a problem in that the luminance distribution changes due to factors other than the occurrence of cracks, resulting in crack detection failure.

  Furthermore, in the method using image data, the feature amount is calculated from the difference between the luminance distribution of the image data photographed in a normal state and the luminance distribution of the image data photographed of the inspection object, and the crack is based on the data. However, due to the influence of the change in the shape of the inspection object and the resulting change in lighting conditions (such as shadows), a feature value that is different from the feature value obtained when the normal state is obtained as a reference is calculated. There is also a problem that it cannot be detected.

  Patent Document 1 discloses a method of facilitating detection of defects by controlling illumination based on an acquired image, acquiring a plurality of image data, generating image data having a uniform gradation, and thereby detecting defects. However, it is difficult to say that the method of Patent Document 1 has sufficient accuracy for detecting a fine crack. Further, Patent Document 2 obtains a luminance distribution while moving an observation part two-dimensionally, calculates a base luminance distribution corresponding to a change component of the gentle luminance distribution from the obtained imaging luminance distribution, and calculates the luminance distribution from the luminance distribution. A method of using the corrected luminance distribution excluding the base luminance distribution for inspection is disclosed. However, in the method of Patent Document 2, since high-resolution shape data is required to detect a fine crack, it is necessary to reduce the scanning speed or to reduce the scanning interval. The time required for this will increase.

JP2013-140040A International Publication No. 2009/139394

  An object of this embodiment is to provide an inspection apparatus control device, an inspection apparatus, an inspection apparatus control method, and a program capable of accurately detecting a crack generated in an inspection object while suppressing an increase in scan time. That is.

  The inspection apparatus control device according to the present embodiment acquires image data and shape data from the surface of the inspection object, moves the sensor part around the inspection object, and rotates the sensor part. An inspection apparatus control device that controls an inspection apparatus main body including a mechanism, a control unit that controls the sensor unit and the moving mechanism, and detects crack candidates from the shape data , Correcting the image data based on the shape data, detecting crack candidates from the corrected image data, and detecting crack candidates detected from the shape data and crack candidates detected from the image data. A crack candidate calculation processing device that integrates and outputs the result as an inspection result, and a detection result output device that outputs the inspection result output from the crack candidate calculation processing device to a user. .

  The inspection apparatus according to the present embodiment acquires image data and shape data from the surface of an inspection object, a sensor unit, a movement mechanism that moves and rotates the sensor unit around the inspection object, From the control unit that controls the sensor unit and the moving mechanism, the crack candidate is detected from the shape data, the image data is corrected based on the shape data, and from the corrected image data, A crack candidate calculation processing device that detects a crack candidate, integrates the crack candidate detected from the shape data and the crack candidate detected from the image data, and outputs the result as an inspection result, and the crack candidate A detection result output device that outputs the inspection result output from the arithmetic processing device to a user.

  The inspection apparatus control method according to the present embodiment acquires image data and shape data from the surface of an inspection object, and moves and rotates the sensor unit and the sensor unit around the inspection object. And a control unit that controls the sensor unit and the moving mechanism, and an inspection device control method for controlling the inspection device main body unit, wherein the moving mechanism and the sensor unit are connected via the control unit. Controlling to acquire the image data and the shape data of the surface of the inspection object; detecting crack candidates from the shape data; and correcting the image data based on the shape data The crack candidate is detected from the corrected image data, and the crack candidate detected from the shape data and the crack candidate detected from the image data are integrated and output as an inspection result. Comprising that the step, and outputting the outputted the inspection results to the user, the.

  The program according to the present embodiment includes a sensor unit that acquires image data and shape data from the surface of an inspection object, a movement mechanism that moves and rotates the sensor unit around the inspection object, and A program that is read and executed by a computer that controls the inspection apparatus main body unit, the control unit controlling the sensor unit and the moving mechanism, and the program is provided to the computer. Via the step of acquiring the image data and the shape data of the surface of the inspection object by controlling the moving mechanism and the sensor unit, detecting crack candidates from the shape data, and the shape Correcting the image data based on the data, detecting a crack candidate from the corrected image data, and detecting the crack candidate detected from the shape data; By integrating the detected crack candidate from the image data, to execute the steps of outputting a test result, and outputting the outputted the inspection results to the user, the.

1 is a block diagram functionally illustrating an inspection apparatus according to a first embodiment. The perspective view explaining an example of the inspection apparatus shown in FIG. 1 and its inspection object. The flowchart explaining the crack inspection process content of the inspection apparatus which concerns on 1st Embodiment. A block diagram functionally explaining an inspection device concerning a 2nd embodiment. The flowchart explaining the crack inspection process content of the inspection apparatus which concerns on 2nd Embodiment. The figure explaining the positional relationship of the sensor part with respect to a test | inspection object, and an example of luminance distribution of image data. The block diagram functionally explaining the inspection device concerning a 3rd embodiment. The flowchart explaining the crack inspection process content of the inspection apparatus which concerns on 3rd Embodiment. The figure explaining the relationship between the image data correct | amended by correction amount data, the luminance distribution, and a crack discrimination threshold value. The block diagram of the inspection apparatus at the time of combining 2nd Embodiment and 3rd Embodiment.

  Hereinafter, an inspection apparatus according to an embodiment will be described with reference to the drawings. In the following description, components having substantially the same functions and configurations are denoted by the same reference numerals, and redundant description will be provided only when necessary.

(First embodiment)
FIG. 1 is a block diagram for functionally explaining the configuration of the inspection apparatus 1 according to the first embodiment. As shown in FIG. 1, in this embodiment, an inspection apparatus 1 that inspects the presence or absence of a crack in an inspection object 100 includes a scanning apparatus 200, a control apparatus 300, an arithmetic processing apparatus 400, and a recording apparatus 500. And a detection result output device 600.

  The inspection object 100 is, for example, a stationary blade of a gas turbine and has a complicated shape, but is a mechanical structure manufactured with high accuracy. The scanning device 200 includes a sensor unit 210 and a moving mechanism 220. The sensor unit 210 includes an image data acquisition unit 212 and a shape data acquisition unit 214, and measures the surface of the inspection object 100 in order to detect a crack from the surface of the inspection object 100. The moving mechanism 220 includes an X moving stage 230, a Y moving stage 232, a Z moving stage 234, an X rotating stage 240, a Y rotating stage 242, and a Z rotating stage 244, and the sensor unit 210 is inspected in various shapes. It is possible to move along the surface of the object 100 and measure the entire surface of the inspection object 100. The scanning device 200 further includes an illumination 250. The illumination 250 moves with the moving mechanism 220, and irradiates the inspection object 100 with light when the sensor unit 210 acquires image data and shape data.

  More specifically, the image data acquisition unit 212 provided in the sensor unit 210 is configured by, for example, a camera that can acquire image data. That is, the image data acquisition unit 212 acquires two-dimensional image data captured by the camera. Similarly, the shape data acquisition unit 214 provided in the sensor unit 210 is configured by, for example, a stereo camera based on the principle of triangulation or a sensor that combines a laser and a camera, and the shape data of the inspection object 100 is obtained. get. That is, the shape data acquisition unit 214 acquires shape data that is an aggregate of three-dimensional coordinate data including X coordinates, Y coordinates, and Z coordinates.

  The moving mechanism 220 performs movement control of the sensor unit 210 in order to acquire the entire surface data including the back surface and the side surface of the inspection object 100. Specifically, the X moving stage 230 of the moving mechanism 220 has a function of moving the sensor unit 210 in the X direction, and the Y moving stage 232 has a function of moving the sensor unit 210 in the Y direction. The Z moving stage 234 has a function of moving the sensor unit 210 in the Z direction.

  Further, the X rotation stage 240 of the moving mechanism 220 has a function of rotating the sensor unit 210 around the X axis, and the Y rotation stage 242 has a function of rotating the sensor unit 210 around the Y axis. The Z rotation stage 244 has a function of rotating the sensor unit 210 around the Z axis.

  That is, the moving mechanism 220 moves the sensor unit 210 along the three axes XYZ and rotates it along the three axes XYZ so that the entire surface of the inspection object 100 can be measured. The movement control 210 is performed. In other words, the moving mechanism 220 is a mechanism that moves the sensor unit 210 including the image data acquisition unit 212 and the shape data acquisition unit 214 by movement control and rotation control of a total of six axes. The movement amount and the rotation amount are controlled by the movement amount and the rotation amount, and the movement amount and the rotation amount can be recorded in the recording apparatus 500. The illumination 250 of the moving mechanism 220 has a role of illuminating the surface of the inspection object 100 and assisting in acquiring image data. The illumination 250 can be additionally provided, for example, in the sensor unit 210, but may be omitted when the inspection apparatus 1 has sufficient brightness for inspecting the inspection object 100. .

  The control device 300 controls communication of the sensor unit 210 and the moving mechanism 220 including the image data acquisition unit 212 and the shape data acquisition unit 214 based on the measurement plan. In the present embodiment, the control device 300 includes an image data acquisition unit controller 312 and a shape data acquisition unit controller 314.

  The image data acquisition unit controller 312 transmits control data to the image data acquisition unit 212 and controls the operation of the image data acquisition unit 212. The shape data acquisition unit controller 314 transmits control data to the shape data acquisition unit 214 and controls the operation of the shape data acquisition unit 214.

  Further, the control device 300 includes an X movement stage controller 330, a Y movement stage controller 332, a Z movement stage controller 334, an X rotation stage controller 340, a Y rotation stage controller 342, a Z rotation stage controller 344, and an illumination. The control controller 350 is provided. The X movement stage controller 330, the Y movement stage controller 332, the Z movement stage controller 334, the X rotation stage controller 340, the Y rotation stage controller 342, and the Z rotation stage controller 344, respectively, Control data is transmitted to the Y movement stage 232, the Z movement stage 234, the X rotation stage 240, the Y rotation stage 242, and the Z rotation stage 244 to perform operation control. The illumination controller 350 controls the current or voltage of the illumination 250 to control the illuminance of the illumination 250 and to control the position and orientation of the illumination 250.

  In order to measure the surface of the inspection object 100 using the scanning device 200 and the control device 300, it is necessary to set a measurement trajectory path to be scanned by the scanning device 200 in advance for the control device 300. The measurement trajectory path can be set, for example, by recording the coordinates and posture angles of the position to be measured and enumerating them, thereby setting the start point and the end point at which the moving mechanism 220 moves. By setting the combination of the start point and the end point for all routes, the control device 300 can automatically issue a command to the scanning device 200.

  Further, when the inspection object 100 moves or is replaced with a separate body, the positional relationship between the inspection object 100 and the scanning device 200 may change. Therefore, the control device 300 determines the current position of the inspection object 100 and the scanning apparatus 200 with respect to the measurement orbit path when the positional relationship between the inspection object 100 that created the measurement orbit path and the scanning apparatus 200 is used as a reference position. It has a function to obtain the position and posture angle from the relationship. The control device 300 corrects the measurement trajectory path to be scanned by the scanning device 200 with respect to the current inspection object 100 by performing coordinate conversion on the reference measurement trajectory route based on the obtained position and orientation angle. It becomes possible to do.

  In order to obtain the current positional relationship between the inspection object 100 and the scanning device 200 and the positional relationship between the inspection object 100 and the scanning device 200 as a reference, for example, a sensor including a shape data acquisition unit 214 is used. The position of the current position with respect to the reference position is obtained by measuring the shape of the inspection object 100 by the unit 210 and performing alignment adjustment calculation using the obtained shape data and the shape data of the inspection object 100 at the reference position. And the posture angle can be obtained.

  The measurement trajectory path to be set for the control device 300 is recorded in advance in the recording device 500, and the control device 300 reads the data of the measurement trajectory route recorded in the recording device 500, whereby the scanning device 200. Control the movement and posture of the camera. Thereby, using the image data acquisition unit 212 and the shape data acquisition unit 214, data obtained by measuring the surface of the inspection object 100 can be automatically acquired.

  The arithmetic processing device 400 includes a shape data crack detection processing unit 410, a calibration unit 420, a surface identification unit 430, an image correction determination unit 440, an image correction unit 450, an image data crack detection unit 460, A crack integration unit 470 and a result output processing unit 480 are provided. Each of these processing units is a process realized when the arithmetic processing device 400 reads and executes a program recorded in the recording device 500. Alternatively, the arithmetic processing device 400 may be configured by an ASIC (Application Specific Integrated Circuit) in which the processing executed by each of these processing units is previously incorporated as a function, and these processing may be executed. Details of these processing units will be described later. The arithmetic processing device 400 constitutes a crack candidate arithmetic processing device in the present embodiment.

  FIG. 2 is a perspective view for explaining the inspection apparatus 1 according to the present embodiment and an example of the inspection object 100 that performs inspection for the presence or absence of a crack using the inspection apparatus 1. As shown in FIG. 2, an inspection object 100 is placed below the inspection apparatus 1. The scanning device 200 is positioned above the inspection object 100.

  The X moving stage 230 in the scanning apparatus 200 is configured by a guide rail extending in the X direction, the Y moving stage 232 is configured by a guide rail extending in the Y direction, and the Z moving stage 234 extends in the Z direction. It is composed of guide rails. A sensor unit 210 is attached to the Z movement stage 234 via a support mechanism, and the sensor unit 210 can be moved in the Z direction. The Z movement stage 234 can be moved in the X direction by the X movement stage 230, and the X movement stage 230 can be moved in the Y direction by the Y movement stage 232. As a result, the sensor unit 210 can move in three directions, the X direction, the Y direction, and the Z direction.

  Further, the sensor unit 210 is provided with an X rotation stage 240, a Y rotation stage 242, and a Z rotation stage 244, and is rotatable about three axes. Further, the sensor unit 210 is also provided with an illumination 250, which can brightly illuminate the inspection object 100 when acquiring image data and shape data.

  As described above, the scanning device 200 is controlled by the control device 300. Overall control of the control device 300 is performed by a computer 610. That is, the arithmetic processing device 400, the recording device 500, and the detection result output device 600 described above are realized by the computer 610. The computer 610 constitutes the inspection apparatus control apparatus in the present embodiment, and the scanning apparatus 200 and the control apparatus 300 controlled by the inspection apparatus control apparatus constitute an inspection apparatus main body in the present embodiment. In addition, the computer 610 reads and executes a predetermined program (for example, a program recorded in the recording apparatus 500), and controls the inspection apparatus main body in the present embodiment.

  Next, the flow of crack inspection processing will be described based on the flowchart shown in FIG. As shown in FIG. 3, in the crack inspection process using the inspection apparatus 1, a user such as an operator of the inspection apparatus 1 inputs a measurement trajectory path to the control apparatus 300 (step S100). Specifically, as described above, the control device 300 reads the measurement trajectory path data recorded in the recording device 500 via the arithmetic processing device 400 based on an instruction from the user.

  Next, the user of the inspection apparatus 1 places the inspection object 100 at the measurement position (step S102). Subsequently, the control device 300 of the inspection apparatus 1 controls the scanning device 200 to acquire the shape data of the inspection object 100 by the shape data acquisition unit 214 (step S104), and further, the image data acquisition unit 212. Thus, the image data of the inspection object 100 is acquired (step S106). Based on the acquired shape data and image data, crack candidate detection described later is performed.

  Next, the shape data crack detection processing unit 410 of the arithmetic processing device 400 detects a crack candidate based on the shape data acquired in step S104 (step S108). That is, the shape data crack detection processing unit 410 detects a portion opened on the surface of the inspection object 100 as a crack candidate from the shape data. In this embodiment, there are two methods for detecting crack candidates from shape data.

  The first method is to obtain new shape data without cracks as a reference in advance, and minimize the difference between the current shape data of the inspection object 100 and the distance between each measurement point. Align by square method etc. and calculate the position and posture angle. The shape data acquired in advance may be measured by the inspection apparatus 1 or may be obtained by calculation from design data or the like.

  Then, from the result of alignment by the least square method or the like, the current shape data with respect to the reference shape data is coordinate-converted using the position and orientation angle. As a result, the coordinates of the measurement points between the reference shape data and the current shape data can be compared and calculated, and an open crack candidate can be detected.

  The second method of crack candidate detection method from shape data is to cut out an arbitrary cross section from the current shape data and determine the gradient angle between measurement points that are continuously adjacent to the measurement points existing along this cross section. Obtained by differential operation. A threshold value determined to have no flatness is set from the obtained differential calculation result, and measurement points that are equal to or greater than the threshold value are extracted. The differential value in the front-rear direction is inspected with respect to the cross-sectional direction with the extracted measurement point as the center, the opening portion that is judged to be flat is extracted, the open section in the cross-sectional direction is extracted, and this opening The section is detected as a crack candidate.

  Next, the calibration unit 420 of the arithmetic processing device 400 performs alignment between the shape data and the image data (step S110). In other words, the calibration unit 420 integrates image data and shape data of the same measurement region. This process is a process of associating pixels of image data with measurement points of shape data. Here, the case where the image data has a finer resolution than the shape data will be described, but the same processing can be performed in the opposite case. As a method for detecting a portion where the pixel of the image data and the measurement point of the shape data coincide with each other, a member whose length or width is known in advance is measured, and the distance from the member at this time to the origin of the scanning device 200 is recorded. To do. The pixel of the image data of the marking part recorded on the both ends or the member at an arbitrary distance and the measurement point of the shape data are extracted.

  For example, if the pixels of the image data are the start point (100, 100) and the end point (500, 100), and the measurement coordinates (10, 10) and end point (50, 10) of the shape data corresponding to the start point of the image data are By recording each coordinate value, a resolution coefficient necessary for associating the image data with the shape data can be obtained, and the data can be interpolated with respect to the lower resolution of the image data or the shape data. . The image data and shape data of the inspection object 100 and the scanning apparatus 200 are desirably acquired at least in two places. Furthermore, in order to obtain a difference depending on the posture, three or more pieces of image data are obtained. It is desirable to acquire measurement data. This is because the three axes of the posture angle are stably acquired by the least square method.

  Next, the surface identification unit 430 of the arithmetic processing unit 400 calculates surface information from the shape data and identifies the surface (step S112). Specifically, the surface identification unit 430 obtains a normal vector of the surface to which each measurement point belongs from the shape data. For the normal vector, a triangular network is created using the measurement points around each measurement point as a reference, and the normal measurement point vector is obtained from the coordinate value of each vertex by a general method. When eight surrounding points are selected, eight normal vectors are obtained, and by synthesizing these normal vectors, it is possible to obtain the normal vector of the surface of the measured measurement point. The normal vector obtained here and the direction of the surface recorded in the reference shape data prepared in advance, that is, the corresponding surface is classified within the range of the angle of the normal vector with respect to the reference coordinate system, and each measurement point is the front or back Identify whether it is a side or a side. Furthermore, the surface identification unit 430 has a function of recording which surface the detected crack candidate belongs to the surfaces classified here.

  Next, the image correction determination unit 440 of the arithmetic processing device 400 calculates correction amount data from the luminance of the image data at the same position as the crack candidate in the shape data and the luminance of the image data in other regions. (Step S114). Specifically, the image correction determination unit 440 first detects a crack candidate opened from the shape data based on the result of association between the obtained image data and the shape data by the method of the calibration unit 420, and responds to it. The luminance data of the image data in the same amount measurement area to be acquired is acquired. Moreover, the correction amount data of the image which detects a crack candidate is calculated by acquiring the brightness | luminance of area | regions other than a crack candidate. For example, from the acquired image data, when the luminance A1 of the crack candidate region and the luminance A2 of the other region are 50 and 100 (when the range of luminance value is 0 to 255), respectively, When the luminance difference between the crack candidate region and the other region is large, it becomes easier to distinguish from the crack candidate. Therefore, ideally, the crack candidate area has a low brightness B1 such as 0 or 10 and the other areas have a high brightness B2 such as 250 or 255. The difference between the actual brightness A1 and the brightness A2 at this time is the brightness difference A. And the difference between the corrected brightness B1 and brightness B2 is the brightness difference B. That is, it can be said that the image correction determination unit 440 has a function of calculating correction amount data in which the ratio of the luminance difference A and the luminance difference B increases.

  The image correction unit 450 of the arithmetic processing device 400 corrects the luminance distribution of the image data from the correction amount data (step S116). Specifically, the image correction unit 450 corrects the luminance of the image acquired by the image data acquisition unit 212 based on the correction amount data obtained by the above method. Here, the method of correcting the luminance difference A between the presence and absence of a crack candidate in the original image to the luminance difference B between the presence and absence of a crack candidate based on the correction amount is performed by a method such as linear interpolation of luminance or gamma correction. The

  Next, the image data crack detection unit 460 of the arithmetic processing device 400 detects a crack candidate based on the corrected image data (step S118). Specifically, the image data crack detection unit 460 reads the image data corrected by the image correction unit 450 and detects a crack candidate based on the image data. Regarding the flow of processing for detecting crack candidates, first, a small region corresponding to a region size arbitrarily set is first extracted from the corrected image data, and edge detection processing is performed on the small region. Note that the small area extraction process is not necessarily required, and thus the process can be omitted.

  In the subsequent edge detection processing, the edge intensity and density gradient are calculated by the edge detection filter (for example, Sobel filter) for the adjacent four or eight adjacent luminance values based on the luminance value of the pixel of interest. To do. In addition, by binarization processing using a threshold value arbitrarily set for the edge strength, it is determined whether or not it is an edge, and an edge region is detected. Further, by using the density gradient result and the edge region detection result, pixels having similar values in the density gradient direction are traced by the inner product between adjacent pixels. By using the fact that the crack candidate on the image becomes a closed region, the tracking result returns to the closed region (the pixel of the same starting point when tracking in either direction from the pixel that is an arbitrary starting point) ) Is detected.

  Here, the detected closed region may include other than the crack candidate. The area other than the crack candidate and the area other than the crack candidate are discriminated from calculation information such as the area of the pixel and the shape of the closed area, and the erroneously detected portion is removed. Although these processes are performed on all image data, the area may be interrupted during the process, and one crack may be recognized as two or more crack candidates. Therefore, if the contours of the closed regions are within the arbitrarily set pixel distance by region connection processing, a blank region that is not recognized as a crack candidate according to the width of the contour is recognized as a crack candidate, and two or more Concatenate the closed areas. The connected crack candidate is a crack candidate detected by the image data crack detection unit.

  Next, the crack integration unit 470 of the arithmetic processing device 400 integrates the crack candidate detected from the shape data and the crack candidate detected from the image data (step S120). Specifically, the crack integration unit 470 treats both the crack candidate detected by the shape data crack detection processing unit 410 and the crack candidate detected by the image data crack detection unit 460 as a crack candidate. To do. Here, according to the resolution of the image data acquisition unit 212 and the shape data acquisition unit 214, the width threshold value of the crack candidate is set to an arbitrary value. And the width | variety of the crack candidate calculated | required by each is measured. The width of the crack candidate is measured by calculating the distance between the start point and the end point of the contact point of the perpendicular and the contour of the crack perpendicular to the center line, and select an arbitrary value from the average, maximum, and minimum values of the crack. Calculate as width.

  However, if there is a branch point in the crack, the obtained crack width may be long, so it is possible to select whether or not to exclude the section before and after the arbitrarily set branch point. Good. If the width of the crack candidate is narrower than the threshold among the crack candidates detected by the image data crack detection unit 460 based on the measurement result of the width, the crack detection result is output as the crack candidate width. If it is wider, erase it. Similarly, out of the crack candidates detected by the shape data crack detection processing unit 410, those wider than the threshold are output as cracks, and narrow ones are deleted. As a result, the same crack candidate is not duplicated and output as a crack.

  In this embodiment, when integrating the crack candidate detected from the shape data and the crack candidate detected from the image data in step S120, the crack candidate detected from the shape data is not output. The crack candidates detected from the image data may be output. This is because the crack candidates detected based on the shape data are often relatively large and are easily recognized by the naked eye, while the crack candidates detected based on the image data are often relatively small. This is because, since it is difficult to recognize, it is considered that it is sufficient for the inspection apparatus 1 to output only crack candidates based on image data that is difficult to recognize with the naked eye.

  Next, the result output processing unit 480 of the arithmetic processing device 400 outputs the inspection result to the detection result output device 600 (step S122). Specifically, the result output processing unit 480 obtains the contour of the shape of the inspection object 100 created from the surface information obtained by the surface identifying unit 430, the crack candidates integrated by the crack integrating unit 470, and the shape data. Combine and create a crack detection result projection image from an arbitrary viewpoint. The method of detecting the shape contour is based on the shape change point serving as the edge and the normal vector or shape gradient amount (differential value around the shape data) for each of the shape data of the inspection object 100 and the reference shape data. Detect from. Using the detected shape change points, only the points where the change points are continuous are detected. In the case of discontinuity, the point is detected as a continuous point if it is within a threshold value that is determined as a continuous store of arbitrarily set shape change points. Thereby, the shape outline of the test object 100 is detected.

  Subsequently, the result output processing unit 480 creates a projection image obtained by projecting the detected contour of the inspection object 100 onto the two-dimensional space from an arbitrarily designated viewpoint on the three-dimensional space. The arbitrarily designated point is, for example, a state viewed from the upper surface or a side surface of the inspection object 100, and the viewpoint can be changed as appropriate. Thus, the crack inspection process executed by the arithmetic processing unit 400 is completed.

  Returning to FIG. 1 again, the detection result output device 600 outputs the inspection result from the result output processing unit 480 to the user. Specifically, the detection result output device 600 includes a means for inputting the coordinate value of the viewpoint and a display means for confirming the shape data of the inspection object 100 from an arbitrary viewpoint in the three-dimensional space. Means provided by the user with a mouse or keyboard that is an input assist device for the device 400 is provided. Further, crack information such as crack width is drawn on the composite image in the crack detection result projection image. With this detection result output device 600, the user can grasp the crack candidates automatically detected by the inspection device 1.

  Next, a modified example of the image correction determination unit 440 in the arithmetic processing device 400 will be described. In the above description, the image correction determination unit 440 generates the correction amount data in which the luminance difference between the region where the crack candidate exists and the other region increases. If there is no crack candidate, this correction amount is generated. Data no longer exists. The processing method in that case will be described below. If no crack candidate is detected in the shape data, correction amount data cannot be obtained. Therefore, the data to which the corresponding image data and shape data belong is collated with the surface information obtained by the surface identifying unit 430, and the correction amount data of the corresponding surface information is extracted from the correction amount data. Further, by extracting the correction amount data having the closest positional relationship that can detect the crack candidate from the image data and the shape data at the position where the crack candidate does not exist, the photographing situation and shape of the image data are extracted. It is possible to acquire correction amount data in a state in which the data measurement state is very similar.

  Furthermore, since the photographing situation differs depending on the distance between the inspection object 100 and the sensor unit 210 of the scanning device 200, the correction amount data of the brightness of the image data in the image correction determination unit 440 described above also differs. . This is because the irradiation intensity by the illumination 250 differs depending on the distance between the inspection object 100 and the sensor unit 210 of the scanning apparatus 200, and the correction amount data needs to be corrected according to the distance at the time of photographing. Therefore, in the present embodiment, the following method is used to calculate correction amount data according to the distance between the inspection object 100 and the sensor unit 210 of the scanning device 200.

  The correction amount data calculation method according to the distance of the image correction determination unit 440 acquires image data in advance by changing the distance between the members simulating the crack, and there is no crack in the region where the crack exists. Correction amount data is calculated from the luminance ratio of the region. At this time, for a distance for which image data has not been acquired, an expression for correcting the correction amount data by linear interpolation or polynomial approximation is obtained from the correction amount data of the preceding and following distances. For the linear interpolation or polynomial approximation, an interpolation method that approximates the measured value according to the irradiation state of the illumination 250 is selected. Finally, the correction amount data is calculated by applying the correction amount data to the equation for correcting the correction amount data according to the distance at the actually measured distance. According to this method, it is possible to calculate correction amount data suitable for an actual photographing situation even when the measurement distances are different.

  Further, the image correction determination unit 440 can cope with a luminance change in the captured image due to the illumination posture of the illumination 250 generated by the inclination of the surface of the inspection object 100 and the sensor unit 210. The shape data acquired by the shape data acquisition unit 214 of the sensor unit 210 can acquire a three-dimensional cross section or plane, and can determine the posture angle of the sensor unit 210 with the surface of the inspection object 100. . Since the measurement direction of the shape data acquisition unit 214 of the sensor unit 210 is constant, the posture angle is obtained by the least square method or the like if the coordinate values of the measured shape data are at least three points. Is possible. Image data is acquired in advance by changing the posture angle between members simulating a crack, and correction amount data is calculated from the luminance ratio of the crack and other regions.

  At this time, for a posture angle for which image data has not been acquired, an equation for correcting the correction amount data by linear interpolation or polynomial approximation is obtained from the correction amount data of the front and rear posture angles. For the linear interpolation or polynomial approximation in the correction formula, an interpolation method that approximates the measured value according to the irradiation state of the illumination 250 is selected. Finally, the correction amount data is calculated by applying the correction amount data to the equation for correcting the correction according to the distance at the posture angle actually being measured. This method makes it possible to calculate correction amount data suitable for the actual shooting situation even when the measurement posture angle is different.

  As described above, according to the inspection apparatus 1 according to the present embodiment, in order to detect a crack on the surface of the inspection object 100, the luminance distribution of the image data is corrected based on the shape data. The image data necessary for the inspection can be acquired without causing the image data of the inspection object 100 to be defective. Thereby, the inspection apparatus 1 with a good detection rate can be realized in a short time by detecting a crack based on the image data.

  More specifically, in the inspection apparatus 1 according to the present embodiment, the luminance of the image data corresponding to the crack candidate region detected by the shape data with respect to the shape data and image data acquired for the inspection object 100. By correcting the distribution, a finer crack that cannot be detected by shape data can be detected. Further, in the crack detection process using only image data, it is possible to detect a crack well even for the inspection object 100 whose shape has changed which has led to detection omission. Furthermore, compared with the crack inspection using only the shape data, the crack inspection process using the image data is performed to automate the crack inspection on the surface of the inspection object 100, and the time required for the crack inspection process is reduced. Shortening and cost reduction can be achieved. Moreover, it can contribute to the improvement of the yield and the process shortening in the crack inspection process.

(Second Embodiment)
In the inspection apparatus 1 according to the second embodiment, the arithmetic processing device 400 controls the position and orientation of the sensor unit 210 and the illumination 250 to obtain image data on the surface of the inspection object 100 in a state more suitable for inspection. It is something that can be done. Hereinafter, a different part from 1st Embodiment mentioned above is demonstrated.

  FIG. 4 is a block diagram for functionally explaining the configuration of the inspection apparatus 1 according to the second embodiment, and corresponds to FIG. 1 in the first embodiment described above.

  As shown in FIG. 4, the arithmetic processing device 400 of the inspection apparatus 1 according to the present embodiment is different from the arithmetic processing device 400 of the first embodiment in a position / orientation calibration unit 700, an illumination control unit 710, and a sensor. The unit position and orientation control unit 720 is further provided.

  The position / orientation calibration unit 700 acquires the relative position and orientation of the illumination unit 250 and the sensor unit 210 that acquires image data and shape data, and calibrates the image data. The illumination control unit 710 controls the position and orientation of the illumination 250 and adjusts the luminance distribution of the image data. The sensor unit position / orientation control unit 720 determines the position and orientation of the sensor unit 210 such that the ratio of the luminance between the crack region and the other region increases based on the luminance correction amount data calculated by the image correction determination unit 440. The sensor unit 210 is controlled based on the calculation result.

  FIG. 5 is a diagram illustrating a flow chart for explaining the flow of the crack inspection process of the inspection apparatus 1 according to the present embodiment, and corresponds to FIG. 3 in the first embodiment described above. Explaining the difference from the first embodiment described above, after acquiring image data in step S106, in this embodiment, the position / orientation calibration unit 700 in the arithmetic processing unit 400 acquires image data and shape data. A relative position and orientation between the sensor unit 210 and the illumination 250 are acquired, and calibration is performed (step S200).

  More specifically, the position / orientation calibration unit 700 is a position facing the illumination 250 in order to obtain the relative position and orientation of the illumination unit 250 and the sensor unit 210 that acquires image data and shape data. A flat plate is placed on the camera, and the irradiation state is imaged by the image data acquisition unit 212 of the sensor unit 210. At this time, as shown in FIG. 6, photographing is performed with a combination of at least three patterns of the irradiation distance L and the relative angle θ of the image data acquisition unit 212, the shape data acquisition unit 214, and the illumination 250. To do. Since the brightness distribution of the captured image data as shown in FIG. 6 is obtained from the image captured at that time, this is referred to as an irradiation intensity distribution (corresponding to an object reflection distribution).

  The irradiation intensity distribution, that is, the luminance distribution is recorded in each combination, and the average luminance of the area divided by the broken line in FIG. 6 is acquired. It is desirable to arbitrarily set the area to be divided and to provide at least four areas. Here, as an example, it will be described as a nine-divided region (L11 to L33) shown in FIG. 6, and the luminance distribution is assumed to be a distribution that is strong from L11 to L13 and weakened from L31 to L33. The coefficient of the correction formula of the luminance distribution starting from L11, L12, L13 and L11, L21, L31 in the L11 to L13 direction and the L11 to L31 directions is calculated from the average luminance of each region. Here, it is assumed that the correction formula of the luminance distribution exists according to the number of divided areas, and the method of calculating the coefficient is based on linear correction or polynomial approximation. For the linear interpolation or polynomial approximation in the correction formula, an interpolation method that approximates the measured value according to the irradiation state of the illumination 250 is selected.

  In the calibration in step S200, the position / orientation calibration unit 700 adjusts the illumination control unit 710 and the sensor unit position / orientation control unit 720, and then returns to step S106 to acquire image data a plurality of times. That is, for the relative positions and orientations of the image data acquisition unit 212, the shape data acquisition unit 214, and the illumination 250, a correction formula is created by combining a plurality of patterns, and the irradiation distance and the relative angle are referred to together. A table is generated and recorded in the recording device 500.

  The sensor unit 210 is moved or rotated by the moving mechanism 220 and is photographed when data is being measured by the sensor unit 210 including the image data acquisition unit 212 and the shape data acquisition unit 214 that measure the surface of the inspection object 100. As shown in FIG. 6, an arbitrarily set number of regions are segmented to obtain a luminance distribution, that is, an illumination intensity distribution. At this time, it is desirable to set the same number as the number of areas at the time of calibration. If they are different, a correction coefficient between the areas is obtained in advance and recorded in the reference table of the recording apparatus 500.

  Through these processes, the relative positions and orientations of the current image data acquisition unit 212, shape data acquisition unit 214, and illumination 250 are obtained by referring to the reference table recorded in the recording apparatus 500. At this time, the illumination 250 is irradiated so that the illumination 250 is uniformly irradiated when the brightness distribution (irradiation intensity distribution of illumination) is at a position and orientation that are estimated to be biased at or above the illumination bias determination threshold value arbitrarily set. To control. The state in which the illumination 250 is uniformly irradiated is based on the reference data of the correction formula in a state where a flat plate is placed at a position facing the illumination 250 at the time of calibration described above. The relative positions and orientations of the image data acquisition unit 212, the shape data acquisition unit 214, and the illumination 250 that coincide with the luminance distribution at this time or fall within the range below the approximate threshold of the arbitrarily set luminance distribution are acquired from the reference data. . Thereby, the distribution of the brightness of the crack candidate opened on the surface of the inspection object 100 and the other area creates a state in which the crack candidate can be easily detected, and the illumination control unit 710 is communication-controlled by the arithmetic processing unit 400. . Then, the process proceeds to step S108 described in the first embodiment.

  Furthermore, in the present embodiment, also in step S114, the image correction determination unit 440 calculates correction amount data related to the brightness of the image data. The image correction determination unit 440 further calculates the correction amount data based on the calculated correction amount data. The position and orientation of the sensor unit 210 are controlled so that the luminance ratio between the crack candidate region and the other region is increased. From the acquired correction amount data, the ratio between the crack candidate region of the current image data acquired by the image data acquisition unit 212 and the region other than the crack candidate is obtained, and correction amount data whose ratio further increases is acquired. At this time, an arbitrarily set threshold for the ratio may be provided, an upper limit may be set, and the correction amount data may be acquired. Accordingly, the position and orientation of the sensor unit 210 corresponding to the acquired correction amount data are acquired, and the sensor unit position / orientation control unit 720 performs communication control as the arithmetic processing device 400. In step S116, the image correction unit 450 corrects the luminance distribution of the image data based on the correction amount data.

  As described above, also by the inspection apparatus 1 according to the present embodiment, the inspection apparatus 1 can be realized in a short time and with a high detection rate by detecting a crack based on the image data. Further, in order to detect a crack on the surface of the inspection object 100, the luminance distribution is corrected, and the position and orientation of the sensor unit 210 and the illumination 250 are controlled during measurement. Data necessary for inspection can be efficiently acquired without causing image data to be defective.

(Third embodiment)
In the inspection device 1 according to the third embodiment, the arithmetic processing device 400 detects a crack by changing a crack determination threshold value of a feature amount for determining a crack based on the calculated correction amount data. The accuracy is improved. Hereinafter, a different part from 1st Embodiment mentioned above is demonstrated.

  FIG. 7 is a block diagram for functionally explaining the configuration of the inspection apparatus 1 according to the third embodiment, and corresponds to FIG. 1 in the first embodiment described above.

  As shown in FIG. 7, the arithmetic processing device 400 according to the present embodiment is configured to further include a crack determination threshold value processing unit 800 with respect to the arithmetic processing device 400 of the first embodiment. The crack determination threshold value processing unit 800 has a function of changing a crack determination threshold value of a feature amount for determining a crack in the luminance distribution of image data corrected based on the correction amount data calculated by the image correction determination unit 440. Have.

  FIG. 8 is a diagram illustrating a flow chart for explaining the flow of the crack inspection process of the inspection apparatus 1 according to the present embodiment, and corresponds to FIG. 3 in the first embodiment described above. Explaining the difference from the first embodiment described above, after correcting the luminance distribution of the image data in step S116, in this embodiment, the crack discrimination threshold processing unit 800 in the arithmetic processing unit 400 discriminates the crack. A crack determination threshold value, which is a luminance threshold value, which is a feature amount to be set, is dynamically set based on the luminance distribution (step S300).

  More specifically, the crack determination threshold value processing unit 800 uses the brightness correction amount data based on the distance and the brightness correction amount data based on the inclination of the inspection object 100 and the sensor unit 210 to display the image data shown in FIG. Control is performed to vary the crack discrimination threshold for the luminance distribution of the image. In the case of a fixed crack discrimination threshold for the luminance distribution of the image data in FIG. 9, it is possible to discriminate both crack candidate areas and areas including other than crack candidates as crack candidates. There is sex. Therefore, the crack determination threshold value processing unit 800 sets a crack determination threshold value associated with the luminance distribution correction amount data, and dynamically sets a crack determination threshold value for an actually captured image.

  For example, the image data as shown in FIG. 9A is corrected based on the correction amount data, and the corrected image data as shown in FIG. 9B is acquired. A luminance distribution as shown in FIG. 9C is extracted from the corrected image data in FIG. 9B, and the two luminance values in the region including the crack candidate and the region not including the crack candidate are extracted. Extract a mountain and a valley. For example, as shown in FIGS. 9A and 9B, when a crack in the image is open, illumination light is difficult to enter to the inside, and thus a dark luminance distribution is often exhibited. On the other hand, the region that does not include the crack candidate is often illuminated with light from the illumination 250 on the surface, and thus often exhibits a bright luminance distribution. In such a case, the mountain closest to the luminance value 0 can be determined as the region including the crack candidate, and the mountain closest to the luminance value 255 can be determined as the region not including the crack candidate.

  Here, it is determined whether the luminance distribution is only a single peak or two or more peaks. In this determination, the peak of the mountain is detected from the frequency number of the number of pixels. It is assumed that there is no peak in the mountain unless the frequency is arbitrarily set.

  Here, when there is one peak and the peak luminance value is close to 255 (for example, within 20% of the maximum luminance value 255), it is assumed that no crack candidate is present in the image data, and the peak is the highest. A luminance value having a low luminance and a frequency equal to or lower than a predetermined frequency number is set as a crack determination threshold value. When there is only one mountain and the peak luminance value is close to 0 (for example, within 20% of the minimum luminance value 0), the highest luminance from the peak is assumed that a crack candidate exists in the image data. In addition, a luminance value whose frequency is equal to or less than a predetermined frequency number is set as a crack determination threshold value.

  When there are a crack candidate region and a region other than the crack candidate, usually two or more mountains exist, and it is determined that the region is bimodal. In this case, the peak of the peak is detected by the above-described method, and a valley exists between the luminance values of the peaks of both peaks. Therefore, the luminance value with the smallest frequency of pixel values is detected. This makes it possible to variably set a crack discrimination threshold at a portion that is a boundary between regions other than a crack candidate and a crack candidate.

  Taking the luminance distribution of FIG. 9C as an example, as shown in FIG. 9D, in the first and second embodiments described above, the crack determination threshold is a fixed value. As shown to (e), in 3rd Embodiment, it moves to the direction of a luminance value with a small crack discrimination threshold value. For this reason, it becomes possible to eliminate the influence of noise or the like and set a crack determination threshold value with appropriate luminance.

  Then, the image data acquisition unit 212 uses the crack determination threshold associated with the correction amount data calculated by the image correction determination unit 440 for the image data captured by the image data acquisition unit 212 of the sensor unit 210. It becomes possible to set a variable crack discrimination threshold for the captured image data, and the image data crack detection unit 460 uses this crack discrimination threshold to detect a crack in the image data. Will be able to. After step S300, the image data crack detection unit 460 uses the set crack determination threshold value to detect a crack candidate based on the image data in step S118.

  As described above, according to the inspection apparatus 1 according to the present embodiment, since the crack determination threshold value processing unit 800 is additionally provided in the arithmetic processing device 400, the correction amount data by the image correction determination unit 440 and the variable amount are variable. The crack determination threshold value processing unit 800 capable of setting a crack determination threshold value can detect the crack on the surface of the inspection object 100 with high accuracy without causing the image data of the inspection object 100 to be defective. become.

  Note that this embodiment can also be applied to the second embodiment. That is, as shown in FIG. 10, by providing a crack determination threshold value processing unit 800 in the arithmetic processing unit 400 of the inspection apparatus 1 according to the second embodiment, when acquiring image data, the sensor unit 210 and the illumination unit While controlling the position and orientation of 250, the crack discrimination threshold for the acquired image data may be changed, and a crack on the surface of the inspection object 100 may be detected with high accuracy.

  Although several embodiments have been described above, these embodiments are presented as examples only and are not intended to limit the scope of the invention. The novel apparatus and methods described herein can be implemented in a variety of other forms. In addition, various omissions, substitutions, and changes can be made to the forms of the apparatus and method described in the present specification without departing from the spirit of the invention. The appended claims and their equivalents are intended to include such forms and modifications as fall within the scope and spirit of the invention.

DESCRIPTION OF SYMBOLS 1: Inspection apparatus, 100: Inspection object, 200: Scan apparatus, 210: Sensor part, 212: Image data acquisition part, 214: Shape data acquisition part, 220: Movement mechanism, 250: Illumination, 300: Control apparatus, 400 : Arithmetic processing device, 500: recording device, 600: detection result output device

Claims (11)

A sensor unit for acquiring image data and shape data from the surface of the inspection object;
A moving mechanism that moves and rotates the sensor unit around the inspection object;
A control unit for controlling the sensor unit and the moving mechanism;
A control device for an inspection apparatus that controls an inspection apparatus main body, comprising:
A crack candidate is detected from the shape data, the image data is corrected based on the shape data, a crack candidate is detected from the corrected image data, and a crack candidate detected from the shape data is detected. A crack candidate calculation processing device that integrates the crack candidates detected from the image data and outputs the result as an inspection result;
A detection result output device for outputting the inspection result output from the crack candidate calculation processing device to a user;
A control apparatus for an inspection apparatus, comprising: The crack candidate calculation processing device includes an image correction determination unit,
The image correction determination unit calculates correction amount data for increasing a ratio between the luminance of the image data in the same measurement region as the crack candidate detected from the shape data and the luminance of the other region. The control device for an inspection apparatus according to claim 1. The crack candidate calculation processing device includes a surface identification unit that calculates surface information and identifies a surface from the shape data,
When there is no crack candidate based on the shape data, the image correction determination unit recognizes the correction amount data of the brightness that is recognized as the same surface by the surface identification unit and has the closest distance, and the crack candidate does not exist. The inspection apparatus control apparatus according to claim 2, wherein the inspection apparatus control apparatus is used as area correction amount data.   The image correction determination unit calculates a distance between the inspection object and the sensor unit based on the shape data, and calculates luminance correction amount data of the image data according to the calculated distance. The control apparatus for an inspection apparatus according to claim 3.   The image correction determination unit calculates the inclination of the inspection object and the sensor unit based on the shape data, and calculates the luminance correction amount data of the image data according to the calculated inclination. The control device for an inspection device according to claim 3, wherein the control device is an inspection device. The inspection apparatus main body further includes illumination for irradiating the inspection object with light,
The crack candidate calculation processing device comprises:
A position and orientation calibration unit that acquires a relative position and orientation of the sensor unit and the illumination and performs calibration;
An illumination control unit that controls the illumination to adjust the distribution of luminance of the image data;
The inspection apparatus control device according to claim 4, further comprising: The crack candidate calculation processing device comprises:
Based on the correction amount data calculated by the image correction determination unit, the position and orientation of the sensor unit that increases the ratio between the luminance of the crack region and the luminance of the other region are calculated, and based on the calculation result The sensor unit position and orientation control unit for controlling the sensor unit
The inspection apparatus control device according to claim 4, further comprising: The crack candidate calculation processing device comprises:
Based on the correction amount data calculated by the image correction determination unit, a crack determination threshold value processing unit that changes a crack determination threshold value of a feature amount for determining a crack,
The inspection apparatus control device according to claim 4, further comprising: A sensor unit for acquiring image data and shape data from the surface of the inspection object;
A moving mechanism that moves and rotates the sensor unit around the inspection object;
A control unit for controlling the sensor unit and the moving mechanism;
A crack candidate is detected from the shape data, the image data is corrected based on the shape data, a crack candidate is detected from the corrected image data, and a crack candidate detected from the shape data is detected. And a crack candidate calculation processing device that integrates the crack candidates detected from the image data and outputs them as inspection results;
A detection result output device for outputting the inspection result output from the crack candidate calculation processing device to a user;
An inspection apparatus comprising: A sensor unit for acquiring image data and shape data from the surface of the inspection object;
A moving mechanism that moves and rotates the sensor unit around the inspection object;
A control unit for controlling the sensor unit and the moving mechanism;
An inspection apparatus control method for controlling an inspection apparatus main body comprising:
Via the control unit, controlling the moving mechanism and the sensor unit, obtaining the image data and the shape data of the surface of the inspection object;
A crack candidate is detected from the shape data, the image data is corrected based on the shape data, a crack candidate is detected from the corrected image data, and a crack candidate detected from the shape data is detected. And integrating the crack candidates detected from the image data and outputting as inspection results;
Outputting the output inspection result to a user;
An inspection apparatus control method comprising: A sensor unit for acquiring image data and shape data from the surface of the inspection object;
A moving mechanism that moves and rotates the sensor unit around the inspection object;
A control unit for controlling the sensor unit and the moving mechanism;
A program that is read and executed by a computer that controls the inspection apparatus main body, and the program is stored in the computer,
Via the control unit, controlling the moving mechanism and the sensor unit, obtaining the image data and the shape data of the surface of the inspection object;
A crack candidate is detected from the shape data, the image data is corrected based on the shape data, a crack candidate is detected from the corrected image data, and a crack candidate detected from the shape data is detected. And integrating the crack candidates detected from the image data and outputting as inspection results;
Outputting the output inspection result to a user;
A program characterized by having executed.