JPH116712A - Image processor for inspection - Google Patents

Abstract

PROBLEM TO BE SOLVED: To build a system at a lower cost without requiring a specialized device, by fetching the characteristic of an inspected object as image data to evaluate and judge the data using an image processing thereof. SOLUTION: An infrared camera 10 photographs a temperature distribution of a paper carbon 1. In an offline type system, the taken image is sent to a video to be stored once and thereafter, a reproduced image data is supplied to an image processing device 14, which measures the temperature of the paper carbon 1. The luminance is measured at measuring points of the paper carton 1 referring the image data in a measuring frame image. Luminance value data obtained at the measuring points are sent to a data evaluating/judging part to judge whether the obtained luminance value thus obtained is within a predetermined luminance range or not. The luminance value is proportional to the temperature of a stuck part 2a of the paper carton 1 roasted with a burner 5 to allow judging of whether heating by the burner 5 is proper or not. The data evaluating/judging part detects the abnormality of the luminance value to perform a processing of emitting an alarm, automatically stopping a line or others.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention belongs to the technical field of image processing, and more particularly to image processing performed for inspection purposes in a product manufacturing line.

[0002]

2. Description of the Related Art In quality inspection of a product, the shape and other characteristics of the product are captured and evaluated as image data. For example, in the production of paper containers such as milk cartons, since the ends of paper materials are bonded to form a three-dimensional shape, it is necessary to inspect whether the work is performed properly. . In this example, bonding is performed by heating the end of the paper container to melt the resin on the surface. For this reason, the temperature is detected by using an infrared camera to check whether or not the bonded portion of the paper container is appropriately heated.

[0003]

However, in the case of a liquid container or the like, a very large number of products are manufactured in a short time.
It is required to process and evaluate the measured data at high speed. Further, if a system for storing and processing image data captured by an infrared camera is to be constructed, an enormous capacity of memory is required, and the apparatus itself must be expensive.

[0004] The present invention has been made in view of the above points, and has been developed in image processing applied to product inspection and the like.
It is an object of the present invention to provide an image processing apparatus which can be constructed at low cost and can analyze data at high speed.

[0005]

According to a first aspect of the present invention, there is provided an image processing apparatus for inspecting an image of an image showing characteristics of an inspection object conveyed in a production line. Means for capturing, a means for detecting the position of the inspection object in the captured image data, and a means for determining a measurement point at which the characteristic of the inspection object is to be measured based on the detected position. Means for measuring characteristics of the inspection object at the measurement point based on the captured image data, and means for determining pass / fail of the object based on the measured characteristics. To be configured.

According to the image processing apparatus for inspection configured as described above, the image data of the image indicating the characteristic of the inspection object conveyed in the manufacturing line is taken in, and the inspection object in the image data is taken. The position is detected. Next, based on the detected position of the inspection object, the position of the measurement point, at which the characteristic should be measured, is determined from the image data, and the characteristic of the inspection object is measured from the image data at the measurement point. I do. Based on the image data indicating the characteristics of the inspection object measured in this way, the quality of the inspection object is determined.

According to a second aspect of the present invention, in the inspection image processing apparatus according to the first aspect, the capturing means collects image data of images of a plurality of portions corresponding to one inspection object. Wherein the detecting means includes means for detecting an image of a first part of the plurality of parts, and means for detecting a tip position of the inspection object included in the detected image of the first part. , The determining means comprises:
A plurality of the measurement points are determined based on the tip position.

According to the inspection image processing apparatus configured as described above, the capturing means captures an image of each part in a state where the inspection object is divided into a plurality of parts. The detecting means detects an image of a first part among the images of the plurality of parts, and further detects a tip position of the inspection object included in the image of the first part. The determining means determines a measurement point based on the tip position of the inspection object detected in this way.

According to a third aspect of the present invention, in the inspection image processing apparatus according to the first aspect, the fetching means fetches images of a plurality of portions corresponding to one inspection object as a group. When the determined measurement point is located between the images of the plurality of continuous parts, the measurement point is located in the image of the closer part of the images of the parts located before and after the measurement point. Configure to change.

According to the inspection image processing apparatus configured as described above, the capturing means captures an image of each part in a state where the inspection object is divided into a plurality of parts. The determining means, when the determined measurement point is located between the images of the plurality of continuous portions, performs the measurement in an image of a portion in a closer direction among the images of the portions located before and after the measurement point. Change the point.

[0011]

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings. In the embodiment described below, in a process of bending a paper carton such as a milk pack and bonding both ends, a case where an inspection as to whether heating of both ends is appropriate is performed by temperature detection with an infrared camera will be described.

First, FIG. 1 shows a method of bonding a paper carton to be inspected. In FIG. 1, a paper carton 1
Is supplied to the bonding step in an open state as shown on the left side in the figure. On the left side of FIG. 1, the paper carton 1 has two ends 2, and a bonding portion 2 a having a predetermined width is formed inward from both ends 2. Three folds 3 are formed near the center of the paper carton 1. The paper carton 1 is a laminate formed of layers of paper, PET, PE, and the like. The resin is melted by heating the surface layer, and is bonded by bonding the portions. As shown in the center of FIG. 1, the left and right end portions 2 of the paper carton 1 are bent along two outer creases 3, and the lower side of the bonding portion 2a on the front side and the bonding portion 2a on the back side are bent. The upper side is bonded as shown in the right side of FIG.

FIG. 2 shows a bonding portion 2a of the paper carton 1.
The method of heating is shown. In FIG. 2, a paper carton 1 is conveyed from left to right by a conveyor (not shown).
A pair of burners 5 are arranged in the transport path of the paper carton 1. The burner 5 on the near side in the figure heats the bonding portion 2a of the paper carton 1 from below. On the other hand, the back side burner 5 heats the bonding part 2a of the paper carton 1 from above. An infrared camera 10 is arranged above the transport path of the paper carton 1 and images the heated paper carton 1 from above. The image acquired by the infrared camera 10 is an image having a luminance corresponding to the temperature distribution of the paper carton 1 as the target.

FIG. 3 shows a processing flow of the image processing system according to the present invention. FIG. 3A shows the configuration of an off-line system, which is composed of an infrared camera 10, a video 12, an image processor 14, and a data evaluation / judgment unit 16.
Having. In the offline type system, the infrared camera 10
Is temporarily stored in the video 12, and the subsequent processing is performed.

FIG. 3B shows the configuration of an in-line system. This system has an infrared camera 10, an image processor 14, and a data evaluation / determination unit 16. An in-line system is a system that immediately performs image processing without storing image data in video. Whether to use the offline type system or the inline type system basically depends on the processing capability (speed) of the image processing machine 14. However, even when the capacity of the image processing machine 14 is sufficient, when it is necessary to store the captured image on a video tape, an offline type system is adopted.
In addition, high-speed processing can be performed.

Next, the operation of the image processing system according to the present invention will be described.

First, as shown in FIG. 2, a temperature distribution of a paper carton 1 is photographed by an infrared camera 10 arranged on a production line. In the off-line system (FIG. 3A), the captured image is sent to the video 12 and temporarily stored. Thereafter, reproduction of the video 12 is controlled by a control signal Sc supplied from the image processor 14, and the reproduced image data is supplied to the image processor 14. on the other hand,
In the in-line system (FIG. 3B), image data output from the infrared camera 10 is directly transmitted to the image processor 14.
Supplied to

The image processor 14 measures the temperature of the paper carton 1 based on the supplied image data. less than,
The image data processing in the image processor 14 will be described with reference to FIGS.

Referring to FIG. 4, first, the image processor 14
Determines various constants as pre-processing (step S1). Specifically, a reference value for a binarization process for determining the presence or absence of an object, an appropriate luminance range indicating that the luminance of the measurement point is appropriate, the number of measurement points of the object, an interval, and the like are determined. .

Next, the supplied image data is fetched (step S3), and a field image division process is performed (step S5). The signal output from the infrared camera 10 is a video signal, and one frame image is composed of two field images. Here, paper carton 1
Is set sufficiently faster than the scanning speed of the video signal, so that two consecutive field images indicate the temporally continuous movement of the paper carton 1. FIG. 6 shows the outline. When the photographed image is viewed in field image units, as shown in the right side of FIG.
The moving state of is shown. Note that the vertically long part in the image in FIG. 6 corresponds to the bonded part 2 a of the paper carton 1.
Since this portion is heated by the burner 5 and has a high temperature, the brightness is higher than that of the surrounding portion.

Since the signal output from the infrared camera 10 is a frame image unit, as shown on the left side of FIG.
Two continuous field images are output as one frame image. Therefore, processing for dividing this frame image into field images is required. FIG. 7 shows a method of division. Since one frame image is configured by alternately arranging odd (ODD) fields and even (EVEN) fields in the vertical direction, the odd field is doubled in the vertical direction to generate one frame image. Similarly, the even field is doubled in the vertical direction to generate the next frame image (step S7). In this way, a temporally continuous frame image (see the right side in FIG. 6) is generated and stored in the frame memory in the image processor 14 (step S9). Hereinafter, the frame image thus obtained is referred to as a “measurement frame image”.

Next, preprocessing of the measurement frame image is performed (step S11). In the pre-processing step, first, the measurement frame image is binarized. As described above, the portion heated by the burner of the paper carton 1 (the bonding portion 2
Since a) is in a high temperature state, the brightness value is high in the image captured by the infrared camera 10. Therefore, the measurement frame image is binarized using the predetermined reference value determined in step S1, and the high-temperature portion is detected.

Next, expansion / compression processing is performed in order to remove noise components of the binarized measurement frame image. The expansion process is a process of changing all pixels surrounding the pixel detected as positive (1) by the binarization process to positive (1). Subsequently, a compression process is performed. The compression process is a process of setting all pixels surrounding the negative (0) pixel of the image after the expansion process to be negative (0). This expansion / compression processing removes noise components from the measurement frame image.

Next, a labeling process is performed on the measurement frame image for which the preprocessing has been completed (step S13).
The labeling process is a process of detecting an object existing in the measurement frame image, and specifically, finds diagonal coordinates when the object is surrounded by a rectangle. FIG. 8 shows a labeling processing method.

First, as shown in FIG. 8A, the origin is located at the upper left of the frame image, and the horizontal scanning direction is the X axis and the vertical scanning direction is the Y axis. Further, as shown in FIG. 8B, the coordinates of the upper left vertex of one measurement frame image are represented by (X
_S, and Y _S), the coordinates of the lower right apex and (X _E, Y _E). Scanning the measurement frame image in the X-axis direction or the Y-axis direction and examining each pixel data reveals that a pixel corresponding to a heated portion of the paper carton 1 is "1" due to binarization. The shape of the heated part (object) can be understood. Therefore, the maximum value and the minimum value of the X coordinate and the maximum value and the minimum value of the Y coordinate of the rectangle surrounding the shape are detected.

For example, when an object having the shape shown in FIG. 8A (the bonding portion 2a) is included in the measurement frame image, the minimum point of the rectangle surrounding the object (the point where both the X coordinate and the Y coordinate are minimum) ) The coordinates of A are A (Xa, Ya),
The maximum point (point at which both the X coordinate and the Y coordinate are maximum) B is (X
b, Yb). Note that Ya <Yb. As described above, the target in each measurement frame image is detected by the labeling process, and the coordinates of the minimum point A and the maximum point B indicating the diagonal points of the rectangle surrounding the target are obtained.

Thus, whether or not an object exists in the measurement frame image, that is, the maximum point A
Then, it is determined whether or not the coordinates of the minimum point B have been obtained (step S15). When the maximum point A and the minimum point B cannot be obtained by the labeling process (step S15: No), it is determined that the object, that is, the heated bonded portion 2a of the paper carton 1 does not exist in the measurement frame image. Therefore, the process returns to step S11, and the same processing is performed for the next measurement frame image. On the other hand, when the maximum point A and the minimum point B are obtained (Step S15: Yes), since the measurement frame image includes the target, it is determined whether the measurement point has already been determined. (Step S1
7). The measurement point is a point at which the luminance is actually measured and evaluated. When the measurement point has already been determined (step S1
7: Yes), the process directly proceeds to the step S21. On the other hand, when the measurement point has not been determined yet (step S1)
7: No), a measurement point determination routine (step S19)
Proceed to.

Before describing the measurement point determination routine, the types of measurement frame images and measurement points will be described with reference to FIG. Since the paper carton 1 as an object has a certain length, the temperature distribution measured by the infrared camera 10 may extend over a plurality of measurement frame images. FIG.
In the example, the temperature distribution measurement data of one paper carton 1 is 4
This shows a case in which the measurement frame image is straddled. here,
The measurement frame images can be classified into four types of images according to the shape of the object included therein. The first type is an image including the tip of the object in the measurement frame image as shown in FIG. 10 (A).
Call. The second type is an image that does not include the front end or the rear end of the target object, as shown in FIG. 10B, and is hereinafter referred to as a “middle image”. The third type is an image including the rear end of the object as shown in FIG. 10C, and is hereinafter referred to as a “bottom image”. The fourth type is an image including both the front end portion and the rear end portion of the object as shown in FIG. 10D, and is hereinafter referred to as an “overall image”. In this way, the generated measurement frame image can be classified into any of the above. Normally, a paper carton 1 of the same size
Is transported at the same transport speed, and the infrared camera 10 performs the measurement. Therefore, depending on the relationship between the length of the object and the transport speed, either the middle image or the entire image does not exist. For example, when the target object is short or the transport speed is very fast, the middle image does not exist. On the other hand, when the target object is long or the transport speed is slow, the entire image does not exist. In the present embodiment, as shown in FIG. 9, one object covers a plurality of measurement frame images. Therefore, it is assumed that the entire image does not exist. The four types of images are based on the coordinates of the minimum point A and the maximum point B obtained in the labeling process, as shown in FIG.
It can be determined by the equations shown in (A) to (D).

Next, the measurement points will be described. Measurement points and
Measures the luminance of the target part in the measurement frame image.
Point. In this embodiment, the method of transporting the paper carton 1
In the direction (see FIG. 2)
Point. " ), Intermediate part (hereinafter referred to as “intermediate measurement point”)
U. ) And the rear end (hereinafter referred to as the "rear end measurement point")
U. 3) are determined. The number of measurement points is
Not limited to this, it can be determined according to the shape of the object
Wear. These three measurement points are each in a paper carton 1
Is defined as a point at a predetermined distance from the tip of
You. Referring to FIG. 9, the length of the paper carton 1 is now L._allWhen
Then, the top measurement point is t from the tip of the paper carton 1 (example
For example, L_all/ 10) distance point, middle measurement
The point is L from the tip_{a ll}/ 2 distance point, bottom measurement point after
It is defined as a point at a distance of t from the end. Therefore,
The y-coordinate of the tip point of the object existing in the image
For example, the y coordinate of each measurement point can be obtained by calculation.
You. In addition, L which is the total length of the paper carton 1 _allThe value of
Use the standard length of the paper carton 1 supplied to the
Can be used. Also, instead obtain by measurement
The distance between the front and rear ends of the object
No.

Next, a measurement point determination routine will be described. FIG. 5 shows the measurement point determination routine. First, with respect to the measurement frame image, it is checked whether or not the object included in the measurement frame image has an appropriate length and width. Since the width and length of the bonded portion 2a of the supplied paper carton 1 are known from the standard of the paper carton 1, if an object having a size far from the size is detected, an imaging error of the infrared camera 10 is detected. Others are possible. Therefore, it is confirmed whether or not the size of the object included in the measurement frame image is within a range in which it can be determined that the object is the bonded portion 2a of the paper carton 1 (step S31). Specifically, the width of the object,
It is determined whether or not the length is within a predetermined reference value.

Next, it is determined whether or not the measurement frame image is a top image (step S33). This determination is made from the coordinates of the minimum point A, the coordinates of the maximum point B, and the vertex coordinates Y _S and Y _{E of} the frame image, as shown in FIG. 10A. If it is the top image (step S33: Yes), the y coordinate of the tip point of the object included in the measurement frame image is obtained (step S35). This can use the y-coordinate of the minimum point A usually determined in the labeling step.

Next, based on the y-coordinate of the front end point, the y-coordinate of the front end measurement point, the intermediate measurement point, and the rear end measurement point is calculated (step S37). As described with reference to FIG. 9, since the positions of these three measurement points are determined as distances from the tip, if the y coordinate of the tip is determined, the y coordinate of each measurement point is determined. You. According to the above example, it is as follows.

Y coordinate of front end measurement point = y coordinate of front end point + t (= L _all / 10) y coordinate of intermediate measurement point = y coordinate of front end point + L _all / 2 y coordinate of rear end measurement point = front end Y coordinate of point + L _all −t (= L _all / 10) Next, the x coordinate of each measurement point thus determined is determined (step S39). In this method, the x-coordinate is changed at the determined y-coordinate position, the average luminance of eight pixels near the pixel is calculated, and the x-coordinate at which the average luminance is maximized is determined as the x-coordinate of the measurement point. The measurement point is preferably near the center of the object. When the average luminance of the neighboring pixels is calculated by moving the x coordinate as described above, the average luminance becomes low near the end of the target object. Therefore, the point at which the maximum average luminance is obtained is estimated to be near the center of the object, and the measurement point is determined at that point.

As described above, when the coordinates of the measurement points of the front end measurement point, the intermediate measurement point, and the rear end measurement point are determined, the measurement point determination routine ends and returns to the main routine. If it is determined in step S31 that the target object is not an appropriate size, and if it is determined in step S33 that the target object is not a top image, the process immediately returns to the main routine.

When the measurement points are determined, it is checked whether or not the determined measurement points are included in the current measurement frame image (step S21). This can be determined from the relationship between the coordinates of the maximum and minimum points obtained by labeling and the coordinates of each measurement point. This will be described with reference to FIG. FIG. 11 shows a case where the current measurement frame image is the top image.

In FIG. 11A, the front end measurement point T (x
_t , y _t ) are included in the range defined by the maximum point and the minimum point. Specifically, Xa ≦ x _t ≦ Xb, and, in the case of Ya ≦ y _t ≦ Yb would forward measuring point in the top image exists. On the other hand, when this condition is not satisfied, it is determined that the measurement point does not exist in the top image in principle. However, as shown in FIG.
It is possible that the measurement points belong to a blank between fields or between frames. As shown in FIG. 9, the shooting of the object is performed using a video signal. Therefore, there is a vertical blanking period between each field image, and no image can be shot during that period. Therefore, depending on the timing, the measurement point may be located between two consecutive field images. Therefore, in this case, the measurement point must be determined in one of the previous and next measurement frame images. FIG.
(B) shows the processing in this case. Referring to FIG. 11B, the coordinates of the maximum point B of the measurement frame (first frame) are known by labeling. The length _Lb (y coordinate) of the blank between fields (or between frames) is determined to be a fixed value based on the transport speed of the object and the vertical blanking period of the video signal. Therefore, the y coordinate of the start point of the next measurement frame is also known. Therefore, it is determined which position in the blank the y coordinate of the measurement point is closer to, specifically, which of the first frame and the second frame is closer by comparing the coordinate values. That is, as shown in FIG. 11B, Xa ≦ _xt ≦ X
b, and Y _E <y _t ≦ Y _E + Lb / 2 <Y _E + Lb, the measurement point is set to T (x _t , Y _E ) in the frame (the first frame in FIG. 11B). ). On the other hand, Xa ≦
xt ≦ Xb, _{and, Y E <Y E + Lb} / 2 <y t ≦ Y E + Lb
If it is, since the measurement points are closer to the next measurement frame image, T in the measurement point of the next frame (the second frame in FIG. _{11 (B)) (x t} , Y S (= Y _C )). The measurement points thus determined are stored in a memory (not shown).

The case where the current frame is the top image has been described above as an example. For frames other than the top image, the coordinates of the minimum and maximum points similarly obtained by labeling are determined in step S17. From the measured measurement point coordinates, it can be confirmed whether or not the measurement point exists in the measurement frame image. In practice, it is preferable to perform the confirmation processing on each of the measurement frame images at the front end measurement point, the intermediate measurement point, and the rear end measurement point.

After the confirmation of the measurement points is completed, the luminance of the coordinates of each confirmed measurement point is measured with reference to the pixel data in the actual measurement frame image (step S23). Since the pixels in each measurement frame image are represented by a plurality of bits of luminance values, the luminance values for the coordinates of each measurement point are obtained. When the measurement of the brightness is completed, the process returns to step S9, and the same process is performed on the next measurement frame image.

The luminance value data for each measurement point thus obtained is sent to the data evaluation / judgment unit 16 (FIG. 3).
reference). The data evaluation / determination unit 16 determines whether the obtained luminance value is within a predetermined luminance range. As described above, since the luminance value in the measurement frame image is proportional to the temperature of the portion of the bonding portion 2a of the paper carton 1 that has been brushed by the burner, by determining this luminance value,
It is determined whether heating by the burner is appropriate.
If the brightness value is smaller than the lower limit of the predetermined brightness range,
It is determined that the heating of the bonding portion 2a is insufficient.
This may be caused by insufficient heating power of the burner, too high a speed of transporting the paper container, too large a distance between the burner and the paper container, and the like. On the other hand, if the luminance value is larger than the upper limit of the predetermined luminance range, it is determined that the bonding section 2a is excessively heated. It is considered that this is caused by an excessive heating power of the burner and an excessively low conveying speed. In addition, when the brightness value varies among a plurality of measurement points of the same paper container, the paper container is lifted up during transportation due to a wind pressure generated at the time of transportation (for example, the upstream side of the transportation path is downstream. It is also conceivable that the heating by the burner is not uniform. Data evaluation / judgment unit 16
Performs such processing as detecting an abnormal state of the luminance value, issuing an alarm, and automatically stopping the line.

As described above, according to the image processing apparatus and method of the present invention, characteristics such as the temperature of an object are acquired as image data, and the quality of the state of the object is determined by the data processing. Even when the object is conveyed at high speed, quick detection and determination can be performed.
Also, since data is collected using a normal video signal,
No special device is required for data recording and processing, and the system can be configured at relatively low cost.

In the above-described embodiment, an example has been described in which a temperature distribution image of a heated portion of a paper carton is taken and image processing is performed, but the application of the present invention is not limited to this. That is, other than the paper carton, when a heating process is involved in the manufacturing process, the same process can be performed using an infrared camera. Further, the image to be acquired is not limited to the temperature distribution image photographed by the infrared camera, and may be applied to a system in which the print surface of a printed matter is photographed by a normal camera and the print density and the position are inspected. Is also possible.

[0042]

As described above, according to the first aspect of the present invention, the characteristics of the object to be inspected are fetched as image data and evaluated and determined by image processing, so that the object is conveyed at high speed. Inspection of the inspection object becomes possible.
Further, since a normal image signal is used, a special dedicated device is not required, and the system can be configured at low cost.

According to the second aspect of the present invention, the object to be inspected is fetched as images of a plurality of portions and processed, so that inspection can be performed regardless of the shape and size of the object to be inspected.

According to the third aspect of the present invention, when the determined measurement point is located between a plurality of continuous images, the measurement point is changed in a closer image, so that the original measurement point Characteristics with similar values can be measured reliably.

[Brief description of the drawings]

FIG. 1 is a diagram showing a method of attaching a paper carton as an object of image processing according to the present invention.

FIG. 2 is a diagram illustrating a method of heating a bonding portion of a paper carton.

FIG. 3 is a diagram showing a flow of processing from measurement of a temperature of a paper carton to evaluation.

FIG. 4 is a flowchart showing a process in the image processing machine shown in FIG. 3;

FIG. 5 is a flowchart showing processing of a measurement point determination routine shown in FIG.

FIG. 6 is a diagram illustrating the field image division processing shown in FIG.

FIG. 7 is another diagram illustrating the field image division processing shown in FIG.

FIG. 8 is a diagram illustrating the labeling process shown in FIG.

FIG. 9 is a diagram illustrating measurement points.

FIG. 10 is a diagram illustrating types of measurement frame images.

FIG. 11 is a diagram for explaining a measurement point confirmation process shown in FIG. 4;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Paper carton 2 ... End part 2a ... Lamination part 5 ... Burner 10 ... Infrared camera 12 ... Video 14 ... Image processing machine 16 ... Data evaluation and judgment part

Claims (3)

[Claims] 1. A means for capturing image data of an image indicating characteristics of an inspection object transported in a production line; a means for detecting a position of the inspection object in the captured image data; Means for determining a measurement point at which the characteristic of the inspection object is to be measured based on the position; means for measuring the characteristic of the inspection object at the measurement point based on the captured image data; Means for determining pass / fail of the object based on the measured characteristics. 2. The image capturing means collects image data of images of a plurality of parts corresponding to one inspection object, and the detecting means extracts an image of a first part of the plurality of parts. Means for detecting, and means for detecting a tip position of the inspection object included in the detected image of the first part, wherein the determining means determines a plurality of the measurement points based on the tip position. The inspection image processing apparatus according to claim 1, wherein the determination is made. 3. The image capturing means collectively captures images of a plurality of portions corresponding to one inspection object, and the determining means determines an interval between images of the plurality of portions in which determined measurement points are continuous. The inspection image processing apparatus according to claim 1, wherein when the position is located in the image, the measurement point is changed in an image of a closer portion of the images of the portion located before and after the measurement point.