Quality evaluation apparatus, quality evaluation system, and quality evaluation method

Abstract

The system efficiently and accurately evaluates the quality of multiple workpieces being transported in multiple rows on a conveyor belt. [Solution] This application relates to a quality evaluation device for evaluating the quality of multiple workpieces transported in multiple rows on a conveyor belt. The quality evaluation device creates multiple image frames from image data continuously captured using a line camera installed so as to intersect the transport direction of the conveyor belt. A composite image frame is synthesized based on a first image frame and a second image frame that are consecutive among the multiple image frames, and abnormal parts are detected based on the composite image frame.

Description

Technical Field

[0001] The present disclosure relates to a quality evaluation apparatus, a quality evaluation system, and a quality evaluation method.

Background Art

[0002] For example, in the quality evaluation of a workpiece such as food shipped as a product, it is required to determine whether there are foreign substances or abnormal parts on its outer surface. Conventionally, such quality evaluation has been performed manually by workers, but there are limitations when performing a full inspection of mass-produced products, etc., so automation has been promoted by analyzing an image obtained by imaging the workpiece using a camera.

[0003] Regarding such an automatic quality evaluation apparatus using image analysis technology, for example, there are Patent Documents 1 and 2. In Patent Document 1, food is handled as a workpiece, and a technique is disclosed in which, by performing binarization processing on an image obtained by imaging the food, it is determined that there are foreign substances mixed in the food in a range exceeding a threshold value among adjacent pixels. Further, in Patent Document 2, although it is not a technique for evaluating foreign substances or abnormal parts, by performing binarization processing on a processed image obtained as the difference between an image using visible light and an image using infrared light for imaging the workpiece, an image processing technique for distinguishing and extracting the regions of the fat part and the bone part of meat is disclosed.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] In processing lines that handle a relatively large volume of workpieces, these workpieces are transported in multiple rows on a conveyor belt to process them efficiently. When introducing a quality evaluation device using the image analysis technology described above into such a processing line, it is conceivable to use a line camera as the camera to continuously capture images of the workpieces being transported on the conveyor belt, and then evaluate the quality of each workpiece contained in the image frame extracted from the captured image data to check for any abnormalities. In this case, the continuous image data captured by the line camera is divided into multiple image frames, but there is a risk that the quality evaluation accuracy will decrease for workpieces that fall on the boundaries between consecutive image frames. In other words, since workpieces that fall on the boundaries between image frames are only partially included in each image frame (the entire workpiece is not included in a single image frame), if there are abnormalities in the area not included in the image frame, false detections such as overlooking the abnormality may occur.

[0006] At least one embodiment of this disclosure has been made in view of the above circumstances and aims to provide a quality evaluation device, a quality evaluation system, and a quality evaluation method that can efficiently and accurately evaluate the quality of multiple workpieces transported in multiple rows on a conveyor belt. [Means for solving the problem]

[0007] A quality evaluation apparatus according to at least one embodiment of this disclosure solves the above problem, A quality evaluation device for evaluating the quality of multiple workpieces transported in multiple rows on a conveyor belt, An image frame creation unit for creating multiple image frames from image data continuously captured using a line camera installed so as to intersect the conveying direction of the conveyor, An image frame compositing unit for compositing a composite image frame based on a first image frame and a second image frame that are consecutive among the plurality of image frames, The system includes a detection unit that detects abnormal parts from the plurality of workpieces based on the composite image frame.

[0008] A quality evaluation method according to at least one embodiment of this disclosure solves the above problems. A quality evaluation method for evaluating the quality of multiple workpieces transported in multiple rows on a conveyor belt, A step of creating multiple image frames from image data continuously captured using a line camera installed so as to intersect the conveying direction of the conveying conveyor, A step of synthesizing a composite image frame based on a first image frame and a second image frame that are consecutive among the plurality of image frames, A step of detecting abnormal parts from the plurality of workpieces based on the composite image frame, It is equipped with. [Effects of the Invention]

[0009] According to at least one embodiment of this disclosure, it is possible to provide a quality evaluation device, a quality evaluation system, and a quality evaluation method that can efficiently and accurately evaluate the quality of multiple workpieces being transported in multiple rows on a conveyor belt. [Brief explanation of the drawing]

[0010] [Figure 1] This is a schematic diagram of a quality evaluation system according to one embodiment. [Figure 2] Figure 1 is a schematic diagram showing a workpiece being transported on a conveyor belt, viewed from above. [Figure 3] Figure 1 is a block diagram of the quality evaluation apparatus. [Figure 4] Figure 3 shows an example of image data acquired by the image data creation unit. [Figure 5A] This is an example of the first image frame GF1. [Figure 5B] This is an example of the second image frame GF2. [Figure 5C] This is an example of a composite image frame created from the first image frame in Figure 5A and the second image frame in Figure 5B. [Figure 6]It is a flowchart showing a quality evaluation method that can be implemented by the quality evaluation apparatus in FIG. 3. [Figure 7] It is a flowchart showing a method of synthesizing a composite image frame in step S102 in FIG. 6 and a method of detecting an abnormal part in S103. [Figure 8] It is a diagram showing a specific example of a workpiece in the composite image frame shown in FIG. 5C. [Figure 9] It is a flowchart showing a method of detecting an abnormal part in step S205 in FIG. 7. [Figure 10] It is a flowchart showing a method of creating position information by the position information creation unit in FIG. 3. [Figure 11] It is an example of an image frame corresponding to the position information created in FIG. 10.

Embodiments for Carrying Out the Invention

[0011] Hereinafter, some embodiments of the present invention will be described with reference to the accompanying drawings. However, the configurations described as embodiments or shown in the drawings are not intended to limit the scope of the present invention, but are merely illustrative examples.

[0012] First, a quality evaluation system 1 including a quality evaluation apparatus 100 according to at least one embodiment of the present disclosure will be described. FIG. 1 is a schematic diagram of a quality evaluation system 1 according to one embodiment, and FIG. 2 is a schematic diagram showing a workpiece w being conveyed on the conveyor 2 in FIG. 1 from above.

[0013] The quality evaluation system 1 is a system for evaluating the quality of the workpiece w. In the quality evaluation system 1, the quality of the workpiece w is evaluated while being conveyed on the conveyor 2. The conveyor 2 includes a pair of pulleys 2a and 2b respectively arranged at both ends, and a belt 2c stretched across the pair of pulleys 2a and 2b. At least one of the pair of pulleys 2a and 2b is rotationally driven by a power source (not shown), so that a plurality of workpieces w placed on the belt 2c can be conveyed along the conveying direction a.

[0014] Furthermore, as shown in Figure 2, the conveyor belt 2 has a predetermined width along a direction that intersects (approximately perpendicular to) the conveying direction a (hereinafter referred to as "width direction b" as appropriate). The workpieces w are conveyed in multiple rows along the conveyor belt 2, which has a certain width along the width direction b, relative to the conveying direction a. This makes it possible to efficiently convey even a relatively large quantity of workpieces w using the conveyor belt 2.

[0015] The workpiece w subject to quality evaluation is not limited, but it could be food products, for example.

[0016] Returning to Figure 1, the quality evaluation system 1 comprises a light source 4, an imaging device 6, and a quality evaluation device 100. The light source 4 is positioned above the conveyor belt 2, enabling illumination light L1 to be shone onto the workpiece w placed on the belt 2c.

[0017] The imaging device 6 is configured to image a workpiece w being transported by the conveyor belt 2. The imaging device 6 is a line camera, and more specifically, a trilinear line sensor. The imaging device 6 obtains image data by receiving imaging light L2 obtained when illumination light L1 from the light source 4 is shone onto the workpiece w. In the configuration example shown in Figure 2, the imaging device 6 is positioned to receive imaging light L2 passing between multiple light sources 4 arranged along the belt 2c.

[0018] The quality evaluation device 100 is a device for evaluating the quality of multiple workpieces w being transported in multiple rows on a transport conveyor 2, based on image data acquired from the imaging device 6. Such a quality evaluation device 100 consists of, for example, a CPU (Central Processing Unit), RAM (Random Access Memory), ROM (Read Only Memory), and a computer-readable storage medium. A series of processes for realizing various functions are stored in the storage medium in the form of a program, for example, and the CPU reads this program into the RAM and executes information processing and calculations to realize the various functions. The program may be pre-installed in ROM or other storage media, provided in a state where it is stored in a computer-readable storage medium, or distributed via wired or wireless communication means. Computer-readable storage media include magnetic disks, magneto-optical disks, CD-ROMs, DVD-ROMs, semiconductor memory, etc.

[0019] Figure 3 is a block diagram of the quality evaluation device 100 shown in Figure 1. The quality evaluation device 100 includes an image data acquisition unit 102, an image frame creation unit 104, an image frame synthesis unit 106, a detection unit 108, a position information creation unit 110, and an output unit 112.

[0020] The image data creation unit 102 is configured to create image data G based on data acquired by continuous imaging by the imaging device 6, which is a line camera. As mentioned above, the imaging device 6 is a camera positioned along the direction (width direction b) that intersects the transport conveyor 2, and by continuously imaging the transport conveyor 2 from above, it images the workpiece w passing directly below the imaging device 6 together with the transport conveyor 2.

[0021] The image data G contains data for one line along the transport direction a, corresponding to the imaging time, and the obtained image data G includes the transport conveyor 2 and the workpiece w that passed through the imaging range of the imaging device 6 during the imaging time.

[0022] The image frame creation unit 104 is configured to create an image frame GF based on image data G created by the image data creation unit 102. Figure 4 shows an example of the creation of an image frame GF in the image frame creation unit 104 based on image data G created by the image data creation unit 102. The creation of the image frame GF is performed by accumulating image data G continuously captured by the imaging device 6. More specifically, multiple image frames GF are created by accumulating image data G created from data continuously captured by the imaging device 6 at predetermined lengths along the transport direction a.

[0023] In the following explanation, two consecutive image frames GF created based on the image data G will be referred to as "first image frame GF1" and "second image frame G2" (specifically, the first image frame GF1 is the image frame corresponding to time axis t1, and the second image frame GF2 is the image frame corresponding to time axis t2, which is consecutive to time axis t1. Note that the relationship between time axis t1 and time axis t2 is such that time axis t1 is in the past).

[0024] The image frame merging unit 106 acquires the first image frame GF1 and the second image frame GF2 from the image frame creation unit 104, and if there is a workpiece w in the lower end region 10 of the first image frame GF1 that overlaps with the boundary 12, it is configured to merge that lower end region 10 into the second image frame GF2. In the following description, the image frame merged in this way by the frame merging unit 106 will be referred to as the "merged image frame CGF".

[0025] Here, with reference to Figures 5A to 5C, the method for combining the composite image frame CGF by the image frame combining unit 106 will be explained. Figure 5A is an example of the first image frame GF1, Figure 5B is an example of the second image frame GF2, and Figure 5C is an example of the composite image frame CGF created from the first image frame GF1 of Figure 5A and the second image frame GF2 of Figure 5B.

[0026] Figure 5A shows an example of the first image frame GF1 corresponding to time axis t1. This first image frame GF1 contains several workpieces w, but there is a workpiece w1 that falls on the boundary 12 of the lower end region 10, shown by the dashed line. When there is a workpiece w1 that falls on the boundary 12 in the first image frame GF1, the lower end region 10 of the first image frame GF1 containing the workpiece w1 is combined with the second image frame GF2 (Figure 5B), which is the next corresponding time axis t2, to create a composite image frame CGF. In the composite image frame CGF, as shown in Figure 5C, the workpiece w1 that fell on the boundary 12 in the first image frame GF1 is interpolated, and its entirety is included. In other words, in the composite image frame CGF, the workpiece w1 that was partially missing in the rear end region 10 of the first image frame GF1 due to falling on the boundary 12 is interpolated through the combination with the second image frame GF2 so that its entirety is included. Furthermore, the length c of the rear end region 10 along the transport direction a may be set to correspond, for example, to the length of the long side of the workpiece w being transported on the transport conveyor 2.

[0027] Returning to Figure 3, the detection unit 108 is configured to detect abnormal parts of the workpiece w based on the composite image frame CGF synthesized by the image frame synthesis unit 106. The abnormal parts are the parts of the workpiece w other than the normal parts, and may be other substances mixed in the workpiece w (e.g., dust, dirt, etc.) or regions where the healthy workpiece w has changed (e.g., rotten parts, discolored parts, etc.). The detection unit 108 detects abnormal parts by performing image analysis on the composite image frame CGF and comparing the brightness value of each pixel contained in the composite image frame CGF with a predetermined threshold, but the specific detection method will be described later.

[0028] The position information creation unit 110 is configured to create position information Ip related to the location of the abnormal part detected by the detection unit 108. In this embodiment, as shown in Figure 2, the transport conveyor 2 includes a first transport area R1 to a third transport area R3 as a plurality of transport areas arranged along the width direction b. These first transport areas R1 to the third transport area R3 may be defined as areas that divide a single transport lane, as in this embodiment, or they may be configured to include three (or more) transport lanes corresponding to the first transport area R1 to the third transport area R3. The position information creation unit 110 creates position information Ip as information indicating which of these transport areas the workpiece w having the abnormal part detected by the detection unit 108 is located in.

[0029] The output unit 112 is configured to output the position information Ip created by the position information creation unit 110 to the downstream device 200 located downstream of the conveyor belt 2. The downstream device 200 can identify which conveying area of ​​the conveyor belt 2 the workpiece w in which an abnormal part has been detected will be transported based on the position information Ip obtained from the output unit 112, and can appropriately determine how to handle the workpiece w with the abnormal part.

[0030] In this embodiment, the downstream device 200 is a sorting device for sorting workpieces w transported by the conveyor belt 2 according to their quality. In this case, the downstream device 200 can distinguish and handle workpieces w with defects from normal workpieces w by identifying the transport area through which workpieces w with defects are flowing based on position information Ip (for example, workpieces w with defects can be picked up or guided to a dedicated discharge lane).

[0031] Furthermore, depending on the downstream device 200, which is a sorting device, the position information Ip may require information about the location of the abnormal part detected on the workpiece w on the conveyor belt 2 (for example, the position coordinates on the conveyor belt 2) instead of information about the transport area through which the workpiece w with the abnormal part described above flows. In this case, the output unit 112 may transmit information about the location of the abnormal part detected on the workpiece w on the conveyor belt 2 to the downstream device 200 as the position information Ip.

[0032] Next, a quality evaluation method that can be performed by the quality evaluation device 100 having the above configuration will be described. Figure 6 is a flowchart showing a quality evaluation method that can be performed by the quality evaluation device 100 shown in Figure 3.

[0033] First, the image data creation unit 102 creates image data G from data for one line continuously captured by the imaging device 6, which is a line camera (step S100). Next, the image frame creation unit 104 creates image frames GF (including at least a first image frame GF1 and a second image frame GF2) based on the image data G created in step S100 (step S101). These image frames GF are created by accumulating the image data G as described above.

[0034] Next, the image frame synthesis unit 106 acquires a first image frame GF1 and a second image frame GF2 corresponding to consecutive time axes t1 and t2, respectively, from the image frames created in step S101, and synthesizes a synthesized image frame CGF (step S102). Subsequently, the detection unit 108 detects abnormal parts based on the synthesized image frame CGF synthesized in step S102 (step S103). If an abnormal part is detected in step S103 (step S104: YES), the location information creation unit 110 creates location information Ip related to the location of the abnormal part (step S105). The location information Ip created in step S105 is output to the downstream device 200 by the output unit 112 (step S106).

[0035] Next, the method for synthesizing the composite image frame CGF in step S102 and the method for detecting abnormal areas in step S103 will be explained in detail. Figure 7 is a flowchart showing the method for synthesizing the composite image frame CGF in step S102 and the method for detecting abnormal areas in step S103 of Figure 6.

[0036] In Figure 7, the image frame synthesis unit 106 first obtains the first image frame GF1 and the second image frame GF2, which correspond to consecutive time axes t1 and t2, respectively, from the multiple image frames GF created in step S101 (step S200).

[0037] Next, the image frame synthesis unit 106 synthesizes a composite image frame GF2 based on the first image frame GF1 and the second image frame GF2 acquired in step S200 (step S201). As described above with reference to Figures 5A to 5C, in this example, the first image frame GF1 includes workpiece w1 that falls on the boundary 12 with the second image frame GF2. Therefore, the rear end region 10 of the first image frame GF1 is synthesized with the second image frame GF2 to create the composite image frame GF2.

[0038] Next, the composite image frame CGF synthesized in step S201 is subjected to image analysis to identify the workpiece w contained in the composite image frame CGF (step S202). The identification of the workpiece w in step S202 may be performed, for example, by comparing the brightness value of each pixel in the composite image frame CGF with a predetermined threshold to detect the workpiece w contained in the composite image frame CGF, or by inputting the composite image frame CGF into a prediction model constructed in advance using machine learning, and identifying the workpiece w contained in the composite image frame CGF as a prediction result of the learned model.

[0039] Figure 8 shows an example of workpiece w identification in the composite image frame CGF shown in Figure 5C. In this example, several workpieces w identified by image analysis of the composite image frame CGF are recognized by being surrounded by roughly square-shaped boxes 8. The shape of these boxes 8 can be arbitrary.

[0040] Next, the image frame synthesis unit 106 determines, based on the identification result of step S202, whether or not there is a workpiece w in the synthesized image frame CGF that falls on the upstream boundary 12a or the downstream boundary 12b with respect to the transport direction a (step S203). The determination in step S203 is performed, for example, by comparing the coordinates of the box 8 surrounding the workpiece w identified in step S202 with the coordinates corresponding to the upstream boundary 12a or the downstream boundary 12b of the synthesized image frame CGF. In the example shown in Figure 8, among the workpieces included in the synthesized image frame CGF, workpiece w3 falls on the upstream boundary 12a and workpiece w2 falls on the downstream boundary 12b. This allows for a suitable determination of whether or not each workpiece w included in the synthesized image frame CGF falls on the upstream boundary 12a or the downstream boundary 12b of the synthesized image frame CGF.

[0041] If it is determined that there is a workpiece w in the composite image frame CGF that is related to the upstream boundary 12a or the downstream boundary 12b (step S203: YES), the image frame compositing unit 106 excludes the workpiece w from the detection of abnormal parts (step S204), and the detection unit 108 performs abnormal part detection on the detection targets (i.e., workpiece w that is not related to the upstream boundary 12a or the downstream boundary 12b in the composite image frame CGF) (step S205). In other words, since the entire workpiece w related to the upstream boundary 12a or the downstream boundary 12b in the composite image frame CGF is not captured, abnormal part detection is not performed. In the example shown in Figure 8, among the workpiece w included in the composite image frame CGF, the workpiece w other than the workpiece w3 related to the upstream boundary 12a and the workpiece w2 related to the downstream boundary 12b are the targets for abnormal part detection.

[0042] On the other hand, if there is no workpiece w that falls within the upstream boundary 12a or downstream boundary 12b of the composite image frame CGF (step S203: NO), all workpieces w included in the composite image frame CGF are targeted for detection (step S205), and the detection unit 208 performs abnormality detection on the target workpieces w included in the composite image frame CGF (step S205).

[0043] Next, we will explain in detail the method for detecting abnormal parts in step S205. Figure 9 is a flowchart showing the method for detecting abnormal parts in step S205 of Figure 7.

[0044] First, the detection unit 108 acquires the image frame GF to be detected (step S300). The image frame GF acquired in step S300 is each image frame that was set as the target for detection in step S205.

[0045] Next, the image frame GF acquired in step S300 is divided into sub-image frames SGF (step S301). In the image frame GF acquired in step S300, each pixel is represented, for example, as a luminance value in the RGB color space. An image frame GF in which each pixel is represented in the RGB color space can be separated into individual color component images (R component image, G component image, and B component image) by image analysis. In step S301, each pixel of the image frame GF is divided into multiple sub-image frames SGF, which are individual color component images in the RGB color space.

[0046] In this embodiment, the example shown is that the image frame GF is divided into multiple sub-image frames SGF corresponding to each color component image in the RGB color space. However, the sub-image frames SGF may be color component images corresponding to other color spaces. For example, the sub-image frames SGF may be color component images (H component image, S component image, and V component image) corresponding to the HSV color space, which is another color space obtained by performing a predetermined conversion operation on the RGB color space.

[0047] Next, the detection unit 108 identifies the color of the abnormal part of the workpiece w that is to be detected (step S302). Based on the color of the abnormal part identified in step S302, it selects a sub-image frame SGF to be used for detecting the abnormal part from among the multiple sub-image frames SGF created by dividing the image frame GF in step S301 (step S303), and compares the brightness value of each pixel of the selected sub-image frame SGF with a threshold (step S304). The detection unit 108 then identifies the set of pixels in the sub-image frame SGF whose brightness value is less than the threshold as the abnormal part (step S305).

[0048] In step S303, a sub-image frame SGF corresponding to a component image that is complementary to the color of the anomaly identified in step S302 may be selected. For example, if the sub-image frame SGF is the R component image in the RGB color space, anomalies with colors close to red will be difficult to distinguish, so a sub-image frame SGF corresponding to the G component image or B component image may be selected. Similarly, if the sub-image frame SGF is the G component image in the RGB color space, anomalies with colors close to green will be difficult to distinguish, so a sub-image frame SGF corresponding to the R component image or B component image may be selected. Furthermore, if the sub-image frame SGF is the B component image in the RGB color space, anomalies with black, red, and green colors will be easier to distinguish. By selecting a sub-image frame SGF that is close to a component that is complementary to the color of the anomaly, anomalies with various colors can be detected with high accuracy.

[0049] Furthermore, in step S303, multiple sub-image frames SGF may be selected. In this case, for each of the selected sub-image frames SGF, the luminance value of each pixel may be compared with a threshold to detect abnormalities. This allows for accurate detection of abnormalities in the workpiece w, even if there are multiple abnormalities with different colors, based on the sub-image frame SGF corresponding to each abnormality.

[0050] Next, we will explain in detail how the location information creation unit 110 creates location information Ip based on the detection results of the detection unit 108. Figure 10 is a flowchart showing how the location information creation unit 110 creates location information Ip as shown in Figure 3, and Figure 11 is an example of an image frame GF corresponding to the location information Ip created in Figure 10.

[0051] First, the position information creation unit 110 acquires width information related to the conveyor belt 2 included in the image frame GF (step S400). In step S400, the unit performs image analysis on the image frame GF to identify the range of the conveyor belt 2 shown in the image frame GF, and acquires width information related to the conveyor belt 2 by counting the number of pixels along the width direction b for that range (in the example in Figure 11, this is 3000 pixels).

[0052] Next, the location information creation unit 110 acquires coordinate information indicating the position of the abnormal part detected by the detection unit 108 in the image frame GF (step S401). In step S401, the location information creation unit 110 identifies the area in the image frame GF where the abnormal part detected by the detection unit 108 is located, and acquires the representative coordinates (for example, the center coordinates of the abnormal part) in that area as coordinate information.

[0053] Next, the position information creation unit 110 acquires the coordinate range of each transport area of ​​the transport conveyor 2 (step S402). As mentioned above with reference to Figure 2, the transport conveyor 2 has a plurality of transport areas arranged along the width direction b, namely the first transport area R1 to the third transport area R3. In step S402, the coordinate ranges representing these first transport area R1 to the third transport area R3 are identified for the transport conveyor 2 as seen in the image frame GF. In the example in Figure 11, it is shown that the coordinate range of the first transport area R1 is 1 to 1000 pixels, the coordinate range of the second transport area R2 is 1001 to 2000 pixels, and the coordinate range of the third transport area R3 is 2001 to 3000 pixels.

[0054] Next, the position information creation unit 110 creates position information Ip indicating the transport area to which the workpiece w with the abnormal part is transported, based on the coordinate information of the abnormal part identified in step S401 and the coordinate range of each transport area identified in step S402 (step S403). In the example in Figure 11, the workpiece w with the abnormal part detected by the detection unit 108 is being transported in the second transport area R2 of the transport conveyor 2. Therefore, the position information creation unit 110 creates position information Ip indicating that the transport area to which the workpiece w with the abnormal part is transported is the second transport area R2 of the transport conveyor 2.

[0055] The position information Ip created by the position information creation unit 110 is output to the downstream device 200 by the output unit 112. The downstream device 200 receives the position information Ip output from the output unit 112 and performs control based on the position information Ip. In the case of the position information Ip created in the example in Figure 11, the downstream device 200 is controlled to sort the workpiece w with an abnormal part, as it is being transported from the second transport area R2, so that it distinguishes the workpiece w from healthy workpieces w transported from other transport areas (first transport area R1 and third transport area R3). Specifically, the downstream device 200 may be controlled to pick up the workpiece w transported from the second transport area R2 from the transport conveyor 2, or it may be controlled to discharge the workpiece w transported from the second transport area R2 from a dedicated discharge lane different from the first transport area R1 and third transport area R3.

[0056] Furthermore, in the downstream device 200, which is a sorting device, instead of information regarding the transport area through which the workpiece w with the aforementioned abnormal part flows, it may be necessary to use information regarding the location of the abnormal part detected on the workpiece w on the transport conveyor 2 (for example, the position coordinates on the transport conveyor 2) as position information Ip. In this case, the output unit 112 may transmit information regarding the location of the abnormal part detected on the workpiece w on the transport conveyor 2 to the downstream device 200 as position information Ip. The downstream device 200 receives such position information Ip and can perform predetermined processing on the abnormal part located at the position on the transport conveyor 2 identified based on the position information Ip.

[0057] As described above, according to each embodiment, an imaging device 6 (line camera) installed so as to intersect the transport direction a of the transport conveyor 2 captures images of multiple workpieces w being transported in multiple rows on the transport conveyor 2. Multiple image frames GF are created by accumulating the image data G captured by the imaging device 6, and abnormal parts are detected from the workpieces w based on these multiple image frames GF. When detecting abnormal parts based on such image frames GF, among the consecutive first image frame GF1 and second image frame GF2, workpieces w1 that are located at the boundary 12 with the second image frame GF2 in the trailing end region 10 of the first image frame GF1 are excluded from the detection of abnormal parts based on the first image frame GF1. By excluding workpieces w1 that are not entirely included in the first image frame GF1 from the detection of abnormal parts in this way, false detections such as overlooking abnormal parts in parts of the workpiece not included in the first image frame GF1 can be effectively prevented.

[0058] Furthermore, it is possible to replace the components in the above-described embodiments with well-known components as appropriate, without departing from the spirit of this disclosure, and the above-described embodiments may also be combined as appropriate.

[0059] The contents described in each of the above embodiments can be understood, for example, as follows:

[0060] (1) A quality evaluation device according to one embodiment is: A quality evaluation device for evaluating the quality of multiple workpieces transported in multiple rows on a conveyor belt, An image frame creation unit for creating multiple image frames from continuous image data captured using a line camera installed so as to intersect the conveying direction of the conveyor, An image frame compositing unit for compositing a composite image frame based on a first image frame and a second image frame that are consecutive among the plurality of image frames, The system includes a detection unit that detects abnormal parts from the plurality of workpieces based on the composite image frame.

[0061] According to the embodiment of (1) above, a line camera installed so as to intersect the conveying direction of the conveyor captures images of multiple workpieces being transported in multiple rows on the conveyor. Multiple image frames are created from the image data continuously captured by the line camera. A composite image frame is created based on the first and second image data, which are consecutive images from the multiple image frames, and an abnormality is detected in the workpiece based on the composite image frame.

[0062] (2) In other embodiments, in the embodiment of (1) above, The image frame compositing unit creates the composite image frame by compositing the rear end region adjacent to the second image frame in the first image frame with the second image frame if the workpiece that is bound to the boundary with the second image frame is included in the rear end region adjacent to the second image frame.

[0063] According to the embodiment of (2) above, if the rear end region of the first image frame adjacent to the second image frame contains a workpiece that falls within the boundary with the second image frame, the rear end region is combined with the second image frame. As a result, even if the entire workpiece is not included in the first image frame because the workpiece included in the first image frame falls within the boundary with the second image frame, the combined image frame created by combining the rear end region with the second image frame can include the entire workpiece. By detecting abnormalities based on the combined image frame created in this way, the accuracy of abnormality detection can be suitably improved.

[0064] (3) In other embodiments, in the embodiment of (1) or (2) above, The detection unit excludes the workpiece that falls within the upstream or downstream boundary of the composite image frame from the detection target of the abnormal part.

[0065] According to the embodiment of (3) above, among the workpieces included in the composite image frame, workpieces that fall on the upstream boundary or the downstream boundary are excluded from the detection of abnormalities. By excluding workpieces that are not entirely included in the composite image frame from the detection of abnormalities in this way, false detections such as overlooking abnormalities in parts of workpieces not included in the composite image frame can be effectively prevented.

[0066] (4) In other embodiments, in the embodiment of (3) above, The detection unit identifies the distribution area of ​​the workpiece in the composite image frame and identifies the workpiece relating to the upstream boundary or the downstream boundary by comparing the coordinates of the distribution area with the coordinates of the upstream boundary or the downstream boundary.

[0067] According to the embodiment of (4) above, by comparing the coordinates of the distribution area of ​​the workpieces identified in the composite image frame with the coordinates of the upstream or downstream boundary of the composite image frame, it is possible to suitably determine whether each workpiece included in the composite image frame falls on the upstream or downstream boundary of the composite image frame.

[0068] (5) In other embodiments, in any one embodiment of (1) to (4) above, The conveying conveyor includes a plurality of conveying regions arranged along the width direction intersecting the conveying direction, The detection unit transmits positional information indicating the transport area in which the workpiece in which the abnormal part was detected is transported, based on the coordinate ranges of each of the plurality of transport areas included in the image data and the coordinate information of the abnormal part detected by the detection unit, to the downstream device of the transport conveyor.

[0069] According to the embodiment of (5) above, the conveyor is configured to include multiple conveying areas. In particular, in this embodiment, abnormal parts are detected for each of the multiple workpieces that are conveyed in multiple rows on the conveyor, as described above, and as a result, position information indicating the conveying area in which the workpiece with the detected abnormal part is being conveyed is transmitted to the downstream device of the conveyor. This makes it possible for the downstream device of the conveyor to identify which conveying area of ​​the conveyor the workpiece with the detected abnormal part is being conveyed in based on the position information, and to handle the workpiece with the abnormal part separately from other workpieces.

[0070] (6) In other embodiments, in the embodiment of (5) above, The downstream device is a sorting device for sorting workpieces in each transport area based on the detection results from the detection unit.

[0071] According to the embodiment of (6) above, the sorting device located downstream of the conveying conveyor can identify which conveying area of ​​the conveying conveyor the workpiece in which an abnormal part has been detected is being transported based on positional information, and can suitably sort the workpiece having the abnormal part from the normal workpiece.

[0072] (7) In other embodiments, in any one embodiment of (1) to (6) above, The detection unit detects the abnormal area by comparing the brightness value of each pixel included in the image frame with a preset threshold.

[0073] According to the embodiment of (7) above, abnormal parts of the workpiece can be suitably detected by comparing the brightness value of each pixel included in the image frame with a preset threshold.

[0074] (8) In other embodiments, in any one embodiment of (1) to (7) above, The aforementioned line camera is a trilinear line sensor.

[0075] According to the embodiment of (8) above, by using a trilinear line sensor as a line camera for capturing image data, high-resolution images can be acquired and images can be processed at high speed.

[0076] (9) A quality evaluation system relating to one aspect is: A line camera for capturing the aforementioned image data, A quality evaluation apparatus relating to any one of the above embodiments (1) to (8), It is equipped with.

[0077] According to the embodiment of (9) above, the quality of the workpieces can be suitably evaluated based on image data captured of workpieces being transported in multiple rows on a conveyor belt.

[0078] (10) A quality evaluation method relating to one aspect is: A quality evaluation method for evaluating the quality of multiple workpieces transported in multiple rows on a conveyor belt, A step of creating multiple image frames from image data continuously captured using a line camera installed so as to intersect the conveying direction of the conveying conveyor, A step of synthesizing a composite image frame based on a first image frame and a second image frame that are consecutive among the plurality of image frames, A step of detecting abnormal parts from the plurality of workpieces based on the composite image frame, It is equipped with.

[0079] According to the embodiment of (10) above, a line camera installed so as to intersect the conveying direction of the conveying conveyor captures images of multiple workpieces being transported in multiple rows on the conveying conveyor. Multiple image frames are created from the continuous image data captured by the line camera. A composite image frame is created based on the first and second image data, which are consecutive images from the multiple image frames, and an abnormality is detected in the workpiece based on the composite image frame. [Explanation of Symbols]

[0080] 1. Quality Evaluation System 2. Conveyor 2a, 2b pulleys 2c belt 4 light source 6. Imaging device 8 boxes 10 Posterior end area 12 boundaries 12a Upstream boundary 12b Downstream boundary 100 Quality evaluation device 102 Image Data Creation Department 104 Image Frame Creation Section 106 Image frame composition unit 108 Detection unit 110 Location Information Creation Department 112 Output section 200 Downstream equipment G Image Data GF Image Frame GF1 First Image Frame GF2 2nd image frame CGF composite image frame SGF Sub-image Frame IP location information L1 illumination light L2 imaging light R1 First transport area R2 Second transport area R3 Third transport area

Claims

1. A quality evaluation device for evaluating the quality of multiple workpieces transported in multiple rows on a conveyor belt, An image frame creation unit for creating multiple image frames from image data continuously captured using a line camera installed so as to intersect the conveying direction of the conveyor, An image frame compositing unit for compositing a composite image frame based on a first image frame and a second image frame that are consecutive among the plurality of image frames, A quality evaluation apparatus comprising a detection unit that detects abnormal parts from the plurality of workpieces based on the composite image frame.

2. The quality evaluation apparatus according to claim 1, wherein the image frame synthesis unit creates the synthesized image frame by synthesizing the rear end region adjacent to the second image frame of the first image frame with the second image frame if the workpiece that is bound to the boundary with the second image frame is included in the rear end region of the first image frame adjacent to the second image frame.

3. The quality evaluation apparatus according to claim 1 or 2, wherein the detection unit excludes the workpiece relating to the upstream boundary or downstream boundary of the composite image frame from the detection target of the abnormal part.

4. The quality evaluation apparatus according to claim 3, wherein the detection unit identifies the distribution area of ​​the workpiece in the composite image frame, and identifies the workpiece relating to the upstream boundary or the downstream boundary by comparing the coordinates of the distribution area with the coordinates of the upstream boundary or the downstream boundary.

5. The conveying conveyor includes a plurality of conveying regions arranged along the width direction intersecting the conveying direction, The quality evaluation apparatus according to claim 1 or 2, wherein the detection unit transmits position information indicating the transport area in which the workpiece in which the abnormal part was detected is transported, based on the coordinate ranges of each of the plurality of transport areas included in the image data and the coordinate information of the abnormal part detected by the detection unit, to a downstream device of the transport conveyor.

6. The quality evaluation apparatus according to claim 5, wherein the downstream device is a sorting device for sorting the workpieces for each transport area based on the detection results by the detection unit.

7. The quality evaluation apparatus according to claim 1 or 2, wherein the detection unit detects the abnormal part by comparing the brightness value of each pixel included in the image frame with a preset threshold.

8. The quality evaluation apparatus according to claim 1 or 2, wherein the line camera is a trilinear line sensor.

9. A line camera for capturing the aforementioned image data, A quality evaluation apparatus according to claim 1 or 2, A quality evaluation system equipped with the following features.

10. A quality evaluation method for evaluating the quality of multiple workpieces transported in multiple rows on a conveyor belt, A step of creating multiple image frames from image data continuously captured using a line camera installed so as to intersect the conveying direction of the conveying conveyor, A step of synthesizing a composite image frame based on a first image frame and a second image frame that are consecutive among the plurality of image frames, A step of detecting abnormal parts from the plurality of workpieces based on the composite image frame, A quality evaluation method comprising the following features.

Images