Pod and fruit sorting system and pod and fruit sorting device

Abstract

To provide a legume selection system and a legume selection device that have higher selection accuracy for legumes by using an artificial intelligence (AI).SOLUTION: A legume selection system is for testing or selecting legumes with shells by analyzing an image of the legumes, and includes: image acquisition means for acquiring an image taken by imaging means; and image analysis means for analyzing the image acquired by the image acquisition means. The image analysis means includes: a legume image extraction means for extracting an image of each legume in the acquired image; and a contraction processing extraction unit for executing contraction processing of reducing an outline region of the images of the extracted legumes.SELECTED DRAWING: Figure 13

Description

[Technical field]

[0001] The present invention relates to a pod sorting system and device for sorting pods such as green soybeans, particularly pods with pods still attached, and also to a pod sorting system and device that uses artificial intelligence (AI) to accurately sort pods with pods still attached. [Background technology]

[0002] Pods are the fruits found in many legume plants, including beans, peas, mimosa, soybeans, redbuds, lupines, clovers, and vetch. These pods are often harvested with the pods still attached, and in the case of edamame, which are harvested from immature soybeans, they are often served with the pods still attached. When these pods are distributed to the market, they are sorted to control their quality and grade, and several technologies have been proposed to use cameras to sort pods, including edamame, with the aim of automating the sorting of pods.

[0003] The applicant of the present application has proposed such a pod and fruit sorting system and device in Patent Document 1 (WO 2016 / 133175). This document realizes a pod and fruit sorting system and device that, when observing the seeds in the pods, does not require photographing with transmitted light, but photographs reflected light to inspect the seed storage condition (particularly the number and size of seeds) in the pods, and preferably inspects or sorts the pods based on the vegetation index value calculated from the photographed image.

[0004] Such a pod and fruit sorting system inspects or sorts pods and fruits by analyzing images of pods with pods, and includes an image acquisition means for acquiring images captured by an image acquisition means, and an image analysis means for analyzing the images acquired by the image acquisition means, and the image analysis means performs a size inspection for measuring the length and width of each pod and fruit based on the captured image, and a seed number inspection for calculating the number of seeds contained in the pod based on the measured pod length, or alternatively, a pod and fruit sorting system that inspects or sorts pods and fruits based on the value of the vegetation index.This document makes it possible to inspect the state of seeds contained in the pods (particularly the number and size of seeds), defects such as black spots, discoloration, or the presence or absence of cracks or hairs.

[0005] Patent Document 2 (JP Patent Publication No. 2008-20347A) proposes a pod / fruit evaluation structure for judging the quality of pods / fruits. This document proposes a pod / fruit evaluation structure that includes an imaging means for capturing an image of the pod / fruit by using light transmitted through the pod with beans contained therein, and that judges the state of beans contained in the pod, the state of damage to the pod / fruit, the size of the pod / fruit, the state of cracks at the longitudinal ends of the pod / fruit, the shape of the pod / fruit, etc., based on the captured image. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] International Publication No. 2016 / 133175 [Patent Document 2] JP 2008-20347 A Summary of the Invention [Problem to be solved by the invention]

[0007] Various technologies have been proposed in the past for sorting pods and fruit with pods using cameras. However, the previously proposed pod and fruit sorting devices still have room for improvement in the accuracy of determining whether a product is good or bad. Therefore, the first objective of the present invention is to provide a pod and fruit sorting system and a pod and fruit sorting device that improve the accuracy of sorting pods and fruit by using artificial intelligence (AI) and the like.

[0008] In addition, while conventional pod and fruit sorting devices were also capable of detecting black spots and discoloration, due to the characteristics of image recognition, it was difficult to accurately detect black spots or discoloration by image inspection due to the influence of the outline area of ​​the pod and fruit. Therefore, in the present invention, another object of the present invention is to provide a pod and fruit sorting system and a pod and fruit sorting device that can accurately detect black spots or discoloration that exist in areas other than the outline of the pod and fruit.

[0009] In addition, in the conventional pod and fruit sorting device, the number of beans present in a pod is also determined, but since natural pods and fruits are the target, it is difficult to improve the accuracy of the appearance inspection. Therefore, in the present invention, another object is to provide a pod and fruit sorting system and a pod and fruit sorting device that can accurately determine the number of beans in a pod from the appearance of pods of various shapes.

[0010] In addition, in the conventional pod / fruit sorting devices, it is possible to sort good pods and bad pods, but it is not possible to check the judgment criteria used at the time of sorting later. Therefore, another object of the present invention is to provide a pod / fruit sorting system and a pod / fruit sorting device that can verify past judgment results later, check the validity of the judgment criteria, and modify the judgment criteria later.

[0011] In addition, in conventional pod and fruit sorting devices, cameras using image sensors such as CCD elements or CMOS elements are used as an imaging means for capturing images of pods and fruit, but due to the characteristics of the camera, distortion occurs in the image of areas away from the center. Such distortion causes errors when determining the size of pods and fruit, and is an obstacle to accurate pod and fruit sorting. Therefore, another object of the present invention is to provide a pod and fruit sorting system and pod and fruit sorting device that eliminates distortion in captured images and improves the sorting accuracy of pods and fruit, taking into consideration that the inspection target is pods and fruit in particular.

[0012] Furthermore, when an area sensor is used as the imaging means, it is possible that the same inspection object may be judged multiple times due to the movement of the pods and fruits to be inspected. If the same inspection object is judged and selected multiple times, the reliability of the good or defective rate of the harvested pods and fruits will decrease. Therefore, another object of the present invention is to provide a pod and fruit sorting system and a pod and fruit sorting device that can calculate accurate values ​​for the good or defective rate of the selected pods and fruits.

[0013] In addition, in the conventional pod and fruit sorting device, chipping was judged based on the shape of the pod and fruit, but it was difficult to accurately judge the presence or absence of chipping for pods of various shapes. Therefore, another object of the present invention is to provide a pod and fruit sorting system and a pod and fruit sorting device that can accurately judge the presence or absence of chipping by taking into account the unique shape characteristics of pods and fruit.

[0014] In addition, the conventional pod and fruit sorting devices require the setting of judgment criteria for each device when they are installed, which makes the initial setup process complicated. This is because the pods and fruit to be inspected vary greatly in the harvested region and growing environment, making it impossible to set uniform judgment criteria. On the other hand, the same user may use multiple pod and fruit sorting devices, and setting judgment criteria for each installed pod and fruit sorting device in such cases is a waste of work. Therefore, in the present invention, another object of the present invention is to provide a pod and fruit sorting system and a pod and fruit sorting device that can apply the judgment conditions of the initially set pod and fruit sorting device to other pod and fruit sorting devices when the pods and fruit to be judged are similar. [Means for solving the problem]

[0015] In order to solve the above problems, the present inventors have further improved the previously proposed pod and fruit sorting system and device, and have reviewed the judgment criteria and conditions to provide a pod and fruit sorting system and device that improves the judgment accuracy of good and bad products or makes it easier to set up at the time of installation.The present inventors also provide a pod and fruit sorting system and device that can preferably use artificial intelligence (AI) when selecting the judgment criteria.

[0016] That is, the present invention provides a pod and fruit sorting system and a pod and fruit sorting device for reliably determining the presence or absence of black spots in pods and fruits. Specifically, the present invention provides a pod and fruit sorting system for inspecting or sorting pods and fruits by analyzing images of pods and fruits with pods, the pod and fruit sorting system comprising an image acquisition means for acquiring an image acquired by an image acquisition means, and an image analysis means for analyzing the image acquired by the image acquisition means, the image analysis means comprising a pod and fruit image extraction unit for extracting images of each pod and fruit from the acquired images, and a contraction processing execution unit for executing a contraction processing to reduce the contour area of ​​the extracted images of each pod and fruit, and a pod and fruit sorting device using the pod and fruit sorting system.

[0017] The present invention also provides a pod / fruit sorting system and a pod / fruit sorting device that can accurately determine the pod / fruit quality of each pod by setting a measurement standard for the pod length while taking into consideration the characteristics of pods, particularly green soybeans.Specifically, the present invention provides a pod / fruit sorting system for inspecting or sorting pods by analyzing images of pods with pods, the system comprising an image acquisition means for acquiring images captured by an image acquisition means, and an image analysis means for analyzing the images acquired by the image acquisition means, the image analysis means comprising a pod / fruit image extraction unit for extracting images of each pod / fruit from the acquired images, and a pod length calculation unit for calculating the pod length from the extracted images of each pod / fruit, the pod length calculation unit calculating the pod length for an area where the pod width, which is calculated from the extracted images of each pod / fruit, is a predetermined value, and a pod / fruit sorting device using the pod / fruit sorting system.

[0018] The present invention also provides a pod and fruit sorting system and a pod and fruit sorting device that can review measurement standards by storing measurement data used in determining whether a pod or fruit is podded as log information. Specifically, the present invention provides a pod and fruit sorting system for inspecting or sorting pods and fruit by analyzing images of pods and fruit, the pod and fruit sorting system comprising an image acquisition means for acquiring images captured by an image acquisition means, an image analysis means for analyzing images acquired by the image acquisition means, and a log information storage unit for storing analysis results of the image analysis means as log information, and a pod and fruit sorting device using the pod and fruit sorting system. In particular, by using the log information, an artificial intelligence device can be constructed as a trained model, and by using the artificial intelligence device, more accurate determination authority can be set. In addition, the log information can be viewed from an external terminal via the Internet, so that the current inspection status can be monitored in real time from a remote location.

[0019] The present invention also provides a pod and fruit sorting system and a pod and fruit sorting device that improves the accuracy of pod and fruit determination by verifying the analysis results of the pod and fruit and the image to be analyzed. Specifically, the present invention provides a pod and fruit sorting system for inspecting or sorting pods and fruits by analyzing an image of the pod and fruit with pods, the pod and fruit sorting system comprising an image acquisition means for acquiring an image acquired by the image acquisition means, an image analysis means for analyzing the image acquired by the image acquisition means, and an image storage unit for storing the image used in the analysis by the image analysis means, and a pod and fruit sorting device using the pod and fruit sorting system.

[0020] The present invention also provides a pod and fruit sorting system and a pod and fruit sorting device that can solve the problem of erroneous judgment caused by a camera that photographs the pods and fruit to be sorted. Specifically, the present invention provides a pod and fruit sorting system for inspecting or sorting pods and fruit by analyzing an image of the pods and fruit, the pod and fruit sorting system comprising an image acquisition means for acquiring an image captured by the image acquisition means and an image analysis means for analyzing the image acquired by the image acquisition means, the image analysis means comprising a distortion correction unit for correcting distortion of the image acquired by the image acquisition means, and a pod and fruit sorting device using the pod and fruit sorting system.

[0021] The present invention also provides a fruit and pod sorting system and a fruit and pod sorting device that can prevent a decrease in the credibility of the proportion of good or defective products caused by discriminating the same fruit and pods multiple times. Specifically, the present invention provides a fruit and pod sorting system for inspecting or sorting fruit and pods by analyzing images of the fruit and pods, the fruit and pod sorting system including an image acquisition means for acquiring images acquired by an image acquisition means and an image analysis means for analyzing the images acquired by the image acquisition means, the image analysis means calculating a movement position for the images acquired by the image acquisition means, and excluding images of fruit and pods with pods present at the movement position from the analysis target, and a fruit and pod sorting device using the fruit and pod sorting system.

[0022] The present invention also provides a pod / fruit sorting system and a pod / fruit sorting device that can easily perform initial settings, i.e., adjustment of judgment criteria and various devices (including cameras, lighting, etc.), when installing multiple pod / fruit sorting systems or pod / fruit sorting devices. Specifically, the present invention provides a pod / fruit sorting system for inspecting or sorting pods and fruits by analyzing images of the pods and fruits, the pod / fruit sorting system comprising an image acquisition means for acquiring images acquired by an image acquisition means and an image analysis means for analyzing the images acquired by the image acquisition means, the image analysis means setting analysis criteria based on analysis information setting information acquired from other pod / fruit sorting systems, and a pod / fruit sorting device using the pod / fruit sorting system.

[0023] The present invention provides a pod and fruit sorting system and a pod and fruit sorting device that can accurately determine whether or not there are chips in pods, particularly in green soybeans.Specifically, the pod and fruit sorting system is for inspecting or sorting pods and fruit by analyzing images of pods with pods, and includes an image acquisition means for acquiring images acquired by the image acquisition means and an image analysis means for analyzing the images acquired by the image acquisition means, and the image analysis means includes a pod and fruit image extraction section for extracting images of each pod and fruit from the acquired images, and divides the extracted images of each pod and fruit into a plurality of parts and calculates the area ratio between the base end and the tip end in the length direction to determine whether or not there are chips in the pod and fruit, and a pod and fruit sorting device using the pod and fruit sorting system.

[0024] The pod / fruit sorting system according to the present invention can be formed using a computer. Specifically, it can be embodied as a computer whose processing contents and processing procedures are controlled by computer software. The image analysis means can be configured with a CPU and memory that perform processing by executing the computer software.

[0025] In addition, the pod and fruit sorting system is configured with a transporting means such as a belt conveyor or a vibrating feeder for transporting the pods and fruits, and a photographing means such as a digital camera or a digital video camera for photographing the pods and fruits being transported, and may further include a removal means for removing those pods and fruits that do not meet the standards based on the judgment results.

[0026] Furthermore, by combining the above-configured fruit sorting system with a fruit and pod transport means, etc., a fruit and pod sorting device capable of continuously sorting fruit and pods can be provided. That is, a fruit and pod sorting device is provided that is equipped with a transport means for transporting the inspection object, a photographing means for photographing the inspection object transported by the transport means, and a fruit and pod sorting means for acquiring the images photographed by the photographing means and analyzing the images, and uses the above-configured fruit and pod sorting system as the fruit and pod sorting means.

[0027] According to this pod and fruit sorting device, the pods and fruits to be inspected are continuously transported by a conveying means, photographed by a photographing means, and analyzed by the pod and fruit sorting system of the present invention, thereby making it possible to find pods and fruits that do not meet the standards and to reject them if necessary.

[0028] In addition, the above-mentioned pod and fruit sorting device according to the present invention is preferably further provided with a discharge means for blowing air onto the inspection objects designated by the pod and fruit sorting means among the inspection objects conveyed by the conveying means, and blowing off the inspection objects that do not meet the standard with air. In this case, the image analysis means in the pod and fruit sorting means can identify the center of length and width from the result of a size inspection that measures the length and width of each pod and fruit from the photographed image. Therefore, it is preferable that the discharge means is configured to blow air toward the center of length and width of the inspection object designated by the pod and fruit sorting means. The pods that are blown toward the center (pods that do not meet the standard) are reliably extracted, and thus the harvested pods and fruits can be inspected and sorted accurately and reliably even when a large amount of sorting is required. Effect of the Invention

[0029] The above-mentioned edamame sorting system and pod / fruit sorting device of the present invention using the same can provide a pod / fruit sorting system and pod / fruit sorting device that can accurately determine black spots or discoloration present in areas other than the pod / fruit contours without being affected by the pod / fruit contour areas when the image analysis means that analyzes the image acquired by the image acquisition means is equipped with a pod / fruit image extraction unit that extracts images of individual pods from the acquired image, and a shrinkage processing execution unit that performs shrinkage processing to reduce the contour areas of each extracted pod / fruit image.

[0030] In addition, the image analysis means that analyzes the images acquired by the image acquisition means is equipped with a pod / fruit image extraction unit that extracts images of individual pods and fruits in the acquired images, and a pod length calculation unit that calculates the pod length from the extracted images of each pod and fruit.When the pod length calculation unit calculates the pod length for an area where the pod width calculated from the extracted images of each pod and fruit is a predetermined value, it is possible to provide a pod / fruit sorting system and pod / fruit sorting device that can clearly determine differences in the number of kernels within a pod even if the pod lengths are the same, particularly in the case of edamame.

[0031] Furthermore, when a log information storage unit that stores the analysis results of the image analysis means as log information is provided, a fruit sorting system and a pod / fruit sorting device can be provided that enable detailed confirmation of the non-defective product rate (standard products) and defective product rate (non-standard products) of the sorting objects, or the reasons for determining that the objects are defective, etc. Furthermore, when an image acquisition means that acquires an image captured by the photographing means and an image analysis means that analyzes the image acquired by the image acquisition means are provided, and an image storage unit that stores the image used by the image analysis means for analysis, a pod / fruit sorting system and a pod / fruit sorting device can be provided that enables later verification of the judgment criteria used at the time of sorting, and enables the validity of the judgment criteria and subsequent correction of the judgment criteria.

[0032] In addition, if the image analysis means is equipped with a distortion correction unit that performs distortion correction on the image acquired by the image acquisition means, it is possible to provide a pod and fruit sorting system and a pod and fruit sorting device that eliminate image distortion in areas away from the center caused by the characteristics of the camera, thereby improving the accuracy of pod and fruit sorting.

[0033] Furthermore, if the image analysis means calculates the movement position for the image acquired by the image acquisition means, and excludes images of pods with pods present at the movement position from the analysis target, it is possible to provide a pod / fruit sorting system and pod / fruit sorting device that can calculate accurate values ​​for the proportion of good or defective pods among the sorted pods / fruit, without having to judge and select the same inspection object multiple times.

[0034] In addition, when the image analysis means includes a pod image extraction unit that extracts images of individual pods from the acquired image, divides the extracted images of each pod into multiple equal parts, and calculates the area ratio between the base end and the tip in the longitudinal direction to determine whether or not there are any chips in the pods, a pod sorting system and pod sorting device can be provided that can accurately determine whether or not there are any chips, taking into account the unique shape characteristics of pods, particularly edamame beans.

[0035] Furthermore, when the image analysis means sets the analysis criteria based on analysis information setting information obtained from other pod and fruit sorting systems, it is possible to eliminate the waste of work of setting judgment criteria for each installed pod and fruit sorting device, and if the pods and fruits to be judged are similar, it is possible to provide a pod and fruit sorting system and a pod and fruit sorting device in which the judgment conditions set for the initially set pod and fruit sorting device can be expanded to other pod and fruit sorting devices. [Brief description of the drawings]

[0036] [Figure 1] An overall configuration diagram showing a soybean sorting device configured using the soybean sorting system according to the present embodiment. [Diagram 2] Block diagram showing the hardware configuration [Diagram 3] Flowchart showing the contents of basic processing [Figure 4] A processing diagram showing the contents of the basic processing [Diagram 5] Flowchart showing the process of inspecting the number of grains [Figure 6] Diagram of the process for inspecting grain count [Figure 7] Diagram of the process for determining pod length in grain number inspection [Figure 8] Flowchart showing the process of single grain inspection [Figure 9] Flowchart showing the process of chipping inspection [Figure 10] Diagram of the process for chipping inspection [Figure 11] Flowchart showing the process of performing black spot inspection [Figure 12] Front view showing black spots [Figure 13] Settings screen showing the reduction process for black spot inspection [Figure 14] Screen showing inspection log [Figure 15] Setting screen to prohibit multiple inspections of the same edamame [Figure 16] Inspection image save setting screen [Figure 17] Camera correction setting screen [Figure 18] Calibration setting screen DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0037] Hereinafter, one embodiment of the fruit pod sorting system according to the present invention will be specifically described with reference to the drawings. In particular, in this embodiment, a fruit pod sorting system for sorting green soybeans among fruit pods will be specifically described. However, the fruit pod sorting system and fruit pod sorting device according to this embodiment can also be used to sort fruit pods other than green soybeans, such as soybeans, kidney beans, chickpeas, adzuki beans, and peanuts. Furthermore, the fruit pod sorting system and fruit pod sorting device according to this embodiment are for inspecting fruit pods with pods, and therefore the fruit pods refer to fruit pods with pods.

[0038] 1 is an overall configuration diagram showing a pod and fruit sorting device (hereinafter referred to as "edamame sorting device") configured using a pod and fruit sorting system (hereinafter referred to as "edamame sorting system 40") according to this embodiment. This edamame sorting device is composed of a hopper section 10 that stores edamame W to be sorted and continuously supplies them to a sorting line, a first conveying means 20 that conveys the edamame W supplied from the hopper section 10, a second conveying means 30 that inverts and conveys the edamame W conveyed by the first conveying means 20, photographing means 21, 31 provided on each conveying means, an edamame sorting system 40 that analyzes the images photographed by the photographing means 21, 31, conveying detection means 22, 32 provided on each conveying means for detecting the conveying speed or conveying amount, and discharge means 23, 33 provided at the drop section of the edamame W at the end side of each conveying means.

[0039] The hopper section 10 has a space for storing the edamame W (pods) to be sorted. It also has a discharge structure for supplying the edamame W stored inside the hopper section 10 to a conveying means consisting of a first conveying means 20 and a second conveying means 30. This discharge structure is a belt conveyor or the like provided on the bottom surface of the hopper section 10, and moves the edamame W stored inside the hopper section 10 forward. However, the hopper section 10 may have a different configuration as long as it can send out the edamame W stored therein to the conveying means where inspection is carried out.

[0040] The soybeans W discharged from the hopper section 10 are supplied to a conveying means consisting of a first conveying means 20 and a second conveying means 30. At that time, it is also preferable to provide a vibration means (not shown) at the falling portion for vibrating and conveying the soybeans W discharged from the hopper section 10 so that the soybeans W do not overlap each other.

[0041] The first conveying means 20 and the second conveying means 30 can each be formed by a belt conveyor, a vibrating feeder, or the like. The edamame W supplied from the hopper section 10 is first conveyed by the first conveying means 20. This first conveying means 20 can preferably be accompanied by an alignment section for aligning the direction of the edamame W being conveyed in a certain direction. Such an alignment section can be formed, for example, by arranging long alignment plates or wires (not shown) at predetermined intervals in the conveying direction of the edamame W. When the edamame W being conveyed hits this alignment plate, its orientation is corrected so that it is vertical in the conveying direction.

[0042] Above the first conveying means 20, there is provided a photographing means 21 for photographing the edamame W being conveyed by the first conveying means 20. A camera or a video camera can be used as the photographing means 21. It is preferable that the photographing means 21 photographs the edamame W being conveyed as still or moving images, and obtains the photographed images as electronic data so that the photographed images can be directly imported into the edamame sorting system 40 for analysis. It is also preferable to provide lighting that stably illuminates the edamame W present within the photographing range of the photographing means 21. This is to stabilize the luminance of the photographed image by making the brightness of the light illuminating the edamame W constant.

[0043] The images captured by the imaging means 21 are sent to an edamame sorting system 40, which is composed of a computer and the like. The edamame sorting system 40 inspects the edamame W to be sorted based on the acquired captured images, inspecting at least the size and number of the edamame W, or the presence or absence of chips, black spots, and / or discoloration, and identifies edamame W that do not meet the standards. The configuration and processing of the edamame sorting system 40 will be described later.

[0044] The first conveying means 20 is provided with a conveying detection means 22 for acquiring the conveying speed and conveying distance. This is to identify the position to which the edamame W selected by the edamame sorting system 40 (non-standard edamame W) has been conveyed by associating the conveying information of the edamame W with the photographed image. By identifying the non-standard edamame W selected by the edamame sorting system 40 during the conveying process, the non-standard edamame W can be discharged outside the conveying line. A rotary encoder, a linear encoder, or the like can be used as the conveying detection means 22.

[0045] In this embodiment, the non-standard edamame W are removed when they fall from the end of the first conveying means 20. That is, a drop section is provided at the end of the first conveying means 20, which allows the conveyed edamame W to fall into the second conveying means 30, and when they fall from the drop section, the edamame W identified by the edamame sorting system 40 (non-standard edamame W) are discharged from the conveying line by the discharge means 23. The discharge means 23 can be configured to blow air against the identified edamame W to blow them out of the conveying line, or can be formed in various configurations that allow the identified edamame W to be removed.

[0046] As described above, the edamame W transported by the first transport means 20 and determined to be out of standard by the edamame sorting system 40 are discharged by the discharge means 23, and the remaining edamame W are transported by the second transport means 30. When transported by this second transport means 30, it is preferable to transport the edamame W in a state inverted from the state in which it was transported by the first transport means 20. Therefore, it is preferable to provide a reversing means for reversing the edamame W to be transported after the first transport means 20 or on the starting end side (first transport means 20 side) of the second transport means 30. Such a reversing means can, for example, reverse the edamame W to be transported by utilizing the drop when the edamame W to be transported is dropped, or can reverse the edamame W by applying a rotational force by a belt or roller to any side of the pods to be transported. In addition, the edamame W that has fallen into the second transport means 30 can be reversed by transporting the second transport means 30 in the opposite direction to the first transport means 20.

[0047] The edamame W transported by this second transport means 30 are also sorted in the same way as those transported by the first transport means 20. That is, the edamame W transported by the second transport means 30 are photographed by the photographing means 31, and the photographed image is analyzed by the edamame sorting system 40 to identify edamame W that do not meet the standard. At that time, the transport speed and distance of the edamame are obtained from the transport detection means 32 consisting of a rotary encoder. The non-standard edamame W are then discharged by blowing air onto them by the discharge means 33 at the end of the transport means. Since the process in the second transport means 30 is the same as that in the first transport means 20, in the drawings, the same components as those in the first transport line are indicated by adding a value of "10" to the reference numerals, and detailed explanations thereof will be omitted.

[0048] Fig. 2 shows an example of the hardware configuration of a computer constituting the pod and fruit sorting system according to the present embodiment. However, the computer 500 in Fig. 2 merely illustrates a typical configuration of the pod and fruit sorting system, and the edamame sorting system may be configured as a dedicated device as long as it has a calculation device, memory, and program for image analysis.

[0049] The computer 500 shown in Figure 2 is composed of a CPU 501, a memory 502, an audio output device 503, a network interface 504, a display controller 505, a display 506, an input device interface 507, a keyboard 508, a mouse 509, an external storage device 510, an external recording medium drive device 511, and a bus 512 that connects these components to each other.

[0050] CPU 501 controls the operation of each component of computer 500, and under the control of the OS, controls the execution of each process of the above-mentioned edamame sorting system, and controls its operation. Memory 502 is usually composed of ROM (Read Only Memory), which is a non-volatile memory, and RAM (Random Access Memory), which is a volatile memory, and corresponds to a storage unit. The ROM stores programs and the like executed by the edamame sorting system, which is executed when computer 500 is started up. The RAM temporarily stores programs executed by CPU 501, which analyze images acquired from a photographing means, calculate the size and number of edamame, and determine the presence or absence of chips, whiskers, black spots, discoloration, cracks, or activity, as well as data used by these programs during their execution.

[0051] The audio output device 503 is a device that outputs audio, such as a speaker, and the network interface 504 is an interface for connecting to a network 520 for exchanging information with various devices. The display controller 505 is a dedicated controller for processing drawing commands issued by the CPU 501, and outputs drawing data to a display 506 as a display unit. The display 506 is a display device constituted by an LCD or the like.

[0052] The input device interface 507 receives signals input from input / output devices such as a keyboard 508, a mouse 509, or a touch pad, and provides a predetermined command to the CPU 501 according to the signal pattern. The keyboard 508 and the mouse 509 are necessary when performing operations such as program execution and setting. The external storage device 510 is also included in the category of storage means in this specification. The external storage device 510 can be configured as a storage device such as a hard disk drive (HDD). The above-mentioned programs and data are recorded in this device, and are loaded from there to the RAM of the memory 502 as necessary during execution. The external recording medium drive device 511 is a device that accesses the recording surface of a portable external recording medium 530 such as a CD (Compact Disc), MO (Magnet-Optical Disc), or DVD (Digital Versatile Disc) and reads the data recorded thereon.

[0053] The green soybean selection system according to this embodiment is formed using a computer configured as described above, and executes the processes shown in the following FIGS.

[0054] <Basic Processing> Fig. 3 is a flow chart showing the contents of the basic processing executed for the photographed image acquired from the photographing means, and Fig. 4 is a process content diagram showing the contents of this basic processing. When the basic processing is executed, the edamame sorting system first acquires a photographed image showing multiple edamame from the photographing means and records it in memory (S11). Then, as shown in Fig. 4(B), a labeling process is performed for each edamame for the image showing multiple edamame (the photographed image acquired from the photographing means), and an image for each edamame is extracted. When analyzing the photographed image, it is desirable that the multiple pods contained in the photographed image are extracted for each edamame, and by performing the labeling process, the outline shape of the pods can be identified.

[0055] Then, an angle correction process is performed on the image of each extracted soybean (S13). The length in the long axis direction (hereinafter also referred to as "pod length") and the width in the short axis direction (hereinafter also referred to as "pod width") of the angle-corrected image are measured. By performing angle correction, the length and width of the pod in the pod can be measured more accurately. In this angle correction process, image processing is performed so that the long direction (length direction) of each soybean image faces the same direction, either vertically or horizontally. Then, as shown in Figure 4(C), a binarization process is performed on the angle-corrected image (S15), so that necessary numerical values ​​and locations can be quickly identified and obtained in each subsequent inspection.

[0056] When sorting pods and fruits automatically, they need to be transported by a conveyor belt or other transporting means, and the transported pods and fruits need to be photographed and inspected. However, it is practically difficult to accurately align the orientation of the pods and fruits transported by the transporting means. Therefore, as described above, the pods and fruits are photographed as they are transported randomly, and their orientation is corrected, resulting in a pod and fruit sorting system that can accurately measure the length and width of the pods and fruits while conforming to the actual inspection conditions.

[0057] <Grain number inspection> Fig. 5 is a flow chart showing the processing for inspecting the number of soybeans, Fig. 6 is a processing diagram for inspecting the number of soybeans, and Fig. 7 is a processing diagram for determining the pod length in inspecting the number of soybeans. In inspecting the number of soybeans, an image that has undergone the basic processing (image after labeling and angle correction) is first obtained as an inspection image. Since this inspection image is created for each soybean transported by the transport means, this inspection is performed for all soybeans.

[0058] The image analysis means performs a "size inspection" to obtain information about the length and width of each pod from the captured image. The information about the length and width can be obtained as the number of dots (or pixels) based on the pixels in the captured image, or as the actual length and width of the object to be inspected. When calculating the actual length and width of the object to be inspected from the captured image, the length identified on the captured image can be multiplied by a certain enlargement or reduction constant.

[0059] That is, when the inspection image is acquired, the CPU of the soybean sorting system, which is configured as a computer, reads the number of pixels in the length direction in the area where the soybeans are shown in the inspection image as shown in Figure 6 (A), and calculates the length of the soybeans based on this (S52). This calculation can be performed by multiplying the captured image by a coefficient to convert it into an actual measurement value.

[0060] When the calculated length of the soybeans is obtained, the number of seeds is calculated based on this length (S54). As shown in FIG. 6, the number of seeds is calculated by preparing a length / number table 41 that associates the number of seeds with the length of the pod, and searching the pod length read by image analysis as a search value to calculate and extract the number of seeds corresponding to that length. Also, the program can be designed to output a fixed value (number of seeds) according to the acquired pod length without searching the file system. In other words, by measuring the length of the pod and using this value, the approximate number of seeds contained in the pod can be calculated.

[0061] When calculating the pod length, a standard value for the pod width can be set, and the pod length can be calculated for the range above that value. That is, as described above, it is possible to estimate the approximate number of seeds from the length of the pod in edamame. However, even with the same pod length, there may be one or two seeds, depending on the size of the seeds inside. Therefore, by excluding the length of the area where no seeds exist, the length of the area where seeds actually exist can be calculated, and the number of seeds can be accurately determined.

[0062] FIG. 7 is a process diagram for pod length determination in the grain number inspection. As shown in FIG. 7(A), even if the pod length is the same, the number of seeds contained in the pod may be one or two. Therefore, a reference value (LB) for the pod width is set, and the pod length for values ​​larger than the reference value is calculated, so that the effective length of the area in which the seeds exist can be calculated. In this embodiment, the reference value (LB) for the pod width is set to 10 mm, and as a result, as shown in FIG. 7(B), the pod length of a pod containing one seed is shorter than that of a pod containing two seeds, and it is determined that there is one seed. On the other hand, in a pod containing two seeds, the pod length of the area equal to or larger than the reference pod width is within the range of the pod length that is determined to be two seeds, so that the number of seeds can be determined accurately.

[0063] In addition, when calculating the pod length, it is possible to determine whether the pod contains one seed by calculating the length of the area below the standard pod width, and it is also possible to calculate the number of seeds in the pod by calculating the number of protrusions (convex parts) that are greater than the standard pod width.

[0064] In the above kernel number inspection, the number of seeds in the pod can be calculated, and the width of the pod, which changes due to the presence of seeds, can be calculated. That is, in the kernel number inspection using the image analysis means, the position where the pod width is maximum is regarded as a convex part, and the position between the convex parts where the width is narrower than the concave part and is the minimum is regarded as a concave part. Such convex parts and concave parts can also be calculated for each seed contained in the pod, and this allows the overall shape of the pod to be grasped and judged in detail.

[0065] In addition, since the outline of the pod is bulged, it is possible to define convex parts where seeds exist, and since there are concave parts between the seeds, by checking the existence of the concave parts, it is possible to confirm that seeds exist in the convex parts. If the number of convex parts is calculated to be two or more, it is also possible to determine the number of seeds in the pod by checking the amount of change in width between the convex parts. This is because if the number of seeds is determined only by the length of the pod, even if the pod length is the same, there may be one or two seeds depending on the size of the seeds inside. Therefore, it is desirable to check that the change in width between the convex parts changes so that one valley (concave) appears.

[0066] As described above, in the grain count inspection, it is also desirable to identify pods containing only one seed. Such pods containing one seed can be identified by the overall length of the pod. That is, pods whose pod length in the area where the pod width is equal to or greater than a certain value is shorter than the standard length can be assumed to contain only one seed. By predicting the number of seeds based on the length of the pod, the processing speed of the grain count inspection can be improved. Then, for pods whose length is equal to or greater than the standard value, it is determined that there are two or more seeds in the pod. However, even if it is determined that there are two or more seeds, if the seed is large, there may be only one seed actually contained. Therefore, the width can be read from the captured image, and the convex and concave parts can be extracted and confirmed to calculate the actual number of seeds. That is, if the length is equal to or greater than the standard value and the number of seeds is set to two, the pod is divided into two areas in the length direction, and the widest convex part in each area is extracted. As a result, if the pod contains two or more seeds, the convex parts should be present in two places, and there should be a concave part between the convex parts. However, if the pod contains one seed, even if there are two convex parts, there is no width between the convex parts that is narrower than both convex parts, and therefore no concave part exists. Therefore, for pods that are long enough to contain one seed, it is possible to determine whether there is one seed by checking whether there is a concave part between the convex parts.

[0067] This process of identifying pods with one seed per pod is particularly effective when selecting edamame, because the number of seeds in a pod also affects commercial value, and identifying and selecting pods with one seed can increase commercial value.

[0068] The pods can be selected by determining whether the number of seeds obtained as a result of the calculation is one or not (S56). This is because soybeans with one seed may have low commercial value, so soybeans with one seed are selected and removed to increase their commercial value. In determining whether the pod contains one seed or not, first, those determined to contain one seed in the above-mentioned number of seeds are removed. Therefore, soybeans determined to contain only one seed in the pod are determined to be substandard and are identified as items to be discharged by the discharge means.

[0069] Then, in the above-mentioned kernel number judgment, when it is judged that there are two or more seeds in the pod, the area is divided according to the number of seeds (S58). FIG. 6 shows an example where it is judged that there are two seeds in particular. For this reason, the inspection image is divided into two areas (A1 and A2) in the length direction as shown in FIG. 6(B). Next, in each area, the point with the largest width value is set as the convex part. In FIG. 6(C), the widest point in the first area A1 is set as the first convex part C1, and the widest point in the second area A2 is set as the second convex part C2. Then, the value (width) of each convex part is calculated (S59).

[0070] Next, as shown in Fig. 6(D), the width SA between each of the convex parts (C1 and C2) is calculated, and as shown in Fig. 6(E), a point where the width is smaller than the width of each convex part and is the minimum value is set as a concave part, and its width is calculated (S60). Then, the presence or absence of a concave part is determined (S61), and if the concave part can be calculated, the pod contains two or more seeds, and the soybean meets the standard, and is determined to be within the standard in this grain number inspection (S62). On the other hand, if the concave part cannot be calculated, the pod is determined to contain one seed, and the soybean is determined to be out of standard, and is specified as a target for discharge by the discharge means. That is, the soybean sorting system executes the process of specifying the convex parts and concave parts in the same manner as above, as a result of determining that the number of seeds in the pod is two from the length of the pod (Fig. 8(A)). Specifically, the image of the edamame is divided into two regions (A1 and A2) (Fig. 8(B)), and the widest area in each region is set as the respective convex portion (C1 and C2) (Fig. 8(C)). Then, the position between these convex portions that is narrower than each of the convex portions and has the narrowest width is set as the concave portion (Fig. 8(D)). However, if there is only one seed in the pod, the width of this concave portion will be the same as that of one of the convex portions, and therefore it is not possible to set the concave portion (Fig. 8(E)). Therefore, in this case, it is determined that there is only one seed in the pod, and the edamame is identified as a substandard product.

[0071] By performing the grain count inspection using the above process, it is possible to speed up processing. That is, because edamame beans with one seed in the pod are initially eliminated based on the pod length, the number of beans to be inspected can be reduced, thereby speeding up processing. Furthermore, by further inspecting the convex and concave parts of edamame beans that are within the standard length, it is possible to extract edamame beans that meet the standard length but actually have one seed in the pod, greatly improving the accuracy of sorting. Therefore, by performing the grain count inspection using the above process, edamame beans (edamame beans with one seed in the pod) can be sorted quickly and accurately.

[0072] <Chip inspection> The green soybean sorting system can also perform chipping inspection by acquiring an inspection image created by performing basic processing. This chipping inspection checks whether or not there are chips in the green soybean pods, and in this embodiment, as shown in Figures 9 and 10, the image of the pod is divided into multiple equal parts in the length direction, and the presence or absence of chipping is judged based on the area ratio of both ends. This is based on the fact that chipping of pods often occurs at the ends in the length direction in green soybeans, in particular.

[0073] In this process, the outline shape is specified by the labeling process, and an inspection image with the long axis direction aligned is obtained by the angle correction process (S91). The obtained inspection image is divided into multiple parts with equal lengths in the long axis direction (S92), and the areas of the divided areas are compared at both ends in the length direction (S93). Then, it is determined whether the comparison result is within a predetermined standard (S94), and if the difference between the two is small and the ratio is almost the same, it is determined that there is a chip in the pod and that it is out of standard (S96). On the other hand, if the difference between the two is large and the ratio is different, it is determined that there is no chip in the pod and that it is within standard (S95). This is because, as shown in Figure 10(A), when there is no chip at the ends in the length direction of the pod, one end (the tip of the pod) has a tapered shape, so the areas of both ends are different, while when there is a chip at one end (especially the end on the tip side of the pod) as shown in Figure 10(B), the areas at both ends are almost the same. The area ratio between these longitudinal ends can be set arbitrarily, but for example, if the area of ​​one end is 70% or less, or 50% or less, of the other end, it can be determined that no pods are missing. The comparison of areas can also be determined by the difference in areas, rather than the ratio. When dividing the pod in the longitudinal direction, it can be divided along the length of the longitudinal axis, and the number of divisions can be three as shown in Figure 10, or it can be two or more equal parts, or even four or more. It is also possible to compare the areas of both ends of the pod by extracting a predetermined length from each end of the pod length without dividing it equally.

[0074] By comparing the areas at both ends in this manner, it is possible to accurately determine whether or not there are any chips, even for soybeans whose pod widths vary in a complex manner, thereby improving the accuracy of soybean sorting.

[0075] <Black spot inspection> FIG. 11 is a flow chart showing the process of inspecting black spots, FIG. 12 is a front view showing the appearance of black spots, and FIG. Contraction (A) is a setting screen for the process, and (B) is a diagram showing the process. In this black spot inspection, first, an image captured by a photographing means such as a camera is subjected to labeling processing and tilt correction processing to obtain an inspection image (S111). Then, the contour (periphery) of the obtained inspection image is removed. Contraction This is because, in normal sorting of green soybeans, discolored areas (black spots) that exist in areas other than the outline of the pod as shown in Figure 12 are judged to be black spots, and discoloration that exists in the outline area is not judged to be black spots. Also, it cannot be denied that shadows, etc., may be cast on the inspection image when it is photographed, and this is to prevent the shadowed areas from being recognized as discolored areas.

[0076] It takes Contraction The process creates an image in which the outline area of ​​the inspection image is removed, as shown in Figure 13(B). The amount of outline removal can be set arbitrarily on the setting screen in Figure 13(A). If there is discoloration over a wide area in the outline area of ​​the pod, Contraction By setting the number of times of processing multiple times, it is possible to set many pixels to be reduced. Contraction You can set the number of pixels to be removed by processing and adjust the number of pixels to be removed by repeating this process multiple times. However, by adjusting the number of pixels to be removed in one go, Contraction The degree of processing can also be adjusted.

[0077] and ContractionFor the processed image, the area (number of pixels) of the discolored area as black spots is calculated (S113), and it is determined whether or not this is within a standard range (S114). If the calculated discolored area is within the standard range, it is determined to be within the standard (S115), and if it is beyond the standard range, it is determined to be out of the standard (S116).

[0078] In addition, when recognizing the discolored area, as shown in Fig. 13, upper and lower limits of the brightness value for determining whether or not it is a black spot can be set for at least one of R (red), G (green), and B (blue) in the inspection image, and anything within that range can be subject to black spot judgment. For example, in Fig. 10, the range of minimum 0 and maximum 60 is set for the brightness value of G (green), and anything with a brightness value of G (green) between 0 and 60 is subject to black spot judgment. As a result, discoloration such as black spots can be accurately judged by excluding discoloration and shadows in the outline area of ​​the pod.

[0079] <Inspection log management> In the pod and fruit sorting system and device according to this embodiment, the inspection results are stored, making it possible to confirm the ratio of good and defective soybeans among the soybeans being inspected, and the causes of defective soybeans. Figure 14 shows a pod and fruit sorting system and device that stores data used to judge soybeans as a log. Such inspection logs can be stored for a certain period of time, such as for each inspection date or inspection time, or for each inspection target lot (Figure 14(A)), and can be stored in a text format such as CSV as shown in Figure 14(B). However, the log may of course be stored in another file format.

[0080] In this embodiment, the log information can hold numerical values ​​resulting from judging the inspection image, such as pod length, pod width, area, or RGB brightness values, as well as other data such as actual numerical values ​​used in the judgment. By holding such log information, it is possible to confirm the status of non-defective products (within specifications) and defective products (outside the vaporizer), their ratio, or a breakdown of defective products from the description contents (raw data) of a specific log file (FIG. 14(C)). Therefore, by holding all data used in sorting edamame as the log information, it is possible to confirm the status of the edamame for each lot categorized by the harvesting farmer, harvesting area, etc.

[0081] In this way, when identifying the proportion of good and defective products and the reasons for products being judged as defective from the inspection log, in order to increase accuracy, it is desirable to configure the system so that edamame that have already been judged as good or defective are not judged again.

[0082] Therefore, as shown in FIG. 15, the movement position of the edamame once judged can be predicted, and the edamame present at the movement position can be excluded from the judgment target. That is, the movement amount in the transport direction (movement amount Y) and its allowable range are set on the setting screen shown in FIG. 15(A), and the allowable range of the movement amount in the direction intersecting the movement direction (movement amount X) is also set. Then, using this value, as shown in FIG. 15(B), if the inspection target once detected has moved to the set position, the inspection target present at that position is excluded from the target. In this embodiment, since the edamame to be inspected is moved by a conveying means such as a conveyor, the system is configured to predict the displacement in the conveyor movement direction in principle. Therefore, for the edamame judged in the photographed inspection image, the movement position in the transport direction after a predetermined time (the timing of photographing the inspection image) is predicted, and the edamame present at that position in the next photographed image is excluded from the inspection target. At this time, the amount of movement in the transfer direction (movement amount Y) can be set based on the moving speed of the transfer means such as a conveyor and the transfer detection means 22, 23, and then an error range for the moving distance can be set. Also, in this embodiment, the amount of movement in a direction intersecting the moving direction by the transfer means (movement amount X) is configured so that an error range can be set. This is based on the premise that, in principle, the edamame do not move in a direction intersecting the moving direction, but if they do move in a direction intersecting the moving direction, it is also possible to set the amount of movement in the direction intersecting the moving direction.

[0083] In this way, by configuring so that edamame that have already been judged are not judged again, it is possible to accurately judge the ratio of good and defective edamame to be inspected, the reason for judging them as defective, and the number of judged edamame to be inspected can be accurately grasped. The log data can also be transferred to an external server device via a network, and in this case, the log information can be checked at any time from an external terminal. In addition, machine learning (particularly deep learning) can be performed based on the log information to build an artificial intelligence (AI) of a trained model. That is, with the log information as the input layer, the computer algorithm can derive features from the data of this input layer, and in the intermediate layer, a network structure showing the connections between the features can be built, and finally output information (judgment criteria) that becomes the output layer can be built. Furthermore, by configuring the log information so that it can be accessed from an external terminal via the Internet, the selection status of edamame based on the latest log information can be checked from a remote location.

[0084] <Checking the criteria> In addition, the pod and fruit sorting system and pod and fruit sorting device according to the present embodiment can be configured to perform more accurate sorting by adjusting the judging criteria for soybeans in detail. That is, the image captured by the photographing means is stored, and the image is judged later to confirm whether the criteria are appropriate or not. Therefore, it is preferable to configure the system to save the photographed image of the object (soybean) once inspected, as shown in the saved image in FIG. 16(B), as opposed to the setting screen in FIG. 16(A). It is possible to save such photographed images indefinitely, but it is also possible to configure the system to save the latest image by overwriting it by setting the number of images to be saved and the storage capacity (file size) in advance.

[0085] By applying the judging algorithm to the saved images later, the actual data (numerical data, etc.) at the time of judging can be obtained, and by checking the validity of the values, the accuracy of the judgment can be confirmed.The accuracy of the edamame sorting can be improved by adjusting the judgment criteria, camera direction, lighting settings, etc.

[0086] In particular, when judging whether edamame beans are good or bad, it is desirable to set the camera and lighting settings and the judgment criteria each time, since harvested edamame beans come in various shapes. At this time, by setting the judgment criteria and the like from the result log based on images that have already been judged, it is possible to set optimal sorting conditions according to the situation. Therefore, saving the captured images is particularly effective in sorting edamame beans, which are natural products.

[0087] <Device settings> It is desirable to configure the pod and fruit sorting system and the pod and fruit sorting device according to the present embodiment so that the setup work at the time of installation is easy. For this reason, it is desirable for the pod and fruit sorting system and the pod and fruit sorting device according to the present embodiment to have a camera correction function and an equipment setup function based on the already set up pod and fruit sorting system and the pod and fruit sorting device.

[0088] Fig. 17(A) shows an actual inspection image before camera correction, and Fig. 17(B) shows an inspection image after camera correction. In particular, in this embodiment, a camera with a wide-angle lens is used, so distortion occurs in the captured image as it moves away from the center. Therefore, by using camera correction, it is possible to eliminate the distortion of the image and obtain an accurate inspection image. In this embodiment, the following calculations can be performed.

[0089] Correction amount = 1 + |(Edamame center - camera center)| ÷ correction coefficient Width after correction = Width ÷ Correction amount The correction coefficient can be a value specific to the camera, and by performing the camera correction on an inspection image in which individual soybeans are extracted, an accurate inspection image can be obtained. Then, by applying the judgment criteria to the accurate inspection image, the accuracy of soybean selection can be improved.

[0090] Also, FIG. 18 shows a setting screen for setting up another pod and fruit sorting system and pod and fruit sorting device based on the already set pod and fruit sorting system and pod and fruit sorting device. FIG. 18(A) shows the values ​​before the settings are adjusted, and FIG. 18(B) shows the values ​​after adjustment by the calibration setting. In particular, FIG. 18 uses the values ​​of a pod and fruit sorting device (standard device) that is already in operation or properly set, and is configured to calibrate, adjust, etc. the settings of a newly installed pod and fruit sorting device based on this. This figure shows a screen for adjusting the focus in order to set the focus of the captured image to the optimum setting. The R (red), G (green), and B (blue) values ​​are obtained from the standard device as the reference "focus value (reference value)", and the camera is adjusted so that the difference (focus value (Δ)) between the current R (red), G (green), and B (blue) values, which are the "focus value (current value)", falls within the range of the reference value. If the current value is within the range of the reference values ​​as a result of the camera adjustment, "OK" is displayed as the adjustment result.

[0091] The settings of the pod and fruit sorting system and pod and fruit sorting device can include, in addition to the focus calibration, brightness calibration of the captured image, tilt calibration of the captured image, and dimensional calibration of the captured image.

[0092] Calibration using such a standard machine can greatly simplify the setup work when setting up the pod and fruit sorting system and pod and fruit sorting device. [Explanation of symbols]

[0093] 10 Hopper section 20 First conveying means 21,31 Shooting method 22,32 Transport detection means 23,33 Ejection means 30 Second conveying means 31 Filming Method 40 Edamame sorting system W Edamame

Claims

1. A fruit sorting system for inspecting or sorting fruit with pods by analyzing an image of the fruit with pods, The apparatus includes an image acquisition means for acquiring an image captured by the image capture means, and an image analysis means for analyzing the image acquired by the image acquisition means, The image analysis means a pod image extraction unit for extracting images of individual pods from the acquired image; a shrinkage processing execution unit that executes a shrinkage processing to reduce a contour area of ​​the extracted individual pod and fruit image, This shrinkage process is characterized in that the number of pixels in the contour area of ​​the image of the pod or fruit to be removed in one shrinkage process is set, and the number of pixels to be removed is adjusted by performing this process multiple times, in this pod or fruit sorting system.

2. Furthermore, the area of ​​the region discolored as black spots is calculated for the image that has been subjected to the shrinking process, and it is determined whether or not this area is within a reference range. The pod and fruit sorting system of claim 1, wherein upper and lower limits of brightness values ​​for determining that the areas discolored as black spots are black spots are set for at least one of R (red), G (green), and B (blue) in the image, and anything within said range is considered to be a black spot.

3. The fruit pod sorting system according to claim 1 or 2, further comprising a log information storage unit that stores analysis results from the image analysis means as log information.

4. The pod and fruit sorting system according to any one of claims 1 to 3, wherein the image analysis means calculates a movement position for the image acquired by the image acquisition means, and excludes images of pods and fruit with pods present at the movement position from the analysis target.

5. Furthermore, the image analysis means a pod image extraction unit for extracting images of individual pods from the acquired image; A pod / fruit sorting system as described in any one of claims 1 to 4, which divides the extracted images of each pod / fruit into multiple equal parts and calculates the area ratio between the base end and the tip end in the longitudinal direction to determine whether or not there are any chips in the pod / fruit.

Images