Tomato incoming material whole-process unmanned processing system and method based on AI
The AI-based unmanned tomato processing system solves the problems of low efficiency and poor quality in traditional tomato processing, enabling precise setting of washing time and accurate identification of damaged fruit, thus improving processing efficiency and quality.
Patent Information
- Application Number
- CN202511369908.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-24
- Publication Date
- 2026-01-23
AI Technical Summary
Traditional tomato processing relies on manual operation, which suffers from low efficiency, strong subjectivity, lack of data transparency, insufficient representativeness of manual sampling, and difficulty in ensuring the impartiality of test results.
An AI-based unmanned tomato processing system is adopted, which includes a cleaning mechanism, an evaluation mechanism, and a sorting mechanism. The system uses image processing technology to determine the cleaning time, evaluate regional growth and defective fruits, and achieve automated sorting.
It improves the efficiency and quality of tomato processing, automates the washing process, determines the washing time through image processing technology, accurately segments tomato images and evaluates defective fruits, and achieves fully automated processing of incoming tomatoes.
Smart Images
Figure CN121369722A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of tomato processing technology, and more specifically, to an AI-based unmanned tomato processing system and method for the entire process of receiving tomatoes. Background Technology
[0002] Tomatoes are an important economic crop in Xinjiang, and their processing quality directly affects the market competitiveness of downstream products and the profitability of enterprises. However, in traditional tomato procurement, the washing, identification, and sorting of tomatoes mainly rely on manual operations, which suffers from low efficiency, strong subjectivity, and lack of data transparency. Manual sampling is easily affected by operational standardization, resulting in insufficient sample representativeness; the testing process relies on experience-based judgment, making it difficult to guarantee the impartiality of the results. Summary of the Invention
[0003] This invention provides an AI-based unmanned tomato processing system for the entire process, addressing the problems of low efficiency and poor quality in existing tomato processing technologies. The system includes:
[0004] The cleaning mechanism acquires images of the tomatoes to be processed and determines the cleaning time based on the changes in the images during the cleaning process. The evaluation mechanism acquires grayscale images of the tomatoes after cleaning, performs region growing on the grayscale images, and evaluates the segmented tomato images based on the region growing results. The sorting mechanism identifies bad and good tomatoes based on the bad fruit evaluation results and sorts and weighs the bad and good tomatoes.
[0005] Further, the cleaning mechanism determines the tomato cleaning time based on the changes in the tomato image to be processed during the cleaning process, including: performing grayscale processing on the tomato image to be processed to obtain a grayscale image of the tomato to be processed; obtaining a preset standard grayscale value for tomato pixels, statistically analyzing the grayscale values of pixels in the grayscale image of the tomato to be processed, and calculating the difference between the grayscale value of each pixel in the grayscale image of the tomato to be processed and the standard grayscale value; selecting pixels with a grayscale difference less than a first preset threshold, and calculating the ratio of pixels with a grayscale difference less than the first preset threshold to all pixels in the grayscale image of the tomato to be processed; obtaining the changes in the pixel ratio during the cleaning process, and plotting a pixel change curve based on the changes in the pixel ratio during the cleaning process; performing curve fitting on the pixel change curve to obtain a pixel prediction curve; obtaining a preset standard pixel threshold, calculating the time required for the pixel ratio in the pixel prediction curve to reach the preset standard pixel threshold, and determining the tomato cleaning time based on the time required for the pixel ratio to reach the preset standard pixel threshold.
[0006] Further, determining the tomato washing time based on the time required for the pixel ratio to reach a preset standard pixel threshold includes: obtaining a preset standard required time ratio; calculating the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio; determining whether the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is greater than a second preset threshold; if the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is greater than the second preset threshold, then setting a preset first time as the tomato washing time; if the pixel ratio reaches the preset standard threshold... If the difference between the time required to reach the quasi-pixel threshold and the ratio of the preset standard required time is less than or equal to the second preset threshold, then it is determined whether the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the ratio of the preset standard required time is greater than the third preset threshold. If the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the ratio of the preset standard required time is greater than the third preset threshold, then the preset second time is set as the tomato washing time. If the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the ratio of the preset standard required time is less than or equal to the third preset threshold, then the preset third time is set as the tomato washing time.
[0007] Furthermore, the evaluation agency performs region growing on the cleaned tomato grayscale image, including: clustering each pixel according to its grayscale value in the cleaned tomato grayscale image, and setting seed points according to the cluster centers of the clustering results; traversing the seed points and recording the positions of the pixels in the eight-connected regions of the seed points; traversing the pixels to be tested in the eight-connected regions and determining whether the grayscale difference between the pixels to be tested and the seed points is within a preset range; if the grayscale difference between the pixels to be tested and the seed points is within the preset range, then setting the pixels to be tested as new growth seed points; using the new growth seed points as the center, continuing to detect new pixels to be tested until the region can no longer grow, thus completing the region growing of the cleaned tomato grayscale image.
[0008] Further, the step of clustering pixels based on their grayscale values in the cleaned tomato grayscale image includes: establishing a sample dataset based on the grayscale values of pixels in the cleaned tomato grayscale image; randomly selecting k initial cluster centers from the sample dataset; calculating the Manhattan distance from the sample data in the sample dataset to the initial cluster centers; dividing each pixel into its corresponding cluster partition based on the Manhattan distance from the sample data in the sample dataset to the initial cluster centers; calculating the mean of the sample data of pixels within each cluster partition; recalculating the cluster centers based on the mean of the sample data of pixels within each cluster partition; repeating the above steps iteratively until the cluster centers no longer change or the number of iterations reaches the preset maximum number of iterations, thereby obtaining the clustering result of the pixels.
[0009] Furthermore, the evaluation agency evaluates the segmented tomato images for defective fruit based on the region growth results, including: segmenting the impurity region and tomato region of the washed tomato grayscale image based on the region growth results; extracting each tomato image in the tomato region based on an AI vision algorithm; determining the edges of each tomato based on each tomato image; performing circle fitting detection on each tomato edge; determining a first evaluation parameter based on the circle fitting degree of the tomato edges; statistically analyzing the difference between the grayscale value of a pixel in the tomato image and the standard grayscale value; selecting pixels with a grayscale value difference less than a fourth preset threshold; calculating the ratio of pixels with a grayscale value difference less than the fourth preset threshold to all pixels in the tomato image to obtain a second evaluation parameter; determining a comprehensive evaluation parameter based on the first evaluation parameter and the second evaluation parameter; and evaluating the tomato images for defective fruit based on the comprehensive evaluation parameter.
[0010] Further, determining the comprehensive evaluation parameter based on the first evaluation parameter and the second evaluation parameter includes: determining the comprehensive evaluation parameter according to the comprehensive evaluation parameter calculation formula, wherein the comprehensive evaluation parameter is,
[0011] E t =αE1+βE2
[0012] Among them, E t The parameters are used for comprehensive evaluation. E1 is the first evaluation parameter, E2 is the second evaluation parameter, α is the preset first weight, β is the preset second weight, and α+β=1.
[0013] Furthermore, the step of evaluating the tomato image as a defective fruit based on the comprehensive evaluation parameters includes: obtaining preset standard evaluation parameters and calculating the difference between the comprehensive evaluation parameters of the tomato image and the preset standard evaluation parameters;
[0014] Determine whether the comprehensive evaluation parameter and the preset standard evaluation parameter are greater than the fifth preset threshold. If the comprehensive evaluation parameter and the preset standard evaluation parameter are greater than the fifth preset threshold, then the tomato corresponding to the tomato image is identified as a good fruit; if the comprehensive evaluation parameter and the preset standard evaluation parameter are less than or equal to the fifth preset threshold, then the tomato corresponding to the tomato image is identified as a bad fruit.
[0015] To achieve the above objectives, the present invention also provides an AI-based fully automated tomato processing method, comprising:
[0016] Acquire images of tomatoes to be processed, and determine the washing time of tomatoes based on the changes in the images during the washing process; acquire grayscale images of tomatoes after washing, perform region growing on the grayscale images of tomatoes after washing, and evaluate the bad fruits of the segmented tomato images based on the region growing results; identify bad and good tomatoes based on the bad fruit evaluation results, and sort and weigh the bad and good tomatoes.
[0017] The beneficial effects of this invention are as follows:
[0018] By applying the above technical solutions, this invention effectively improves cleaning efficiency and results by extracting tomato images to be processed and setting the cleaning time based on the cleaning changes of the tomato images. By performing region growing on the grayscale images of the cleaned tomatoes, the tomato images are accurately segmented and bad fruits are evaluated. This achieves fully automated processing of the entire process of cleaning, identifying, and sorting incoming tomatoes, improving the efficiency and quality of incoming tomato processing. Attached Figure Description
[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0020] Figure 1 The diagram shows the overall structure of an AI-based unmanned tomato processing system according to an embodiment of the present invention.
[0021] Figure 2 The diagram shows a flowchart of an AI-based unmanned tomato processing method proposed in an embodiment of the present invention. Detailed Implementation
[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.
[0023] This application provides an AI-based unmanned tomato processing system for the entire process of receiving and processing tomatoes, such as... Figure 1 As shown, it includes:
[0024] The cleaning mechanism acquires images of the tomatoes to be processed and determines the cleaning time based on the changes in the images during the cleaning process. The evaluation mechanism acquires grayscale images of the tomatoes after cleaning, performs region growing on the grayscale images, and evaluates the segmented tomato images based on the region growing results. The sorting mechanism identifies bad and good tomatoes based on the bad fruit evaluation results and sorts and weighs the bad and good tomatoes.
[0025] In this embodiment, after receiving the tomatoes, the tomatoes are transported to the cleaning mechanism for cleaning and filtration. The cleaning time is set according to the changes in the tomato images to be processed during the cleaning process. After cleaning, the tomatoes are transported to the conveyor belt. The evaluation mechanism collects grayscale images of the cleaned tomatoes on the conveyor belt, segments the tomato image of each tomato based on the region growing algorithm, and evaluates the bad fruits. The sorting mechanism selects and sorts the good and bad fruits, and after weighing, they are transported to different production lines.
[0026] In some embodiments of this application, the cleaning mechanism determines the tomato cleaning time based on the changes in the tomato image to be processed during the cleaning process, including: performing grayscale processing on the tomato image to be processed to obtain a grayscale image of the tomato to be processed; obtaining a preset standard grayscale value for tomato pixels, statistically analyzing the grayscale values of pixels in the grayscale image of the tomato to be processed, and calculating the difference between the grayscale value of each pixel in the grayscale image of the tomato to be processed and the standard grayscale value; selecting pixels with a grayscale difference less than a first preset threshold, and calculating the ratio of pixels with a grayscale difference less than the first preset threshold to all pixels in the grayscale image of the tomato to be processed; obtaining the changes in the pixel ratio during the cleaning process, and plotting a pixel change curve based on the changes in the pixel ratio during the cleaning process; performing curve fitting on the pixel change curve to obtain a pixel prediction curve; obtaining a preset standard pixel threshold, calculating the time required for the pixel ratio in the pixel prediction curve to reach the preset standard pixel threshold, and determining the tomato cleaning time based on the time required for the pixel ratio to reach the preset standard pixel threshold.
[0027] In this embodiment, a standard grayscale value for tomato pixels is set based on historical experience. The standard grayscale value is the grayscale value of the pixels in the tomato grayscale image after cleaning. The proportion of pixels in the tomato grayscale image to be processed that are similar to the standard grayscale value is selected. The pixel change curve is plotted based on the change of the pixel proportion during the cleaning process. The pixel change curve is fitted based on the least squares method. The tomato cleaning time is set by the time required for the pixel ratio in the fitted pixel prediction curve to reach the preset standard pixel threshold, so as to achieve precise setting of the cleaning time.
[0028] In some embodiments of this application, determining the tomato washing time based on the time required for the pixel ratio to reach a preset standard pixel threshold includes: obtaining a preset standard required time ratio; calculating the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio; determining whether the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is greater than a second preset threshold; if the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is greater than the second preset threshold, then setting a preset first time as the tomato washing time; if the pixel ratio reaches... If the difference between the time required to reach the preset standard pixel threshold and the ratio of the preset standard required time is less than or equal to the second preset threshold, then it is determined whether the difference between the time required to reach the preset standard pixel threshold and the ratio of the preset standard required time is greater than the third preset threshold. If the difference between the time required to reach the preset standard pixel threshold and the ratio of the preset standard required time is greater than the third preset threshold, then the preset second time is set as the tomato washing time. If the difference between the time required to reach the preset standard pixel threshold and the ratio of the preset standard required time is less than or equal to the third preset threshold, then the preset third time is set as the tomato washing time.
[0029] In this embodiment, the tomato washing time is set by the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time. The larger the difference, the more impurities there are in the tomato feed, and the longer the corresponding washing time.
[0030] In some embodiments of this application, the evaluation agency performs region growing on the cleaned tomato grayscale image, including: clustering each pixel according to the grayscale value of the pixels in the cleaned tomato grayscale image, and setting seed points according to the cluster centers of the clustering results; traversing the seed points and recording the positions of the pixels in the eight-connected regions of the seed points; traversing the pixels to be tested in the eight-connected regions and determining whether the grayscale difference between the pixels to be tested and the seed points is within a preset range; if the grayscale difference between the pixels to be tested and the seed points is within the preset range, then setting the pixels to be tested as new growth seed points; using the new growth seed points as the center, continuing to detect new pixels to be tested until the region can no longer grow, thus completing the region growing of the cleaned tomato grayscale image.
[0031] In this embodiment, by performing region growing on the seed points in the cleaned tomato grayscale image, the tomato region and the background region are accurately segmented.
[0032] In some embodiments of this application, the step of clustering pixels based on the grayscale values of pixels in the cleaned tomato grayscale image includes: establishing a sample dataset based on the grayscale values of pixels in the cleaned tomato grayscale image; randomly selecting k initial cluster centers in the sample dataset; calculating the Manhattan distance from the sample data in the sample dataset to the initial cluster centers; dividing each pixel into a corresponding cluster partition based on the Manhattan distance from the sample data in the sample dataset to the initial cluster centers; calculating the mean of the sample data of pixels in each cluster partition; recalculating the cluster centers based on the mean of the sample data of pixels in each cluster partition; repeating the above steps iteratively until the cluster centers no longer change or the number of iterations reaches a preset maximum number of iterations, thereby obtaining the clustering result of the pixels.
[0033] In this embodiment, the pixels in the cleaned tomato grayscale image are clustered based on the k-means clustering algorithm. In this embodiment, the k value is set to 4, and two seed points close to the tomato grayscale value and two seed points close to the background grayscale value are selected through clustering.
[0034] In some embodiments of this application, the evaluation agency evaluates the segmented tomato images as defective based on region growth results, including: segmenting the impurity region and tomato region of the washed tomato grayscale image based on the region growth results; extracting each tomato image in the tomato region based on an AI vision algorithm; determining the edges of each tomato based on each tomato image; performing circle fitting detection on each tomato edge; determining a first evaluation parameter based on the circle fitting degree of the tomato edges; statistically analyzing the difference between the grayscale value of a pixel in the tomato image and the standard grayscale value; selecting pixels with a grayscale value difference less than a fourth preset threshold; calculating the ratio of pixels with a grayscale value difference less than the fourth preset threshold to all pixels in the tomato image to obtain a second evaluation parameter; determining a comprehensive evaluation parameter based on the first evaluation parameter and the second evaluation parameter; and evaluating the tomato images as defective based on the comprehensive evaluation parameter.
[0035] In this embodiment, after extracting the tomato region, the tomato image of each tomato in the tomato region is extracted by AI vision algorithm. The edge circle fitting degree of the tomato image is used as the first evaluation parameter, and the second evaluation parameter is calculated by the gray value of the tomato image, so as to comprehensively calculate the comprehensive evaluation parameter.
[0036] In some embodiments of this application, determining the comprehensive evaluation parameter based on the first evaluation parameter and the second evaluation parameter includes: determining the comprehensive evaluation parameter according to the comprehensive evaluation parameter calculation formula, wherein the comprehensive evaluation parameter is,
[0037] E t =αE1+βE2
[0038] Among them, E tThe parameters are used for comprehensive evaluation. E1 is the first evaluation parameter, E2 is the second evaluation parameter, α is the preset first weight, β is the preset second weight, and α+β=1.
[0039] In this embodiment, the first evaluation parameter and the second evaluation parameter are weighted and summed using the first weight α and the second weight β to achieve a comprehensive evaluation of the tomato image.
[0040] In some embodiments of this application, the step of evaluating the tomato image as a bad fruit based on the comprehensive evaluation parameters includes: obtaining preset standard evaluation parameters; calculating the difference between the comprehensive evaluation parameters of the tomato image and the preset standard evaluation parameters; determining whether the comprehensive evaluation parameters and the preset standard evaluation parameters are greater than a fifth preset threshold; if the comprehensive evaluation parameters and the preset standard evaluation parameters are greater than the fifth preset threshold, then the tomato corresponding to the tomato image is identified as a good fruit; if the comprehensive evaluation parameters and the preset standard evaluation parameters are less than or equal to the fifth preset threshold, then the tomato corresponding to the tomato image is identified as a bad fruit.
[0041] In this embodiment, the difference between the comprehensive evaluation parameters of the tomato image and the preset standard evaluation parameters is used to evaluate bad fruits, which facilitates the accurate sorting of tomatoes.
[0042] To achieve the above objectives, the present invention also provides an AI-based fully automated tomato processing method, comprising:
[0043] S101, Obtain the tomato image to be processed, and determine the tomato cleaning time based on the changes in the tomato image to be processed during the cleaning process;
[0044] S102, acquire the grayscale image of the cleaned tomato, perform region growing on the grayscale image of the cleaned tomato, and evaluate the bad fruit in the segmented tomato image based on the region growing results.
[0045] S103, based on the evaluation results of the damaged fruit, identify the damaged and good tomatoes, and sort and weigh the damaged and good tomatoes.
[0046] By applying the above technical solutions, this invention utilizes a cleaning mechanism to acquire images of tomatoes to be processed and determines the cleaning time based on changes in the images during the cleaning process; an evaluation mechanism to acquire grayscale images of the tomatoes after cleaning, performs region growing on these images, and evaluates the segmented tomatoes as defective based on the region growing results; and a sorting mechanism to identify defective and good tomatoes based on the defective fruit evaluation results, and sorts and weighs the defective and good tomatoes. Through fully automated processing of incoming tomatoes—cleaning, identification, and sorting—the efficiency and quality of incoming tomato processing are improved.
[0047] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. An AI-based fully automated tomato processing system, characterized in that, include: The cleaning mechanism is used to acquire images of tomatoes to be processed and to determine the tomato cleaning time based on the changes in the images of tomatoes to be processed during the cleaning process. The evaluation agency is used to acquire grayscale images of cleaned tomatoes, perform region growing on the grayscale images of cleaned tomatoes, and evaluate the bad fruit in the segmented tomato images based on the region growing results. The sorting mechanism is used to identify bad tomatoes and good tomatoes based on the evaluation results, and to sort and weigh the bad tomatoes and good tomatoes.
2. The AI-based unmanned tomato processing system for the entire process of receiving materials as described in claim 1, characterized in that, The cleaning mechanism determines the tomato cleaning time based on the changes in the tomato image to be processed during the cleaning process, including: The tomato image to be processed is converted to grayscale to obtain the grayscale image of the tomato to be processed. Obtain the standard grayscale value of the preset tomato pixel, count the grayscale values of the pixels in the tomato grayscale image to be processed, and calculate the difference between the grayscale value of each pixel in the tomato grayscale image to be processed and the standard grayscale value. Pixels with grayscale differences less than the first preset threshold are selected, and the ratio of the pixels with grayscale differences less than the first preset threshold to all pixels in the tomato grayscale image to be processed is calculated. Obtain the changes in pixel ratio during the cleaning process, and plot the pixel change curve based on the changes in pixel ratio during the cleaning process; By performing curve fitting on the pixel change curve, the pixel prediction curve is obtained; Obtain a preset standard pixel threshold, calculate the time required for the pixel ratio in the pixel prediction curve to reach the preset standard pixel threshold, and determine the tomato washing time based on the time required for the pixel ratio to reach the preset standard pixel threshold.
3. The AI-based unmanned tomato processing system for the entire process of receiving materials as described in claim 2, characterized in that, Determining the tomato washing time based on the time required for the pixel ratio to reach a preset standard pixel threshold includes: Obtain the time ratio required for the preset standard, and calculate the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the time ratio required for the preset standard. Determine whether the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is greater than the second preset threshold. If the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is greater than the second preset threshold, then set the preset first time as the tomato washing time. If the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is less than or equal to the second preset threshold, then it is determined whether the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is greater than the third preset threshold. If the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is greater than the third preset threshold, then the preset second time is set as the tomato washing time. If the difference between the time required for the pixel ratio to reach the preset standard pixel threshold and the preset standard required time ratio is less than or equal to the third preset threshold, then the preset third time is set as the tomato washing time.
4. The AI-based unmanned tomato processing system for the entire process of receiving materials as described in claim 1, characterized in that, The evaluation agency performs region growing on the cleaned tomato grayscale image, including: Based on the gray values of the pixels in the cleaned tomato grayscale image, each pixel is clustered, and seed points are set according to the cluster centers of the clustering results. Traverse the seed point and record the position of the pixel in the 8-connected region of the seed point; Traverse the pixels to be tested within the 8-connected region and determine whether the grayscale difference between the pixel to be tested and the seed point is within the preset range. If the grayscale difference between the pixel to be tested and the seed point is within the preset range, then the pixel to be tested will be set as the new seed point for growth. Centered on the new growth seed point, continue to detect new pixels to be tested until the region can no longer grow, thus completing the region growth of the cleaned tomato grayscale image.
5. The AI-based unmanned tomato processing system for the entire process according to claim 4, characterized in that, The step of clustering pixels based on their grayscale values in the cleaned tomato grayscale image includes: A sample dataset is established based on the gray values of pixels in the cleaned tomato grayscale image, and k initial cluster centers are randomly selected from the sample dataset. Calculate the Manhattan distance from the sample data in the sample dataset to the initial cluster center, and divide each pixel into the corresponding cluster partition based on the Manhattan distance from the sample data in the sample dataset to the initial cluster center; Calculate the mean value of the sample data of pixels within each cluster partition, and recalculate the cluster centers based on the mean value of the sample data of pixels within each cluster partition; Repeat the above steps until the cluster centers no longer change or the number of iterations reaches the preset maximum number of iterations, and obtain the clustering results of the pixels.
6. The AI-based unmanned tomato processing system for the entire process of receiving materials as described in claim 4, characterized in that, The evaluation agency evaluates the segmented tomato images for defective fruit based on the region growth results, including: Based on the region growth results, the impurity region and tomato region of the washed tomato grayscale image are segmented, and each tomato image in the tomato region is extracted based on AI vision algorithm; The edges of each tomato are determined based on each tomato image, and a circle fitting detection is performed on each tomato edge. The first evaluation parameter is determined based on the circle fitting degree of the tomato edge. The difference between the gray values of pixels in the tomato image and the standard gray values is statistically analyzed. Pixels with gray value differences less than the fourth preset threshold are selected. The ratio of the gray value differences less than the fourth preset threshold to all pixels in the tomato image is calculated to obtain the second evaluation parameter. The comprehensive evaluation parameters are determined based on the first and second evaluation parameters, and the tomato images are evaluated for defective fruit based on the comprehensive evaluation parameters.
7. The AI-based unmanned tomato processing system for the entire process of receiving materials as described in claim 6, characterized in that, The determination of the comprehensive evaluation parameters based on the first evaluation parameter and the second evaluation parameter includes: The comprehensive evaluation parameters are determined according to the comprehensive evaluation parameter calculation formula, wherein the comprehensive evaluation parameters are: E t =αE1+βE2 Among them, E t The parameters are used for comprehensive evaluation. E1 is the first evaluation parameter, E2 is the second evaluation parameter, α is the preset first weight, β is the preset second weight, and α+β=1.
8. The AI-based unmanned tomato processing system for the entire process of receiving materials as described in claim 6, characterized in that, The evaluation of defective tomato images based on comprehensive evaluation parameters includes: Obtain preset standard evaluation parameters and calculate the difference between the comprehensive evaluation parameters of the tomato image and the preset standard evaluation parameters; Determine whether the comprehensive evaluation parameter and the preset standard evaluation parameter are greater than the fifth preset threshold. If the comprehensive evaluation parameter and the preset standard evaluation parameter are greater than the fifth preset threshold, then the tomato corresponding to the tomato image is identified as a good fruit. If the comprehensive evaluation parameter and the preset standard evaluation parameter are less than or equal to the fifth preset threshold, then the tomato corresponding to the tomato image will be identified as a bad fruit.
9. A fully automated, AI-based method for handling incoming tomato feed, characterized in that: include: Acquire images of tomatoes to be processed, and determine the tomato washing time based on the changes in the images of tomatoes to be processed during the washing process; Obtain the grayscale image of the cleaned tomato, perform region growing on the grayscale image of the cleaned tomato, and evaluate the bad fruit in the segmented tomato image based on the region growing results. Based on the evaluation results of the damaged fruit, damaged and good tomatoes were identified, and the damaged and good tomatoes were sorted and weighed.