Intelligent sorting system for vegetable and fruit goods

Abstract

The invention is applicable to the technical field of automatic detection and sorting of vegetables and fruits, and provides an intelligent sorting system for vegetable and fruit goods, which comprises a data acquisition module used for acquiring an acquired image of pitaya to be detected through polarized light imaging, determining an initial abnormal index of a spot area in the acquired image, and sending the initial abnormal index to a server; meanwhile, acquiring surface attributes of a specific area of the acquired image; and the image analysis module is used for analyzing an acquired image obtained by polarized light imaging based on a machine vision technology, judging whether a liquid drop phenomenon exists in the spot region or not, and determining a corresponding liquid drop parameter if the liquid drop phenomenon exists in the spot region. According to the method, deep learning and polarized light imaging technologies are adopted, machine vision is combined, a spot area in a pitaya collection image is subjected to fine analysis, the influence of liquid drops on spot reflection components is dynamically detected, and therefore accurate correction of the initial abnormal index is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of automatic detection and sorting of fruits and vegetables, and particularly relates to an intelligent goods sorting system for fruits and vegetables. Background Art

[0002] Currently, in the field of intelligent goods sorting of fruits and vegetables, it mainly relies on traditional machine vision technology. By collecting the overall reflection characteristics of pitayas, an abnormal index is generated to judge the abnormal conditions in the spotted area. This method usually makes judgments based on fixed thresholds and overall image reflection parameters, ignoring the differences in subtle local features in the spotted area of pitayas. Due to the natural color and texture variations on the surface of pitayas themselves, combined with the influence of environmental factors such as lighting, angle, and equipment parameters during the collection process, the overall reflection parameters often cannot accurately reflect the true abnormal situation in the local spotted area.

[0003] More critically, the traditional method fails to fully consider the influence of tiny water droplets on the reflection components. After the droplets form on the surface of the pitaya, they will change the local reflection characteristics, resulting in an offset of the reflection parameters. The fixed threshold cannot dynamically adapt to this change, thus causing a certain deviation in the calculation of the abnormal index. This defect is particularly prominent in practical applications, which may lead to some abnormal spots not being accurately identified, thereby affecting the overall detection accuracy of the sorting system and the product quality.

[0004] Therefore, the existing technology is difficult to achieve a fine analysis of the abnormal situation in the spotted area of pitayas when facing complex environments and variable conditions. There is an urgent need for a new method to dynamically capture the local reflection characteristics and accurately quantify the influence of factors such as droplets to make up for the deficiencies of the traditional fixed threshold method. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent goods sorting system for fruits and vegetables, aiming to solve the problems raised in the background art.

[0006] The present invention is implemented as follows. An intelligent goods sorting system for fruits and vegetables, the system includes:

[0007] A data acquisition module, configured to obtain a captured image of the to-be-inspected pitaya through polarized light imaging, determine the initial abnormal index of the spotted area in the captured image, and simultaneously obtain the surface attributes of a specific area of the captured image;

[0008] An image analysis module, configured to analyze the captured image obtained by polarized light imaging based on machine vision technology, and judge whether there is a droplet phenomenon in the spotted area. If so, determine the corresponding droplet parameters;

[0009] A model application module, which is used to retrieve a predefined reference model, first input the surface properties of a specific area into it to obtain an initial reference image, and then input the surface properties of the specific area and droplet parameters into it simultaneously to obtain an additional reference image;

[0010] An abnormal index correction module, which is used to determine the difference amount of the reflection components in the initial reference image and the additional reference image, and convert it into a correction coefficient to correct the initial abnormal index.

[0011] As a further limitation of the technical solution of the embodiment of the present invention, the image analysis module specifically includes:

[0012] An image preprocessing unit, which is used to preprocess the acquired image obtained by polarized light imaging by using machine vision technology, including image denoising, enhancement, and edge detection, so as to highlight the detailed features of the speckle area and eliminate background interference;

[0013] A droplet phenomenon judgment unit, which is used to analyze the speckle area based on the preprocessed acquired image by applying feature extraction and pattern recognition algorithms, judge whether there is a droplet phenomenon by analyzing the color, texture, and reflected light characteristics, and confirm the existence of droplets by using a set threshold;

[0014] A droplet evaluation unit, which is used to quantitatively evaluate the droplets in the speckle area after confirming the existence of the droplet phenomenon, and extract the droplet density and thickness parameters.

[0015] As a further limitation of the technical solution of the embodiment of the present invention, the droplet evaluation unit specifically includes:

[0016] A droplet contour extraction sub-unit, which is used to accurately segment the droplet candidate area for the speckle area in the preprocessed acquired image by using an adaptive threshold segmentation or watershed algorithm combined with morphological operations to realize the extraction of the droplet contour;

[0017] A droplet density determination sub-unit, which is used to perform shape analysis on the segmented droplets by using the Canny edge detection and Hough circle transformation methods, identify and measure the edges and diameters of the droplets to determine the size of the droplets, and determine the droplet density by statistically calculating the ratio of the number of droplets in the speckle area to the area of the region;

[0018] A droplet thickness calculation sub-unit, which is used to detect the droplets by using a surface curvature analysis method based on three-dimensional reconstruction technology to estimate the average thickness of the droplets.

[0019] As a further limitation of the technical solution of the embodiment of the present invention, the predefined reference model refers to an image generation and feature analysis model constructed based on a deep learning algorithm, which can generate a reference image based on the surface attributes of pitayas, generate another reference image by combining the surface attributes of pitayas and droplet parameters, generate corresponding reference reflection component parameters according to different surface attributes of pitayas, and generate corrected reflection component parameters corresponding to the reference images generated after applying different droplet parameters.

[0020] As a further limitation of the technical solution of the embodiment of the present invention, the specific area of the acquired image refers to the annular area outside the spot area in the acquired image to ensure that the surface attributes input to the predefined reference model are consistent with the surface attributes of the spot area.

[0021] As a further limitation of the technical solution of the embodiment of the present invention, the abnormal index correction module specifically includes;

[0022] A reflection component parameter determination unit for respectively determining the reference reflection component parameters and corrected reflection component parameters corresponding to the initial reference image and the additional reference image based on the predefined reference model;

[0023] A correction coefficient acquisition unit for quantifying the difference between the reference reflection component parameters and the corrected reflection component parameters and converting the difference into a correction coefficient;

[0024] An abnormal index correction unit for retrieving a preset correction formula and correcting the initial abnormal index according to the correction coefficient.

[0025] As a further limitation of the technical solution of the embodiment of the present invention, the preset correction formula is: , where refers to the corrected abnormal index, refers to the initial abnormal index, refers to the corrected reflection component parameter, refers to the reference reflection component parameter, refers to the difference amount of the reflection components in the initial reference image and the additional reference image, that is, the correction coefficient, refers to the adjustment coefficient of the correction coefficient.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] The present invention adopts deep learning and polarized light imaging technology, combines machine vision, and finely analyzes the spotted areas in the collected images of pitayas, dynamically detecting the influence of droplets on the reflected components of the spots, thereby realizing the precise correction of the initial anomaly index. This technology overcomes the problem that traditional methods only rely on overall reflection parameters and cannot accurately reflect the abnormal conditions of the spots. By comparing the differences in the reflected components of the spotted areas in the initial reference image and the additional reference image, it effectively quantifies the influence of droplets and generates a correction coefficient. After implementing the present invention, the sorting system can significantly improve the detection accuracy and robustness of pitaya spot anomalies, ensure product quality, reduce losses caused by detection errors, and provide a reliable and advanced technical solution for agricultural intelligent sorting. Description of the Drawings

[0028] Figure 1 It is the application architecture diagram of the system provided by the embodiment of the present invention;

[0029] Figure 2 It is the structural block diagram of the image analysis module in the system provided by the embodiment of the present invention;

[0030] Figure 3 It is the structural block diagram of the droplet evaluation unit in the system provided by the embodiment of the present invention;

[0031] Figure 4 It is the structural block diagram of the anomaly index correction module in the system provided by the embodiment of the present invention. Detailed Embodiments

[0032] In order to make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0033] Furthermore, Figure 1 It shows the application architecture diagram of the system provided by the embodiment of the present invention.

[0034] Among them, in another preferred embodiment provided by the present invention, an intelligent sorting system for fruits and vegetables goods includes:

[0035] A data acquisition module 100, configured to obtain a collected image of a to-be-inspected pitaya through polarized light imaging, determine the initial anomaly index of the spotted area in the collected image, and obtain the surface attributes of a specific area of the collected image.

[0036] In an embodiment of the present invention, the data acquisition module 100 adopts polarized light imaging technology. By equipping a high-resolution polarized camera with a specific polarization filter and illuminating with a polarized light source, it captures the acquisition image of the pitaya to be inspected, thereby simultaneously obtaining the intensity and polarization information of the image. This technical means relies on the collaborative work of an optical system, a polarization filter, and image acquisition software, and can effectively capture the information of the microscopic structure and moisture state of the pitaya surface.

[0037] The reason for adopting the polarized light imaging method is mainly that it can effectively suppress environmental light interference and glare, while enhancing the contrast of the target area in the image, providing richer surface details than traditional imaging, which is of great significance for distinguishing natural anomalies and anomalies caused by droplets in the pitaya spot area.

[0038] The initial anomaly index in the acquisition image of the pitaya is obtained by analyzing the color, texture, and reflection characteristics of the spot area based on existing mature machine vision and image processing technologies. This method has been widely used in fields such as agricultural product quality inspection, automatic sorting systems, and food processing.

[0039] The surface attributes of a specific area in the acquisition image include color distribution, texture features, brightness distribution, and reflected light characteristics, and also include polarization parameters (such as polarization angle and degree of polarization) obtained by polarized light imaging. These parameters can reflect the micro-structure and humidity information of the pitaya surface, providing data support for subsequent droplet parameter extraction and anomaly index correction.

[0040] Furthermore, the intelligent sorting system for fruits and vegetables goods further includes:

[0041] An image analysis module 200, which is used to analyze the acquisition image obtained by polarized light imaging based on machine vision technology, and determine whether there is a droplet phenomenon in the spot area. If so, the corresponding droplet parameters are determined.

[0042] Specifically, Figure 2 Fig. shows the structural block diagram of the image analysis module 200 in the system provided by the embodiment of the present invention.

[0043] Among them, in the preferred embodiment provided by the present invention, the image analysis module 200 specifically includes:

[0044] An image preprocessing unit 201, which is used to preprocess the acquisition image obtained by polarized light imaging by using machine vision technology, including image denoising, enhancement, and edge detection, so as to highlight the detailed features of the spot area and eliminate background interference;

[0045] The droplet phenomenon judgment unit 202 is used to analyze the speckle area based on the pre - processed acquired image by applying feature extraction and pattern recognition algorithms, judge whether there is a droplet phenomenon by analyzing color, texture and reflected light characteristics, and confirm the existence of droplets using a set threshold;

[0046] The droplet evaluation unit 203 is used to quantitatively evaluate the droplets in the speckle area after confirming the existence of the droplet phenomenon, and extract droplet density and thickness parameters.

[0047] In the embodiment of the present invention, the image pre - processing unit 201 first performs noise reduction processing on the acquired image obtained by polarized light imaging, and usually uses mature algorithms such as Gaussian filtering or median filtering to remove image noise; then, the overall brightness and contrast of the image are enhanced by histogram equalization or adaptive contrast enhancement technology to further highlight the detailed features of the speckle area; finally, the Canny edge detection algorithm is used to extract the edge information in the image to effectively eliminate background interference and clarify the contour of the speckle area. These technologies are all existing mature machine vision means and have been widely used in the fields of industrial and agricultural detection.

[0048] After receiving the pre - processed image, the droplet phenomenon judgment unit 202 first extracts the image features of the speckle area by means of color space conversion, local texture analysis (such as using local binary pattern or Gabor filter), and reflected light characteristic analysis, etc.; then, pattern recognition is performed on the extracted features according to the set threshold or using a trained deep - learning model to judge whether there is a droplet phenomenon in the speckle area; when a droplet phenomenon is detected, the corresponding droplet parameters are further confirmed and determined. The technical means based on this process also belong to existing mature and realizable technologies and have been successfully applied in fields such as automatic sorting and quality inspection.

[0049] Specifically, Figure 3 The structural block diagram of the droplet evaluation unit 203 in the system provided by the embodiment of the present invention is shown.

[0050] Among them, in the preferred embodiment provided by the present invention, the droplet evaluation unit 203 specifically includes:

[0051] The droplet contour extraction sub - unit 2031 is used to accurately segment the droplet candidate area of the speckle area in the pre - processed acquired image by using adaptive threshold segmentation or watershed algorithm combined with morphological operations to achieve the extraction of the droplet contour;

[0052] The droplet density determination subunit 2032 is used to perform shape analysis on the segmented droplets through the Canny edge detection and Hough circle transformation methods, identify and measure the edges and diameters of the droplets to determine the droplet sizes, and determine the droplet density by statistically calculating the ratio of the number of droplets in the spot area to the area of the region.

[0053] The droplet thickness calculation subunit 2033 is used to detect the droplets by using the surface curvature analysis method based on the three-dimensional reconstruction technology to estimate the average thickness of the droplets.

[0054] In the embodiment of the present invention, the droplet contour extraction subunit 2031 first uses the adaptive threshold segmentation method to preliminarily distinguish the candidate area where droplets may exist in the acquired image from the background, and then further refines the segmentation result through the watershed algorithm to solve the problem of droplet overlap or adhesion. Then, combined with morphological operations (such as dilation, erosion, opening operation, and closing operation), the edges of the candidate area are smoothed and noise is suppressed, so as to accurately extract the droplet contour. The above technologies are all image segmentation methods widely used in the existing industrial and academic circles, and can effectively cope with the challenges of complex textures and noise interference on the surface of pitayas.

[0055] The droplet density determination subunit 2032 uses the Canny edge detection algorithm to carefully extract the edges of the segmented droplets, and then uses the Hough circle transformation method to identify the circular features of the droplets, measure their diameters and edge information to determine the size of each droplet. By statistically calculating the number of droplets in the spot area and calculating its ratio to the total area of the region, the droplet density can be obtained. This process relies on mature edge detection and shape recognition algorithms to ensure that the distribution of droplets on the surface of pitayas can be accurately reflected under different image acquisition conditions.

[0056] The droplet thickness calculation subunit 2033 uses the surface curvature analysis method based on the three-dimensional reconstruction technology, combined with multi-angle image acquisition or structured light technology, to reconstruct the three-dimensional shape of the droplets, analyze the curvature changes on the surface of the droplets, and thus estimate the average thickness of the droplets. By introducing calibration parameters, this subunit can compensate for errors caused by changes in viewing angle, illumination, and droplet shape, and improve the accuracy of thickness measurement. The overall technical solution is based on existing mature three-dimensional imaging and analysis means, providing reliable support for the quantitative evaluation of the thickness of pitaya droplets.

[0057] Furthermore, the intelligent sorting and picking system for fruit and vegetable goods further includes:

[0058] The model application module 300 is used to retrieve a predefined reference model, first input the surface attributes of a specific region into it to obtain an initial reference image, and then input the surface attributes and droplet parameters of the specific region into it simultaneously to obtain an additional reference image.

[0059] The predefined reference model refers to an image generation and feature analysis model constructed based on deep learning algorithms. It can generate reference images based on the surface attributes of pitayas, combine the surface attributes of pitayas with droplet parameters to generate another reference image, generate corresponding reference reflection component parameters according to different surface attributes of pitayas, and generate corrected reflection component parameters corresponding to the reference images generated after applying different droplet parameters.

[0060] The specific area of the collected image refers to the circular area outside the spot area in the collected image to ensure that the surface attributes input to the predefined reference model subsequently are consistent with the surface attributes of the spot area.

[0061] In the embodiments of the present invention, the predefined reference model is constructed based on deep learning algorithms. This model is trained using a large amount of pitaya image data to achieve the functions of image generation and feature analysis. The model can adopt mature technologies such as convolutional neural networks (CNNs) or generative adversarial networks (GANs), and is constructed based on existing deep learning frameworks such as TensorFlow and PyTorch. By collecting and annotating image data of pitayas under different surface attribute conditions, the model can learn the reference reflection component parameters corresponding to different surface attributes; when droplet parameters are input, the model generates corrected reference images reflecting the influence of droplets and outputs the corresponding corrected reflection component parameters.

[0062] The model application module 300 first retrieves the predefined reference model and inputs the surface attributes reflected by the circular area outside the spot area in the collected image into the model to generate an initial reference image, which reflects the reference reflection component parameters of the pitaya without the influence of droplets; subsequently, the same surface attributes and droplet parameters are input into the model simultaneously to generate an additional reference image, which reflects the influence of the presence of droplets on the reflection components. By comparing the reflection component parameters in the initial reference image and the additional reference image, the correction effect of droplets on the surface reflection characteristics of pitayas can be quantitatively calculated.

[0063] Furthermore, the intelligent sorting and grading system for fruit and vegetable products further includes:

[0064] An abnormal index correction module 400, which is used to determine the difference amount of the reflection components in the initial reference image and the additional reference image and convert it into a correction coefficient to correct the initial abnormal index.

[0065] Specifically, Figure 4 shows the structural block diagram of the abnormal index correction module 400 in the system provided by the embodiments of the present invention.

[0066] Among them, in the preferred embodiment provided by the present invention, the abnormal index correction module 400 specifically includes:

[0067] A reflection component parameter determination unit 401 is configured to determine corresponding reference reflection component parameters and corrected reflection component parameters in an initial reference image and an additional reference image respectively based on a predefined reference model;

[0068] A correction coefficient acquisition unit 402 is configured to quantify the difference between the reference reflection component parameters and the corrected reflection component parameters, and convert the difference into a correction coefficient;

[0069] An anomaly index correction unit 403 is configured to retrieve a preset correction formula and correct the initial anomaly index according to the correction coefficient.

[0070] The preset correction formula is: where refers to the corrected anomaly index, refers to the initial anomaly index, refers to the corrected reflection component parameters, refers to the reference reflection component parameters, refers to the difference in the reflection components in the initial reference image and the additional reference image, that is, the correction coefficient, refers to the adjustment coefficient of the correction coefficient.

[0071] In an embodiment of the present invention, in the prior art, generating the initial anomaly index usually only depends on the overall reflection characteristics of the pitaya spot area, without fully considering the influence of tiny water droplets on the reflection components, or unable to accurately quantify this influence, resulting in a possible deviation between the anomaly index and the actual anomaly situation.

[0072] To solve this problem, the anomaly index correction module 400 of the present invention uses a predefined reference model to extract corresponding reflection component parameters from an initial reference image that only reflects the surface properties of the pitaya and an additional reference image that reflects both the surface properties and droplet parameters respectively.

[0073] By comparing the difference between the two, that is, quantifying the change in the reflection components caused by the presence of droplets, this difference is converted into a correction coefficient to dynamically correct the initial anomaly index, so as to more accurately reflect the actual anomaly situation. This method takes advantage of the deep learning model to generate reference images, makes up for the deficiency in the prior art that it is impossible to accurately capture and quantify the influence of tiny water droplets on the reflection characteristics, and significantly improves the scientificity of the anomaly index and the robustness of the system.

[0074] The basic idea of the preset correction formula is to first calculate the relative difference between the reference reflection component parameters extracted from the initial reference image and the corrected reflection component parameters extracted from the additional reference image. This difference reflects the influence of factors such as droplets on the reflection characteristics of pitayas. Then, multiply this relative difference (correction coefficient) by a preset adjustment coefficient to obtain a correction factor, which is then used to dynamically correct the initial anomaly index, so that the corrected anomaly index can more accurately reflect the actual anomaly situation. Such a setting is because the prior art usually fails to fully consider the influence of tiny droplets on the reflection components when generating the initial anomaly index, while this method can intuitively reflect the deviation caused by droplets by quantitatively comparing the reflection parameters under droplet-free and droplet conditions, thus reasonably correcting the anomaly index.

[0075] It should be noted that this correction formula is just a simple and intuitive calculation method, and more advanced and complex calculation methods can be adopted according to requirements in actual applications. For example, a non-linear regression model or a dynamic adjustment algorithm based on neural networks can be used to more finely model and predict the influence of droplets through machine learning means, or methods such as adaptive filtering can be used to update the adjustment coefficient in real time to further improve the accuracy and robustness of anomaly index correction.

[0076] It should be understood that although the steps in the flowcharts of the embodiments of the present invention are shown in sequence according to the indication of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least a part of the steps in each embodiment may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0077] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or an external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0078] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0079] The above-described embodiments merely represent several implementation manners of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent for the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention should be subject to the appended claims.

[0080] The above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent sorting system for fruits and vegetables products, characterized in that, The system includes: A data acquisition module, which is used to obtain a collected image of the pitaya to be inspected through polarized light imaging, determine the initial anomaly index of the spot area in the collected image, and at the same time obtain the surface properties of a specific area of the collected image; An image analysis module, which is used to analyze the collected image obtained by polarized light imaging based on machine vision technology, and judge whether there is a droplet phenomenon in the spot area. If so, determine the corresponding droplet parameters; A model application module, which is used to retrieve a predefined reference model, first input the surface properties of the specific area into it, and obtain an initial reference image, then input the surface properties of the specific area and the droplet parameters into it at the same time, and obtain an additional reference image; An anomaly index correction module, which is used to determine the difference amount of the reflection components in the initial reference image and the additional reference image, and convert it into a correction coefficient to correct the initial anomaly index.

2. The intelligent sorting system for fruits and vegetables goods according to claim 1, wherein The image analysis module specifically includes: An image preprocessing unit, which is used to preprocess the collected image obtained by polarized light imaging by using machine vision technology, including image denoising, enhancement and edge detection, to highlight the detailed features of the spot area and eliminate background interference; A droplet phenomenon judgment unit, which is used to analyze the spot area based on the preprocessed collected image by applying feature extraction and pattern recognition algorithms, judge whether there is a droplet phenomenon by analyzing color, texture and reflected light characteristics, and confirm the existence of droplets by using a set threshold; A droplet evaluation unit, which is used to quantitatively evaluate the droplets in the spot area after confirming the existence of the droplet phenomenon, and extract the droplet density and thickness parameters.

3. The intelligent sorting system for fruits and vegetables goods according to claim 2, characterized in that, The droplet evaluation unit specifically includes: A droplet contour extraction subunit, which is used to accurately segment the droplet candidate area of the spot area in the preprocessed collected image by using an adaptive threshold segmentation or watershed algorithm combined with morphological operations to extract the droplet contour; A droplet density determination subunit, which is used to analyze the shape of the segmented droplets by using the Canny edge detection and Hough circle transformation methods, identify and measure the edges and diameters of the droplets to determine the size of the droplets, and determine the droplet density by statistically calculating the ratio of the number of droplets in the spot area to the area of the region; A droplet thickness calculation subunit, which is used to detect the droplets by using a surface curvature analysis method based on three-dimensional reconstruction technology to estimate the average thickness of the droplets.

4. The intelligent sorting system for fruits and vegetables goods according to claim 1, wherein The predefined reference model refers to an image generation and feature analysis model constructed based on a deep learning algorithm, which can generate a reference image based on the surface properties of the pitaya, generate another reference image by combining the surface properties of the pitaya and the droplet parameters, and at the same time generate corresponding reference reflection component parameters according to different surface properties of the pitaya, and generate correction reflection component parameters corresponding to the reference images generated after applying different droplet parameters.

5. The intelligent sorting system for fruits and vegetables goods according to claim 4, wherein, The specific area of the collected image refers to the circular area outside the spot area in the collected image to ensure that the surface properties input into the predefined reference model subsequently are consistent with the surface properties of the spot area.

6. The intelligent sorting system for fruits and vegetables goods according to claim 5, wherein The anomaly index correction module specifically includes; A reflection component parameter determination unit for respectively determining a reference reflection component parameter and a corrected reflection component parameter corresponding to an initial reference image and an additional reference image based on a predefined reference model; A correction coefficient acquisition unit for quantifying the difference between the reference reflection component parameter and the corrected reflection component parameter and converting the difference into a correction coefficient; An abnormal index correction unit for retrieving a preset correction formula and correcting the initial abnormal index according to the correction coefficient.

7. The intelligent sorting system for fruits and vegetables goods according to claim 6, characterized in that, The preset correction formula is as follows: , where refers to the corrected anomaly index, refers to the initial anomaly index, refers to the corrected reflection component parameter, refers to the reference reflection component parameter, refers to the difference in the reflection components between the initial reference image and the additional reference image, i.e., the correction coefficient, refers to the adjustment coefficient of the correction coefficient.

Images