A method for identifying and detecting the crushing state of cherry juice processing raw materials

Abstract

The present application relates to the technical field of image processing, and more particularly to a cherry juice processing raw material crushing state recognition and detection method, which obtains a gray-scale image of cherry juice processing raw material, detects the kernel in the gray-scale image to obtain at least one kernel region, removes all kernel regions in the gray-scale image to obtain a pulp region, and judges whether the crushing state of the cherry juice processing raw material meets the standard according to the gray-scale distribution characteristics of the pixel points in the pulp region; if the crushing state of the cherry juice processing raw material does not meet the standard, the pulp region is subjected to region segmentation to obtain at least one suspected kernel fragment region, the comprehensive completeness of the pulp and the kernel in the gray-scale image is obtained according to the area characteristics of each kernel region and the shape distribution characteristics of each suspected kernel fragment region, so as to determine the reason why the crushing state of the cherry juice processing raw material does not meet the standard, adjust the equipment parameters of the crusher, and improve the qualified rate of cherry juice.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a method for identifying and detecting the broken state of raw materials used in cherry juice processing. Background Technology

[0002] In cherry juice processing, the crushing state of the raw materials determines the juice yield, nutrient retention, and final product quality. The ideal crushing state requires simultaneously achieving moderate pulp crushing and maintaining the integrity of the pit. The pulp is crushed into small, uniform particles to efficiently release sugars and anthocyanins, but pit breakage introduces bitter substances such as cyanogenic glycosides, leading to increased astringency and color deterioration in the juice. In actual production, due to factors such as cherry variety differences and fluctuations in crushing equipment parameters, the crushing state of cherry juice raw materials easily deviates from the ideal range. Insufficient crushing reduces juice yield and wastes raw materials, while excessive crushing causes pit damage, increasing the fineness of the pulp and making subsequent filtration more difficult, affecting juice clarity and flavor stability. Therefore, accurate identification of the crushing state of cherry juice raw materials is of critical engineering significance for ensuring process stability and improving the yield of high-quality products.

[0003] Existing technologies mainly employ a recognition path based on global feature analysis to identify the breakage state of cherry juice processing raw materials: First, a grayscale image of the cherry juice processing raw material to be detected is acquired. The acquired grayscale image is then segmented using color feature segmentation technology to distinguish between the pulp area and the pit area in the grayscale image. Then, the color distribution characteristics of the pulp area are analyzed to complete the overall assessment of the degree of breakage of the raw material.

[0004] However, the image features of the pulp and pit after crushing are highly mixed, and color-based segmentation analysis methods have limitations. Specifically, pit fragments and insufficiently crushed pulp are extremely similar in color and texture, making accurate differentiation difficult. This leads to difficulties in tracing the specific reasons for substandard cherry juice (i.e., insufficient or excessive crushing). When the pit breaks, its fragments are easily misidentified as large particles from the pulp area, resulting in an incorrect assessment of insufficient pulp crushing. If the crushing intensity is increased accordingly, it will actually exacerbate pit crushing. Conversely, large pieces of pulp may also be misidentified as pit fragments, leading to an unnecessary reduction in crushing intensity and a decrease in juice yield.

[0005] Therefore, accurately identifying the crushed state of raw materials for cherry juice processing, accurately obtaining the specific reasons for cherry juice defects, and thus improving the pass rate of cherry juice have become urgent problems to be solved. Summary of the Invention

[0006] In view of this, embodiments of the present invention provide a method for identifying and detecting the broken state of raw materials for cherry juice processing, in order to solve the problem of how to accurately identify the broken state of raw materials for cherry juice processing, accurately obtain the specific reasons for the failure of cherry juice, and thus improve the pass rate of cherry juice.

[0007] This invention provides a method for identifying and detecting the broken state of raw materials used in cherry juice processing, the method comprising the following steps:

[0008] During the crushing process of cherry juice processing raw materials using a crusher, images of the crushed raw materials are captured, and the crushed images are processed into grayscale images.

[0009] The grayscale image is subjected to fruit pit detection to obtain at least one fruit pit region. All fruit pit regions in the grayscale image are removed to obtain a fruit pulp region. Based on the grayscale distribution characteristics of the pixels in the fruit pulp region and the texture fluctuation characteristics of the fruit pulp region, the roughness of the fruit pulp region is obtained. Based on the roughness of the fruit pulp, it is determined whether the crushing state of the raw materials for cherry juice processing meets the standard.

[0010] If the crushing state of the raw materials for cherry juice processing does not meet the standard, the pulp area is segmented to obtain at least one suspected pit fragment area. Based on the area characteristics of each pit area in the grayscale image and the shape distribution characteristics of each suspected pit fragment area, the overall integrity of the pulp and pit in the grayscale image is obtained.

[0011] By setting a comprehensive integrity threshold and using the difference between the comprehensive integrity of the pulp and pit in the grayscale image and the comprehensive integrity threshold, the reasons why the crushing state of the raw materials for cherry juice processing does not meet the standard can be determined, and the equipment parameters of the crusher can be adjusted accordingly.

[0012] Preferably, the step of obtaining the roughness of the fruit pulp region based on the grayscale distribution characteristics of pixels in the pulp region and the texture fluctuation characteristics of the pulp region includes:

[0013] Obtain the grayscale entropy of the pixels in the pulp region, normalize the grayscale entropy, and obtain the first pulp roughness of the pulp region.

[0014] Obtain the standard deviation of the gray values ​​of the pixels in the pulp region, and normalize the standard deviation to obtain the second pulp roughness of the pulp region;

[0015] The roughness of the first pulp and the roughness of the second pulp are weighted and summed to obtain the roughness of the pulp region.

[0016] Preferably, the step of determining whether the crushing state of the cherry juice processing raw material meets the standard based on the roughness of the fruit pulp includes:

[0017] Set a threshold for the roughness of the pulp. If the roughness of the pulp in the pulp area is less than the threshold, then the crushing state of the raw materials for cherry juice processing is confirmed to meet the standard.

[0018] If the roughness of the pulp in the pulp area is greater than or equal to the pulp roughness threshold, it is confirmed that the crushing state of the raw materials for cherry juice processing does not meet the standard.

[0019] Preferably, the step of obtaining the overall integrity of the pulp and the pit in the grayscale image based on the area characteristics of each pit region in the grayscale image and the shape distribution characteristics of each suspected pit fragment region includes:

[0020] For any kernel region in the grayscale image, the actual area and convex hull area of ​​the kernel region are obtained, the ratio of the actual area to the convex hull area is calculated to obtain the integrity of the kernel region, the integrity of each kernel region in the grayscale image is obtained, the mean integrity value is obtained, and the mean integrity value is normalized to obtain the kernel integrity of the grayscale image.

[0021] Obtain the equivalent circle diameter of each suspected fruit pit fragment region in the grayscale image, form an equivalent circle diameter set, obtain the coefficient of variation of the equivalent circle diameter set, normalize the coefficient of variation, and obtain the fruit pulp integrity of the grayscale image.

[0022] The integrity of the fruit pit and the integrity of the fruit pulp are weighted and summed to obtain the overall integrity of the fruit pulp and the fruit pit in the grayscale image.

[0023] Preferably, the step of determining the reason why the breakage state of the cherry juice processing raw materials does not meet the standard by utilizing the difference between the overall integrity of the pulp and pit in the grayscale image and the overall integrity threshold includes:

[0024] Reasons for substandard crushing of raw materials for cherry juice processing include insufficient crushing of pulp and excessive crushing of pits;

[0025] If the overall integrity of the pulp and pit in the grayscale image is greater than or equal to the overall integrity threshold, then the reason why the crushing state of the raw materials for cherry juice processing does not meet the standard is that the pulp is not crushed enough.

[0026] If the overall integrity of the pulp and pit in the grayscale image is less than the overall integrity threshold, then the reason why the crushed state of the cherry juice processing raw material does not meet the standard is that the pit is excessively crushed.

[0027] Preferably, the step of performing kernel detection on the grayscale image to obtain at least one kernel region includes:

[0028] A set of cherry pit images is obtained, a YOLO detection model is constructed based on the set of cherry pit images, and the YOLO detection model is used to detect the pits in the grayscale images to obtain at least one pit region.

[0029] Preferably, the step of segmenting the pulp region to obtain at least one suspected pit fragment region includes:

[0030] The pulp region is marked with foreground and background to obtain a marked image. The gradient magnitude map of the pulp region is obtained. The marked image is used as a constraint condition for the watershed algorithm. The watershed algorithm is used to segment the gradient magnitude map to obtain at least one suspected kernel fragment region.

[0031] Preferably, the device parameters include:

[0032] The working time, working pressure, speed of the crushing rollers and the gap between the rollers in the crusher.

[0033] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows:

[0034] In this invention, all pit regions in the grayscale image are removed to eliminate interference from the pit regions on the detection of the broken state of cherry juice processing raw materials; the roughness of the pulp region is obtained to quickly determine whether the broken state of the cherry juice processing raw materials meets the standard; when the broken state of the cherry juice processing raw materials does not meet the standard, the overall integrity of the pulp and pit in the grayscale image is obtained, and the correlation between the microscopic features of the pit and pulp is constructed. By quantifying the structural integrity of the pit and the distribution characteristics of the pulp, the specific reasons for the cherry juice's failure are accurately obtained, thereby improving the pass rate of cherry juice. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a flowchart of a method for identifying and detecting the broken state of raw materials for cherry juice processing, provided in Embodiment 1 of the present invention. Detailed Implementation

[0037] Embodiments of this disclosure are described in detail below, with examples of these embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting it.

[0038] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0039] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0040] See Figure 1 This is a flowchart of a method for identifying and detecting the broken state of raw materials for cherry juice processing, provided in Embodiment 1 of the present invention. Figure 1 As shown, the method may include:

[0041] Step S101: During the crushing process of cherry juice processing raw materials using a crusher, a crushing image of the cherry juice processing raw materials is acquired, and the crushing image is processed into grayscale to obtain a grayscale image.

[0042] In cherry juice processing, a crusher is used to break down the raw cherry juice material. The degree of crushing of the material determines the juice yield, nutrient retention, and final product quality. Therefore, identifying and detecting the crushing state of the raw cherry juice material and adjusting the crusher parameters promptly based on the detection results is of critical engineering significance for ensuring process stability and improving the yield of high-quality products.

[0043] Because crushing cherry pulp into small, uniform particles efficiently releases sugars and anthocyanins, while breaking the cherry pit introduces bitter substances like cyanogenic glycosides, leading to increased astringency and color deterioration in the juice, the ideal crushing state for cherry juice processing raw materials must simultaneously meet the dual requirements of moderate pulp crushing and pit preservation. However, in actual production, factors such as differences in cherry varieties and fluctuations in crushing equipment parameters can cause the crushing state of cherry juice processing raw materials to deviate from the ideal range. Insufficient crushing will reduce juice yield and cause raw material waste, while excessive crushing will cause pit damage, leading to increased pulp fineness, which in turn exacerbates the difficulty of subsequent filtration and affects the clarity and flavor stability of the juice.

[0044] Existing technologies mainly employ a recognition path based on global feature analysis to identify the breakage state of cherry juice processing raw materials: First, a grayscale image of the cherry juice processing raw material to be detected is acquired. The acquired grayscale image is then segmented using color feature segmentation technology to distinguish between the pulp area and the pit area in the grayscale image. Then, the color distribution characteristics of the pulp area are analyzed to complete the overall assessment of the degree of breakage of the raw material.

[0045] However, the image features of cherry flesh and pits after crushing are highly mixed, and color-based segmentation analysis methods have limitations. Specifically, pit fragments and insufficiently crushed flesh are extremely similar in color and texture, making accurate differentiation difficult. This leads to difficulties in tracing the specific reasons for substandard cherry juice (i.e., insufficient or excessive crushing). When the pit breaks, its fragments are easily misidentified as large particles from the flesh area, resulting in an incorrect assessment of insufficient flesh crushing. If the crushing intensity is increased accordingly, it will exacerbate pit crushing. Conversely, large pieces of flesh may also be misidentified as pit fragments, leading to an incorrect assessment of excessive crushing and an unnecessary reduction in crushing intensity, resulting in a decrease in juice yield.

[0046] Therefore, in this embodiment of the invention, the grayscale image of the collected cherry juice processing raw material is divided into a pit region and a pulp region. By obtaining the roughness of the pulp in the pulp region, it is quickly determined whether the crushing state of the cherry juice processing raw material meets the standard. When the crushing state of the cherry juice processing raw material does not meet the standard, by obtaining the overall integrity of the pulp and pit in the grayscale image, the correlation between the microscopic features of the pit and pulp is constructed, the specific reasons for the cherry juice being unqualified are accurately obtained, and the pass rate of cherry juice is improved.

[0047] First, during the crushing process of cherry juice processing raw materials using a crusher, images of the crushed raw materials are captured. Since the cherry juice processing raw materials will be crushed into pulp during processing, in order to avoid interference from the pulp state on the imaging, the collected pulp samples are diluted and then evenly spread on the surface of a transparent substrate, with the spreading thickness controlled at 3-8mm. The aim is to make the pulp particles and pits as discretely distributed as possible without stacking or obstruction, so as to reduce mutual interference between the pulp particles and pits. Then, a backlighting method is used to enhance the grayscale difference between the target and the background, and a camera fixed directly above the substrate is used to take pictures to obtain the crushed images.

[0048] After obtaining the broken image, the broken image is preprocessed to obtain a grayscale image, which is used to accurately identify the broken state of the raw materials for cherry juice processing. The specific steps of the preprocessing are as follows: (1) Geometric correction: Based on the camera parameters obtained by the calibration plate, the acquired broken image is subjected to distortion correction and perspective transformation to eliminate the geometric error of the system and obtain a standard view without distortion and conforming to the orthographic projection relationship; (2) Image grayscale: The RGB-HSV fusion weighting method is used to perform grayscale processing on the broken image after geometric correction, retaining the grayscale gradient features and edge details of the linear defects, and avoiding feature loss caused by traditional single-channel grayscale; (3) Contrast enhancement: The grayscale image is contrast stretched to linearly map the image grayscale value to the full range of [0, 255] to enhance the contrast between the defect and the background and ensure that the low-contrast linear defect edges can be captured by subsequent detection. Among them, the distortion correction and perspective transformation, RGB-HSV fusion weighting method and contrast stretching of the acquired broken image are existing technologies and will not be described in detail here.

[0049] Step S102: Perform fruit pit detection on the grayscale image to obtain at least one fruit pit region. Remove all fruit pit regions from the grayscale image to obtain a pulp region. Based on the grayscale distribution characteristics of the pixels in the pulp region and the texture fluctuation characteristics of the pulp region, obtain the pulp roughness of the pulp region. Determine whether the crushing state of the cherry juice processing raw material meets the standard based on the pulp roughness.

[0050] When cherry pulp exhibits extremely high fineness, it usually indicates that the crushing process has been fully completed. In order to better evaluate the fineness of the crushed cherry pulp, this embodiment uses the YOLO object detection algorithm to detect the pits in the obtained grayscale image, obtains at least one pit region, and removes all pit regions in the grayscale image to obtain a pulp region, so as to eliminate the interference of the pit region when evaluating the fineness of the cherry pulp.

[0051] Using the YOLO object detection algorithm to detect cherry pits in grayscale images is an existing technology, and is briefly described here: Images of cherry pits from different varieties are collected to form a cherry pit image set. These images are then divided into training, validation, and test sets according to a set ratio. The model is initialized using the training set, and its parameters are optimized using the validation set. After testing on the test set, a YOLO detection model is obtained. This model is then used to detect cherry pits in grayscale images, identifying at least one pit region. The ratio of the training, validation, and test sets needs to be set according to the actual situation; no restriction is imposed here, and it can be set based on the specific implementation scenario.

[0052] Since the macroscopic texture uniformity of cherry pulp can directly reflect the distribution of pulp particles (density and size difference) in the pulp, that is, the more uniform the texture, the more thorough the crushing and the higher the fineness, after obtaining the pulp area in the grayscale image, the roughness of the pulp area can be obtained based on the grayscale distribution characteristics of the pixels in the pulp area and the texture fluctuation characteristics of the pulp area. This can be used to analyze the fineness of cherry pulp and thus determine whether the crushing state of the raw materials for cherry juice processing meets the standards.

[0053] The method for obtaining the roughness of the fruit pulp region based on the grayscale distribution characteristics of pixels in the pulp region and the texture fluctuation characteristics of the pulp region is as follows:

[0054] Obtain the grayscale entropy of the pixels in the pulp region, normalize the grayscale entropy, and obtain the first pulp roughness of the pulp region.

[0055] Obtain the standard deviation of the gray values ​​of the pixels in the pulp region, and normalize the standard deviation to obtain the second pulp roughness of the pulp region;

[0056] The roughness of the first pulp and the roughness of the second pulp are weighted and summed to obtain the roughness of the pulp region.

[0057] In one embodiment, the formula for calculating the roughness of the pulp in the pulp area is:

[0058]

[0059] in, H represents the roughness of the pulp in the pulp region; H is the grayscale entropy of the pixels in the pulp region. represents the standard deviation of the grayscale values ​​of pixels in the pulp region; This is the normalization function; This is the weighting coefficient for the roughness of the first fruit pulp; This is the weighting coefficient for the second degree of pulp roughness; since the global statistical characteristics (first degree of pulp roughness) and local variation characteristics (second degree of pulp roughness) of the pulp region are equally important for analyzing the pulp roughness of the pulp region, this embodiment sets... There are no restrictions here; settings can be made according to the specific implementation scenario.

[0060] It should be noted that, The value of H represents the roughness of the pulp. The larger the value of H, the more chaotic the gray distribution in the pulp area, the greater the difference in pulp texture, and the coarser the pulp particles. The larger it is, the more... The larger it is; The second aspect is the roughness of the pulp. The larger the value, the more uneven the distribution of the pulp particles in the fruit pulp area, meaning the coarser the pulp particles. The larger it is, the more... The larger it is.

[0061] The less rough the pulp area, the more concentrated the grayscale distribution and the smoother the grayscale fluctuations, consistent with the characteristics of uniform texture and fine particles. Conversely, the greater the roughness of the pulp area, the more scattered the grayscale distribution and the more intense the local fluctuations, usually corresponding to the presence of large pulp particles or agglomeration, indicating insufficient fineness. Therefore, a pulp roughness threshold is set. If the pulp roughness of the pulp area is less than the threshold, it is confirmed that the pulp area has high fineness, the pulp is sufficiently broken, and the crushing state of the cherry juice processing raw material meets the standard. If the pulp roughness of the pulp area is greater than or equal to the threshold, it is confirmed that the pulp area lacks fineness, and the crushing state of the cherry juice processing raw material does not meet the standard.

[0062] In this embodiment, the threshold for fruit pulp roughness is set to 0.3. The main purpose is to ensure that only when the fineness is extremely high can the crushed state of cherry juice processing raw materials be directly judged as meeting the standard. This reduces unnecessary in-depth detection of the crushed state of cherry juice processing raw materials that meet the standard, allowing computing resources to be focused on the reasons why the crushed state of cherry juice processing raw materials does not meet the standard. There are no restrictions here. In actual applications, the threshold for fruit pulp roughness can be adjusted according to the needs of the scenario.

[0063] Thus, the determination of the crushing state of the cherry juice processing raw materials is obtained. If the crushing state of the cherry juice processing raw materials meets the standard, unnecessary in-depth testing is unnecessary, saving computational resources.

[0064] Step S103: If the crushing state of the cherry juice processing raw material does not meet the standard, the pulp area is segmented to obtain at least one suspected pit fragment area. Based on the area characteristics of each pit area in the grayscale image and the shape distribution characteristics of each suspected pit fragment area, the overall integrity of the pulp and pit in the grayscale image is obtained.

[0065] If the crushing state of the cherry juice processing raw materials does not meet the standard, it may be due to excessive crushing caused by pit breakage (damaged pit + uniform pulp particles), meaning that in the pulp area obtained in step S102 above, the pulp is fully crushed but contains small pit fragments; or it may be due to insufficient crushing caused by large pulp particles (intact pit + uneven pulp particles), meaning that in the pulp area obtained in step S102 above, there are no small pit fragments, but the pulp is not sufficiently crushed. Therefore, in this embodiment, a label-controlled watershed algorithm is used to segment the pulp area to obtain at least one suspected pit fragment area, which is used to analyze the specific reasons why the crushing state of the cherry juice processing raw materials does not meet the standard.

[0066] Among them, the use of a label-controlled watershed algorithm for region segmentation of the pulp area is an existing technology, which is briefly described here: First, the pulp area is marked with foreground and background to obtain a labeled image. Then, the gradient magnitude map of the pulp area is obtained. The labeled image is used as a constraint condition for the watershed algorithm. Finally, the watershed algorithm is executed to segment the gradient magnitude map to obtain at least one suspected kernel fragment region.

[0067] Because the image features of cherry pulp and pit after crushing are highly mixed—that is, pit fragments and insufficiently crushed pulp are extremely similar in color and texture, making accurate differentiation difficult—and because pit fragments detach from the pit during the crushing process, the presence of pit fragments in the pulp area means that the pit in the pit area is not a complete pit. Therefore, in this embodiment, the area features of each pit area in the grayscale image and the shape distribution features of each suspected pit fragment area are combined to obtain the overall integrity of the pulp and pit in the grayscale image, construct the correlation between the microscopic features of the pit and pulp, and thus accurately obtain the specific reasons for the cherry juice being substandard.

[0068] The method for obtaining the overall integrity of the pulp and pit in a grayscale image by combining the area features of each pit region in the grayscale image and the shape distribution features of each suspected pit fragment region is as follows:

[0069] For any kernel region in the grayscale image, the actual area and convex hull area of ​​the kernel region are obtained. The method for obtaining the convex hull area is existing technology and will not be described in detail here. The ratio of the actual area to the convex hull area is calculated to obtain the integrity of any kernel region. The integrity of each kernel region in the grayscale image is obtained, and the mean integrity value is obtained. The mean integrity value is normalized to obtain the kernel integrity of the grayscale image.

[0070] The equivalent circle diameter of each suspected fruit pit fragment region in the grayscale image is obtained, forming an equivalent circle diameter set. The coefficient of variation of the equivalent circle diameter set is obtained, and the coefficient of variation is normalized to obtain the fruit pulp integrity of the grayscale image. The equivalent circle diameter and coefficient of variation are existing technologies and will not be described in detail here.

[0071] The integrity of the fruit pit and the integrity of the fruit pulp are weighted and summed to obtain the overall integrity of the fruit pulp and the fruit pit in the grayscale image.

[0072] In one embodiment, the formula for calculating the overall integrity of the pulp and pit in a grayscale image is as follows:

[0073]

[0074] Where F represents the overall integrity of the pulp and pit in the grayscale image; Let be the actual area of ​​the j-th kernel region in the grayscale image; Let be the convex hull area of ​​the j-th kernel region in the grayscale image; M be the number of kernel regions in the grayscale image; and C be the coefficient of variation of the equivalent circle diameter set. This is the normalization function; The weighting coefficient for the integrity of the fruit pit; This is a weighting coefficient for the integrity of the fruit pulp; since the integrity of the pit is direct evidence of whether the raw materials for cherry juice processing have been excessively broken, it should be assigned a higher weighting coefficient. Therefore, in this embodiment, it is set to... There are no restrictions here; settings can be made according to the specific implementation scenario.

[0075] It should be noted that, Let represent the degree of integrity of the j-th kernel region. When the kernel is intact, its outline is full, and its actual area is close to the area of ​​the convex hull. The ratio approaches 1. When the pit breaks, a depression or cavity forms on the pit, causing the actual area to be significantly smaller than the area of ​​the convex hull, and the ratio approaches 0. To assess the integrity of the fruit pit, the overall integrity level of the pit in the grayscale image was quantified. The larger the value, the greater the integrity of the kernel in the grayscale image, and the larger the value of F. The coefficient of variation (C) represents the degree of fruit pulp integrity. It reflects the relative dispersion of the size of suspected pit fragments in the fruit pulp area and quantifies the uniformity of the size of suspected pit fragments. The larger C is, the more uneven the size distribution of suspected pit fragments is, which is more consistent with the uneven distribution of fruit pulp particle size in the fruit pulp area. That is, the fruit pulp is not fully broken. The greater the integrity of the fruit pulp in the grayscale image, the larger F is.

[0076] At this point, the overall integrity of the pulp and pit in the grayscale image is obtained.

[0077] Step S104: Set a comprehensive integrity threshold. Utilize the difference between the comprehensive integrity of the pulp and pit in the grayscale image and the comprehensive integrity threshold to determine the reason why the crushing state of the cherry juice processing raw materials does not meet the standard, and use this information to adjust the equipment parameters of the crusher.

[0078] The reasons for substandard crushing of cherry juice processing raw materials include insufficient pulp crushing and excessive pit crushing. The greater the overall integrity of the pulp and pit in the grayscale image, the more complete the pulp and pulp particles in the corresponding image are. An intact pit excludes excessive pit crushing as the cause of substandard crushing of cherry juice processing raw materials, while an uneven distribution of pulp particles indicates insufficient pulp crushing. Conversely, a smaller overall integrity of the pulp and pit in the grayscale image indicates pit damage and relatively uniform pulp particle distribution in the corresponding image. This usually indicates that small fragments from pit breakage are mixed with pulp particles, suggesting that excessive pit crushing is the cause of substandard crushing of cherry juice processing raw materials.

[0079] Therefore, in this embodiment, the overall integrity threshold is set to 0.5, which is at the critical transition state between the integrity of the pit and the uniformity of the pulp. It is a natural dividing point between the two types of defects, and can provide a clear decision boundary for equipment control while balancing classification accuracy. No restrictions are imposed here; in practical applications, dynamic calibration can be performed according to the production line process requirements. If the overall integrity of the pulp and pit in the grayscale image is greater than or equal to 0.5, it is confirmed that the reason for the substandard crushing state of the cherry juice processing raw materials is insufficient pulp crushing; if the overall integrity of the pulp and pit in the grayscale image is less than 0.5, it is confirmed that the reason for the substandard crushing state of the cherry juice processing raw materials is excessive pit crushing.

[0080] Furthermore, after accurately identifying the specific reasons why the crushing state of the raw materials for cherry juice processing did not meet the standards, these specific reasons were transformed into executable control instructions. By adapting differentiated parameter adjustment strategies to different reasons for the failure of the crushing state of the raw materials for cherry juice processing to meet the standards, the crusher was precisely controlled, eliminating the process factors that cause quality fluctuations from the source, and ultimately achieving the stability and optimization of the crushing state of the raw materials for cherry juice processing.

[0081] The steps for adapting differentiated parameter adjustment strategies to different reasons why the crushing state of cherry juice processing raw materials does not meet the standards are as follows: The working time, working pressure, rotational speed, and inter-roller gap of the crusher.

[0082] If the reason why the crushing state of the raw materials for cherry juice processing does not meet the standard is that the pulp is not crushed enough, the working time of the crusher should be increased proportionally to promote the full crushing of large pulp particles by extending the physical action time; at the same time, the speed or pressure of the crushing roller in the crusher can be slightly increased to enhance the crushing efficiency per unit time; and the pit integrity monitoring should be used to prevent the pit from being over-crushed, thereby improving the fineness of the pulp and the juice yield.

[0083] If the reason why the raw materials for cherry juice processing are not crushed to the standard is that the pits are excessively crushed, then the working pressure of the crusher or the gap between the crushing rollers should be reduced first to fundamentally reduce the mechanical damage to the pits. At the same time, the working time of the crusher should be shortened slightly to avoid excessive grinding of the crushed material after the pressure is reduced, thereby reducing the pit crushing rate, reducing the precipitation of bitter substances, improving the problem of excessively fine fruit residue caused by pit fragments, and improving the subsequent filtration efficiency of cherry juice.

[0084] It is worth noting that adjusting the parameters of the crusher (adapting differentiated parameter adjustment strategies to different reasons why the crushing state of the raw materials for cherry juice processing does not meet the standards) is existing technology. The focus of this invention is to determine the reasons why the crushing state of the raw materials for cherry juice processing does not meet the standards.

[0085] In summary, in this embodiment of the invention, all pit regions in the grayscale image are removed to eliminate interference from the pit regions on the detection of the broken state of cherry juice processing raw materials; the roughness of the pulp region is obtained to quickly determine whether the broken state of the cherry juice processing raw materials meets the standard; when the broken state of the cherry juice processing raw materials does not meet the standard, the overall integrity of the pulp and pit in the grayscale image is obtained, and the correlation between the microscopic features of the pit and pulp is constructed. By quantifying the structural integrity of the pit and the distribution characteristics of the pulp, the specific reasons for the cherry juice's failure are accurately obtained, thereby improving the pass rate of cherry juice.

[0086] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention 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 the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for identifying the state of crushing of a cherry juice processing raw material, characterized by, The cherry juice processing raw material crushing state recognition detection method comprises: In the process of crushing the cherry juice processing raw material using a crusher, a crushing image of the cherry juice processing raw material is collected, and the crushing image is subjected to grayscale processing to obtain a grayscale image; Nucleus detection is performed on the grayscale image to obtain at least one nucleus region, and all nucleus regions in the grayscale image are removed to obtain a pulp region. The roughness of the pulp region is obtained according to the gray level distribution characteristics of the pixel points in the pulp region and the texture fluctuation characteristics of the pulp region. Whether the crushing state of the cherry juice processing raw material meets the standard is determined according to the roughness of the pulp region; If the crushing state of the cherry juice processing raw material does not meet the standard, region segmentation is performed on the pulp region to obtain at least one suspected nucleus fragment region. The comprehensive integrity of the pulp and the nucleus in the grayscale image is obtained according to the area characteristics of each nucleus region and the shape distribution characteristics of each suspected nucleus fragment region in the grayscale image; A comprehensive integrity threshold is set, and the difference between the comprehensive integrity of the pulp and the nucleus in the grayscale image and the comprehensive integrity threshold is used to determine the reason why the crushing state of the cherry juice processing raw material does not meet the standard, which is used to adjust the equipment parameters of the crusher; The roughness of the pulp region is obtained according to the gray level distribution characteristics of the pixel points in the pulp region and the texture fluctuation characteristics of the pulp region, comprising: The gray level entropy of the pixel points in the pulp region is obtained, and the gray level entropy is subjected to normalization processing to obtain the first roughness of the pulp region; The standard deviation of the gray level values of the pixel points in the pulp region is obtained, and the standard deviation is subjected to normalization processing to obtain the second roughness of the pulp region; The first roughness of the pulp region and the second roughness of the pulp region are subjected to weighted summation to obtain the roughness of the pulp region; The comprehensive integrity of the pulp and the nucleus in the grayscale image is obtained according to the area characteristics of each nucleus region and the shape distribution characteristics of each suspected nucleus fragment region, comprising: For any nucleus region in the grayscale image, the actual area and the convex hull area of the any nucleus region are obtained, the ratio of the actual area to the convex hull area is calculated to obtain the integrity of the any nucleus region, the integrity of each nucleus region in the grayscale image is obtained, and the mean value of the integrity is obtained. The integrity mean value is subjected to normalization processing to obtain the nucleus integrity of the grayscale image; The equivalent circle diameters of each suspected nucleus fragment region in the grayscale image are obtained to form an equivalent circle diameter set, the coefficient of variation of the equivalent circle diameter set is obtained, and the coefficient of variation is subjected to normalization processing to obtain the pulp integrity of the grayscale image; The nucleus integrity and the pulp integrity are subjected to weighted summation to obtain the comprehensive integrity of the pulp and the nucleus in the grayscale image.

2. The method according to claim 1, wherein Whether the crushing state of the cherry juice processing raw material meets the standard is determined according to the roughness of the pulp region, comprising: setting a flesh roughness threshold value, and if the flesh roughness of the flesh region is less than the flesh roughness threshold value, it is determined that the crushing state of the cherry juice processing material meets the standard; if the flesh roughness of the flesh region is greater than or equal to the flesh roughness threshold value, it is determined that the crushing state of the cherry juice processing material does not meet the standard.

3. The method according to claim 1, wherein The difference between the comprehensive integrity of the flesh and the pit in the gray-scale image and the comprehensive integrity threshold value is used to determine the reason why the crushing state of the cherry juice processing material does not meet the standard, including: The reason why the crushing state of the cherry juice processing material does not meet the standard includes insufficient crushing of the flesh and excessive crushing of the pit; if the comprehensive integrity of the flesh and the pit in the gray-scale image is greater than or equal to the comprehensive integrity threshold value, it is determined that the reason why the crushing state of the cherry juice processing material does not meet the standard is insufficient crushing of the flesh; if the comprehensive integrity of the flesh and the pit in the gray-scale image is less than the comprehensive integrity threshold value, it is determined that the reason why the crushing state of the cherry juice processing material does not meet the standard is excessive crushing of the pit.

4. The method according to claim 1, wherein The pit detection on the gray-scale image includes: obtaining a set of cherry pit images, constructing a YOLO detection model according to the set of cherry pit images, and using the YOLO detection model to detect the pit in the gray-scale image to obtain at least one pit region.

5. The method according to claim 1, wherein The region segmentation on the flesh region includes: performing foreground labeling and background labeling on the flesh region to obtain a labeled image, obtaining a gradient amplitude graph of the flesh region, using the labeled image as a constraint condition of a watershed algorithm, and using the watershed algorithm to perform region segmentation on the gradient amplitude graph to obtain at least one suspected pit fragment region.

6. The method for identifying and detecting the crushing state of cherry juice processing raw material according to claim 1, characterized in that, The device parameters include: the working time and pressure of the crusher, and the rotating speed of the crushing roller and the gap between the rollers in the crusher.

Images