Blueberry classification and cleaning method and system based on appearance defects and fruit diameter

Abstract

The application discloses a blueberry classification and cleaning method and system based on appearance defects and fruit diameter. A first detection device obtains fruit diameter information and appearance information of blueberries before cleaning, and a second detection device obtains appearance information of the blueberries after cleaning. A control device correlates the appearance information before and after cleaning of the same blueberry according to the release order, fruit diameter, outer contour, fruit stem end characteristics and surface texture characteristics of the blueberries, establishes a surface area correspondence relationship, identifies the skin defects exposed after the removal of the attached matter as the original skin defects exposed after cleaning, and determines the residual attached matter index and the newly added damage index after cleaning. The final grade is determined according to the fruit diameter, appearance information after cleaning, original skin defects exposed after cleaning and appearance re-inspection standard parameters, and the cleaning intensity parameters and conveying speed parameters of the same subsequent processing sub-batch are adjusted according to the two indexes. The application can reduce the misjudgment of defect sources before and after cleaning, and improve the accuracy of the classification and cleaning parameter matching of the blueberries after cleaning.

Description

Technical Field

[0001] This invention relates to the field of fruit and vegetable detection, grading and processing control technology, and in particular to a method and system for classifying and cleaning blueberries based on appearance defects and fruit size. Background Technology

[0002] Blueberries are small fruits, and different fruits vary in diameter, skin integrity, stem fragments, and surface attachments. During blueberry processing, they are typically graded based on diameter and appearance quality, and then quick-frozen, stored, or subjected to other subsequent processing to improve product quality consistency and the utilization rate of the blueberry raw materials.

[0003] Several blueberry quality grading and processing schemes have been proposed in the existing technology. For example, Chinese invention patent application CN106373019A discloses a method for controlling the quality of fresh blueberries. This method classifies blueberries into quality grades based on their diameter and determines the quality grade of the raw blueberry material by statistically analyzing different types of defects such as rot and mold, skin damage, stem defects, and color defects. While this scheme can evaluate blueberry quality based on diameter and appearance defects, it is mainly used for the acceptance and screening of raw blueberry material. It does not involve correlation analysis of the appearance information before and after washing, nor does it adjust the washing parameters based on changes in defects before and after washing.

[0004] For example, Chinese invention patent application CN108391720A discloses a blueberry processing method. Blueberries are fed into a screening device via a conveyor line and sorted into different grades according to fruit diameter, then transported, washed, collected, and cold-stored separately. This method enables continuous grading and classification of blueberries; however, the processing parameters for different grades are mainly executed according to a pre-set process. There is no comparison of the appearance before and after washing, nor is there any feedback adjustment to the washing intensity and conveying speed based on the amount of residue on different categories of blueberries or the additional damage caused by washing.

[0005] In actual processing, dirt, leaf debris, and other adhering substances on the blueberry surface may obscure pre-existing defects such as scratches, cracks, and skin damage, preventing the pre-wash appearance inspection results from fully reflecting the actual skin condition of the blueberries. New defects revealed after washing may be either pre-existing defects exposed after the removal of adhering substances, or damage that occurred or worsened during washing and transportation. If the appearance information of the same blueberry before and after washing cannot be correlated, pre-existing defects exposed after washing may be misjudged as newly added damage, thus affecting the final grading results and the adjustment of processing parameters.

[0006] Meanwhile, blueberries of different initial grades and with different defect types adapt differently to washing intensity and conveying speed. Blueberries with more surface deposits require relatively thorough washing, while blueberries with abrasions, damage, or missing bloom are prone to further damage if the washing is too strong or the conveying time is too long. Existing processing methods typically use preset parameters to continuously process blueberries, or adjust parameters uniformly based on the overall test results of all blueberries. It is difficult to form corresponding processing sub-batches according to the initial grade and main defect types of blueberries, and it is also difficult to adjust the processing parameters of subsequent processing sub-batches of the same type according to the residual deposits and new damage after washing of various types of blueberries.

[0007] Therefore, existing blueberry cleaning methods still struggle to accurately distinguish between pre-existing skin defects and newly caused damage after cleaning, leading to inaccurate assessments of post-cleaning appearance and grading. Furthermore, using uniform cleaning parameters for blueberries with different initial grades and major defect types can easily result in residue buildup or skin damage. Therefore, it is necessary to improve existing blueberry sorting and cleaning methods and systems. Summary of the Invention

[0008] The technical problem this invention aims to solve is that existing blueberry cleaning methods struggle to correlate the appearance information of the same blueberry before and after cleaning, making it difficult to distinguish between pre-existing defects exposed after cleaning and damage generated or amplified during the cleaning process. This leads to inaccurate final grade determination and makes it difficult to match appropriate cleaning intensity and conveying speed parameters to blueberries based on different initial grades and main defect types. Therefore, this invention provides a blueberry classification and cleaning method and system based on appearance defects and fruit diameter.

[0009] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for classifying and cleaning blueberries based on appearance defects and fruit size. The method employs a classification and cleaning system comprising a conveying device, a first detection device, a batch buffering device, a cleaning device, a second detection device, a sorting device, and a control device to process the blueberries, including the following steps: S1. Obtain blueberry diameter information and first appearance information before cleaning through the first detection device. The first appearance information includes information on removable attachments, skin defects, and stem residue.

[0010] S2. The control device determines the initial grade and main defect type of blueberries based on the fruit diameter information and the first appearance information. Based on the initial grade and main defect type, it determines the parameter matching category and appearance re-inspection standard parameters, and controls the batch buffering device to divide blueberries of the same parameter matching category into multiple processing sub-batches.

[0011] S3. The control device matches the cleaning intensity parameters and conveying speed parameters for the current processing sub-batch, and controls the cleaning device and conveying device to clean and convey the blueberries in the current processing sub-batch according to the matched parameters, and keeps the blueberries conveyed in a single layer between the detection position of the first detection device and the detection position of the second detection device.

[0012] S4. Obtain the second appearance information of blueberries after washing in the current processing sub-batch using the second detection device. Determine the candidate matching range based on the release order of blueberries in the current processing sub-batch. Establish a correlation between the first and second appearance information corresponding to the same blueberry based on fruit diameter, outer contour, stem end characteristics, and surface texture characteristics, and establish a correspondence between the surface areas of the blueberries before and after washing. Identify skin defects shown in the second appearance information, whose corresponding surface areas are covered by removable adhering substances in the first appearance information, as original skin defects exposed after washing. Determine the residual adhering substance index of the current processing sub-batch based on the removable adhering substances remaining after washing, and determine the newly added damage index of the current processing sub-batch after washing based on the damage newly identified after washing that does not belong to the original skin defects exposed after washing, and the expansion of skin defects identified before washing after washing.

[0013] S5. Determine the final grade of blueberries based on fruit diameter information, secondary appearance information, original skin defects exposed after washing, and appearance re-inspection standard parameters, and control the sorting device to sort blueberries according to the final grade.

[0014] S6. Based on the residual deposit index and the newly added damage index during cleaning, adjust at least one of the cleaning intensity parameter and conveying speed parameter used in subsequent processing sub-batches of the same parameter matching category, or keep the cleaning intensity parameter and conveying speed parameter unchanged.

[0015] S7. Apply the cleaning intensity parameters and conveying speed parameters determined in step S6 to the next processing sub-batch of the same parameter matching category, so that the next processing sub-batch adopts the parameters determined in step S6 and is cleaned and conveyed according to the cleaning and conveying requirements in step S3. Then, execute steps S4 to S6.

[0016] As a preferred embodiment of the present invention, the first detection device and the second detection device respectively acquire appearance information of blueberries from multiple directions. In step S4, a candidate matching range is determined according to the release order of blueberries in the current processing sub-batch and a preset time interval; the matching similarity between each candidate first appearance information and the second appearance information to be associated is determined according to the fruit diameter, outer contour, stem end features and surface texture features; and the first appearance information with the highest matching similarity and reaching the preset matching threshold is associated with the second appearance information to be associated.

[0017] Furthermore, in step S4, a correspondence between the surface areas of the blueberries before and after washing is established based on the direction of the stem end, the position of the outer contour, and the position of surface features.

[0018] In a preferred embodiment of the present invention, the blueberry diameter grade is determined based on a preset diameter threshold, and the blueberry defect grade is determined based on the proportion of the skin defect area to the effective surface area of ​​the blueberry. The lower quality grade between the diameter grade and the defect grade is determined as the initial grade of the blueberry. When a blueberry has at least one of the following defects: mold, rot, oozing, and open cracks, or when the proportion of the skin defect area to the effective surface area of ​​the blueberry exceeds a preset maximum defect threshold, the blueberry is determined to be rejected.

[0019] In a preferred embodiment of the present invention, the residual residue index is denoted as R, and the residual residue index R is determined according to R = Ar / A0 × 100%. Wherein, A0 is the total area of ​​removable residue on the blueberries in the current processing sub-batch before washing, and Ar is the total area of ​​removable residue remaining on the blueberries after washing. When there are blueberries in the processing sub-batch that have established associations, and A0 is 0, the residual residue index R is 0.

[0020] The newly added damage index after cleaning is denoted as W, and is determined according to the formula W = (An + ΔAd) / At × 100%. Here, An is the total area of ​​the damaged area newly identified after cleaning in the current processing sub-batch of blueberries that are not part of the original epidermal defects exposed after cleaning. For epidermal defects identified before cleaning, if the area of ​​the corresponding defect after cleaning is greater than the area before cleaning, the difference is included in ΔAd; if the area of ​​the corresponding defect after cleaning is not greater than the area before cleaning, the increase in the area of ​​the corresponding defect is 0. ΔAd is the sum of the above area increases, and At is the total effective surface area of ​​the blueberries that have been linked.

[0021] As a preferred embodiment of the present invention, the cleaning intensity parameters include at least one of spray pressure and spray flow rate. The residual threshold R0 of the residual deposit index R is 3% to 8%, and the damage threshold W0 of the newly added cleaning damage index W is 0.5% to 2.0%. The adjustment ratios of the following parameters are all based on the corresponding parameter values ​​of the current processing sub-batch.

[0022] When W is greater than W0, perform at least one of the following adjustments for the next processing sub-batch of the same parameter matching category: reduce the value of at least one parameter in the cleaning intensity parameter by 5% to 20%; increase the conveying speed by 5% to 20%.

[0023] When W is not greater than W0 and R is greater than R0, perform at least one of the following adjustments for the next processing sub-batch of the same parameter matching category: increase the value of at least one parameter in the cleaning intensity parameter by 5% to 15%; decrease the conveying speed by 5% to 15%.

[0024] When W is not greater than W0 and R is not greater than R0, the cleaning intensity parameter and conveying speed parameter used in the next processing sub-batch of the same parameter matching category remain unchanged.

[0025] As a preferred technical solution of the present invention, the main defect types include removable attachment type, epidermal damage type and stem residue type, wherein the epidermal damage type is a type having at least one of the following characteristics: abrasion, epidermal damage and absence of fruit powder.

[0026] Based on the preset basic processing parameters corresponding to the initial grade of blueberries, when the blueberries are of the type with removable attachments, perform at least one of the following adjustments: increase at least one parameter value in the cleaning intensity parameters; reduce the conveying speed.

[0027] When blueberries are of the epidermal fragile type, perform at least one of the following adjustments: reduce the value of at least one of the cleaning intensity parameters; increase the conveying speed.

[0028] When blueberries are classified as stem-damaged, the stem-damaged threshold in the appearance re-inspection standard parameters should be lowered.

[0029] As a preferred technical solution of the present invention, the highest quality grade of blueberries is determined according to the fruit diameter information, the second appearance information, the original skin defects exposed after washing, and the appearance re-inspection standard parameters, in descending order of grade, and the highest quality grade is determined as the final grade of blueberries; when blueberries do not meet the grade standard of the lowest usable grade, or have at least one of mold, rot, oozing, and open cracks, the blueberries are determined as the rejection grade.

[0030] As a preferred technical solution of the present invention, the processing sub-batch is used as the feedback unit, and the cleaning intensity parameter and conveying speed parameter determined according to the feedback result of the current processing sub-batch are used for the next processing sub-batch of the same parameter matching category; the cleaning intensity parameter and conveying speed parameter corresponding to different parameter matching categories are saved respectively, and updated independently according to the feedback result of the corresponding parameter matching category.

[0031] The present invention also provides a blueberry sorting and cleaning system based on appearance defects and fruit size, including a conveying device, a detection device, a cleaning device, a sorting device and a control device, wherein the control device is connected to the conveying device, the detection device, the cleaning device and the sorting device respectively.

[0032] The detection device includes a first detection device and a second detection device respectively located upstream and downstream of the cleaning device along the conveying direction of the conveying device. A batch buffer device is provided between the first detection device and the cleaning device. A sorting device is located downstream of the second detection device. The batch buffer device is connected to a control device. The batch buffer device includes multiple buffer channels and controllable baffles located at the outlets of each buffer channel.

[0033] The conveying device is used to convey blueberries at a single-layer interval between the detection positions of the first detection device and the second detection device.

[0034] The first detection device is used to obtain the diameter information of the blueberries and the first appearance information before washing. The first appearance information includes information on removable attachments, skin defects, and stem residue. The second detection device is used to obtain the second appearance information of the blueberries after washing.

[0035] The control device is used to determine the initial grade and main defect type of blueberries based on fruit diameter information and first appearance information, determine the parameter matching category and appearance re-inspection standard parameters based on the initial grade and main defect type, control the batch buffer device to divide blueberries of the same parameter matching category into multiple processing sub-batches, match the cleaning intensity parameters and conveying speed parameters for the current processing sub-batch, and control the cleaning device and conveying device to operate according to the matched parameters.

[0036] The control device is also used to determine the candidate matching range according to the release order of blueberries in the batch buffer device, and to establish a correlation between the first appearance information and the second appearance information corresponding to the same blueberry according to the fruit diameter, outer contour, stem end features and surface texture features, and to establish the correspondence between the surface areas of the blueberries before and after cleaning; to identify the skin defects shown in the second appearance information and whose corresponding surface areas are covered by removable attachments in the first appearance information as the original skin defects exposed after cleaning; to determine the residual attachment index of the current processing sub-batch according to the removable attachments remaining after cleaning, and to determine the new damage index of the current processing sub-batch after cleaning according to the damage newly identified after cleaning that does not belong to the original skin defects exposed after cleaning and the expansion of the skin defects identified before cleaning after cleaning.

[0037] The control device is also used to determine the final grade of blueberries based on fruit diameter information, second appearance information, original skin defects exposed after washing, and appearance re-inspection standard parameters, and to control the sorting device to sort blueberries according to the final grade; it is also used to adjust or maintain the washing intensity parameters and conveying speed parameters used in the next processing sub-batch of the same parameter matching category based on residual attachment index and washing new damage index.

[0038] As a preferred embodiment of the present invention, both the first detection device and the second detection device include an image acquisition component for acquiring appearance information of blueberries from multiple directions.

[0039] Upstream of the first detection device is a flattening guide section or a vibration dispersion section for forming a single-layer conveying state for blueberries. There is a mutually separated conveying channel between the first detection device and the second detection device. The conveying channel is used to restrict blueberries from passing each other or changing positions across the channel.

[0040] Multiple buffer channels are single-column buffer channels corresponding to parameter matching categories. A diversion execution mechanism connected to the control device is set between the first detection device and the batch buffer device. The diversion execution mechanism is used to send blueberries into the corresponding single-column buffer channel. The controllable material stop is used to release blueberries to form a processing sub-batch according to a preset quantity or preset buffer time.

[0041] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention maintains a single-layer interval during the transport of blueberries between the detection positions of the first and second detection devices. It combines the release sequence, fruit diameter, outer contour, stem end characteristics, and surface texture features to correlate the appearance information of the same blueberry before and after washing. By establishing a correspondence between surface areas before and after washing, it is possible to distinguish between pre-existing epidermal defects exposed after the removal of adhering substances and damage generated or amplified during the washing process. This reduces misjudgment of defect sources and improves the accuracy of appearance defect identification and final grade determination.

[0042] 2. This invention forms processing sub-batches with different parameter matching categories based on the initial grade and main defect types of blueberries, and matches the cleaning intensity parameters and conveying speed parameters respectively, so that the processing parameters can adapt to the actual state of blueberries of different qualities, which is beneficial to reduce blueberry skin damage while ensuring the removal of attached substances.

[0043] 3. This invention uses the residual residue index and the newly added damage index of the processing sub-batch as feedback basis to adjust or maintain the cleaning intensity and conveying speed parameters for subsequent processing sub-batches with the same parameter matching category, and uses the determined parameters for the next processing sub-batch. The processing parameters for different parameter matching categories are saved and updated independently, which is beneficial to balance the degree of cleaning, fruit integrity, and classification cleaning stability. Attached Figure Description

[0044] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the system structure of the present invention; Figure 2This is a schematic diagram of defect analysis of the present invention; Figure 3 This is a schematic diagram of the batch processing of the present invention; Figure 4 This is a flowchart of the parameter feedback process of the present invention; Figure 5 This is a flowchart of the method of the present invention. Detailed Implementation

[0045] The preferred embodiments of the present invention will now be described with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the scope of the invention.

[0046] Example 1 like Figures 1 to 5 As shown, this invention provides a blueberry sorting and cleaning system based on appearance defects and fruit size. The system includes a conveying device, a first detection device, a batch buffer device, a cleaning device, a second detection device, a sorting device, and a control device.

[0047] The first detection device, the batch buffer device, the cleaning device, the second detection device, and the sorting device are arranged sequentially along the conveying direction of the conveying device. The control device is connected to the conveying device, the first detection device, the batch buffer device, the cleaning device, the second detection device, and the sorting device, respectively.

[0048] like Figure 1 As shown, the conveying device extends sequentially along the first detection device, the batch buffer device, the cleaning device, the second detection device, and the sorting device, and is used to carry blueberries through the above devices in sequence. Figure 1 In the diagram, solid arrows between adjacent devices indicate the direction of blueberry transport, while dashed lines indicate control or communication connections between the control device and the corresponding device.

[0049] The conveying device can be a conveyor belt, conveyor rollers, or other mechanisms suitable for low-damage blueberry transport. A leveling guide section or a vibration dispersion section can be installed upstream of the first detection device, in conjunction with a single-row guide channel, to gradually form a single-layer conveying state for the stacked blueberries. Separate conveying channels are provided between the first and second detection devices. Limiting guides or partition structures can be installed within these channels to maintain the spacing between adjacent blueberries and prevent them from overtaking or crossing channels, thus ensuring that the blueberries pass through the two detection positions in a substantially consistent order.

[0050] The first detection device is used to acquire information about the diameter of the blueberries and their initial appearance before washing. The initial appearance information includes information about removable attachments, skin defects, and stem fragments.

[0051] Removeable deposits include mud, leaf debris, and other substances adhering to the blueberry surface that can be removed by washing. Skin defects include at least one of the following: abrasions, skin damage, cracks, loss of bloom, mold, rot, and oozing.

[0052] Both the first and second detection devices include an image acquisition component for acquiring appearance information of blueberries from multiple directions. The first detection device also includes an image analysis component, which determines the fruit diameter based on the blueberry's outer contour and identifies areas where attachments can be removed, areas with skin defects, and areas with residual stems based on the color, brightness, texture, and edge features of different regions in the image.

[0053] The effective area of ​​the blueberry surface referred to in this invention is the blueberry skin area obtained by segmenting from the appearance images of the blueberry from multiple directions, after deducting the image background, the stalk area, and the area of ​​ineffective imaging.

[0054] The first detection device can also be used in conjunction with a rolling conveyor mechanism to make the blueberries rotate during the detection process in order to obtain appearance information of the blueberries from multiple directions.

[0055] The batch buffering device is equipped with multiple single-column buffer channels corresponding to parameter-matched categories. Channel sidewalls or spacer limiters prevent blueberries from stacking or moving across channels during the buffering process. A diversion mechanism controlled by a control device, such as an air nozzle, a feeding plate, or a switching guide plate, is installed between the first detection device and the batch buffering device to feed blueberries into the corresponding buffer channels. Each buffer channel has a controllable stop at its outlet, releasing blueberries according to a preset quantity or preset buffering time, thus forming sequentially processed sub-batches while maintaining the relative release order of the blueberries within the channels.

[0056] The batch buffering device can also be equipped with a rejection exit. When blueberries are moldy, rotten, oozing, or have open cracks, or when the proportion of the surface defect area to the effective area of ​​the blueberry surface exceeds the preset maximum defect threshold, the control device can determine the blueberry as a rejection level and control the batch buffering device to discharge the blueberry before it enters the washing device.

[0057] The cleaning device is used to spray and clean the blueberries in each processing batch. Cleaning intensity parameters include at least one of spray pressure and spray flow rate. Conveyor speed parameters are used to control the speed at which the blueberries pass through the cleaning device.

[0058] By changing the cleaning intensity parameter, the spraying effect on the blueberries can be altered. By changing the conveyor speed parameter, the residence time of the blueberries within the cleaning area can be changed.

[0059] The second detection device is used to acquire second appearance information of the blueberries after washing. The second detection device and the first detection device can use corresponding image acquisition directions to establish the correspondence between the surface areas of the blueberries before and after washing.

[0060] The sorting device is used to transport blueberries to different sorting outlets according to their final grade. The sorting device can adopt an airflow sorting structure, a baffle sorting structure, a flip-plate sorting structure, or a diversion conveying structure.

[0061] The control device includes an initial grading and parameter matching module, a defect change analysis module, a grading adjustment module, and a parameter feedback module.

[0062] The initial grading and parameter matching module sends the blueberry diameter information, initial grade, and appearance re-inspection standard parameters to the grade adjustment module, and sends the parameter matching category, cleaning intensity parameters, and conveying speed parameters of the current processing sub-batch to the parameter feedback module.

[0063] The defect change analysis module sends the original skin defects exposed after cleaning and the corresponding defect change analysis results to the grade adjustment module, and sends the residual attachment index R and the newly added damage index W after cleaning to the parameter feedback module.

[0064] The grading module determines the final grade of the blueberry based on the received information.

[0065] The parameter feedback module determines the cleaning intensity and conveyor speed parameters for the next processing sub-batch within the same parameter matching category, based on the parameter matching category, the processing parameters used in the current processing sub-batch, the residual deposit index R, and the newly added damage index W. The parameter feedback module then returns the determined parameters to the initial grading and parameter matching module for use by the next processing sub-batch.

[0066] The initial grading and parameter matching module receives fruit diameter and appearance information acquired by the first detection device. This module determines the fruit diameter grade of the blueberries based on a preset fruit diameter threshold, and determines the defect grade based on the proportion of skin defects to the effective surface area of ​​the blueberry. The lower grade between the fruit diameter and defect grades is designated as the initial grade of the blueberries. This initial grade is only used for pre-washing condition evaluation, parameter matching, and batch processing, and does not represent the commercial grade of the blueberries after all processing steps.

[0067] In this embodiment, blueberries are divided into four grades: Grade 1, Grade 2, Grade 3, and Grade 4 (rejected). Grade 1 corresponds to blueberries with larger fruit diameter and fewer skin defects; Grade 2 corresponds to blueberries with fruit diameter and skin condition in the middle range; Grade 3 corresponds to blueberries that still have usable value but have smaller fruit diameter or more skin defects.

[0068] The specific fruit diameter threshold and defect threshold corresponding to each grade can be preset according to the blueberry variety, harvest time and subsequent processing requirements.

[0069] The initial grading and parameter matching module also determines the main defect types based on the initial appearance information of the blueberries. The main defect types include removable attachments, fragile skin, and stem remnants.

[0070] The "Removable Attachment" type indicates that the blueberry surface has a preset proportion of removable attachments. The "Skin Fragile" type indicates that the blueberries have at least one of the following characteristics: abrasions, skin damage, and loss of bloom. The "Stemless" type indicates that the blueberries have a residual stem.

[0071] When the same blueberry meets the criteria for two or more major defect types, the initial grading and parameter matching module determines the major defect type to be used to match the cleaning intensity parameter and the conveying speed parameter according to the preset priority.

[0072] Among these, the type with easily damaged skin has a higher priority than the type with removable attachments, to avoid adjusting parameters in opposite directions simultaneously on the same blueberry. When blueberries also have stem remnant characteristics, the stem remnant judgment threshold in the appearance re-inspection standard parameters is lowered, based on matching the cleaning intensity parameters and conveying speed parameters according to other major defect types.

[0073] like Figure 3 As shown, the initial grading and parameter matching module determines the parameter matching category based on the initial grade and main defect type of the blueberries. Different combinations of initial grades and main defect types correspond to parameter matching categories A, B, and C, respectively. Blueberries within the same parameter matching category have similar or identical requirements for cleaning intensity and conveying speed parameters, although their fruit diameter and appearance may differ. Blueberries that reach the rejection grade are discharged before entering the cleaning device and do not participate in the formation of normal processing sub-batches.

[0074] Each parameter matching category has corresponding basic cleaning intensity parameters, conveying speed parameters, and appearance re-inspection standard parameters. For each parameter matching category, the basic cleaning intensity parameters and conveying speed parameters are determined in advance through cleaning tests on the same variety of blueberries. The first processing sub-batch uses this set of parameters, and subsequent processing sub-batches adjust or maintain the corresponding parameters based on the results of the residual attachment index R and the newly added damage index W after cleaning. The appearance re-inspection standard parameters include at least one of the following: the threshold for the proportion of skin defects after cleaning, the threshold for judging residual attachments, and the threshold for judging stem residue, used to determine whether the appearance of the blueberries after cleaning meets the grade standard for the corresponding level.

[0075] For example, parameter matching category A corresponds to blueberries with a large amount of removable attachments, and its basic processing parameters use a higher cleaning intensity and a lower conveying speed. Parameter matching category B corresponds to blueberries with fragile skin, and its basic processing parameters use a lower cleaning intensity and a higher conveying speed. Parameter matching category C corresponds to blueberries with damaged stems, and its appearance re-inspection standard parameters use a lower stem damage threshold to improve the identification and sorting requirements for blueberries with damaged stems.

[0076] The initial grading and parameter matching module uses the preset basic processing parameters corresponding to the initial grade of blueberries as a benchmark. When the blueberries are of the type with removable attachments, at least one of the following adjustments is performed: increasing at least one parameter value in the cleaning intensity parameter; or decreasing the conveying speed.

[0077] When blueberries are of the epidermal fragile type, perform at least one of the following adjustments: reduce the value of at least one of the cleaning intensity parameters; increase the conveying speed.

[0078] When blueberries are classified as stem-damaged, the stem-damaged threshold in the appearance re-inspection standard parameters should be lowered.

[0079] The defect variation analysis module is used to establish a correlation between the first appearance information and the second appearance information corresponding to the same blueberry.

[0080] During the association process, a candidate matching range is first determined based on the release order of blueberries in the batch buffer device and a preset time interval. Then, within this range, the fruit diameter, outer contour, stem end features, and surface texture features of each candidate first appearance information are compared with the second appearance information to be associated. To reduce the impact of changes in surface moisture and reflectivity before and after blueberry washing on the matching results, the first and second appearance images can be calibrated for size, normalized for brightness, or have reflective areas processed before feature comparison. Since blueberries may rotate during washing and transportation, images with corresponding or small viewing angle differences can be selected from appearance images acquired from multiple directions based on the stem end direction and outer contour for feature comparison. For surface texture features, local feature points and corresponding feature descriptors can be extracted within the effective area of ​​the blueberry surface, and the surface texture feature similarity is determined based on the degree of matching between feature descriptors. The local feature points and feature descriptors can be obtained using SIFT feature extraction, ORB feature extraction, or other feature extraction methods capable of characterizing local image features. When using SIFT feature descriptors, the degree of matching can be determined based on the Euclidean distance between feature descriptors; when using ORB feature descriptors, the degree of matching can be determined based on the Hamming distance between binary feature descriptors. The similarities of fruit diameter, outer contour, stem end features, and surface texture features are normalized and then synthesized into a matching similarity score according to preset weights. The relevant weights and preset matching thresholds can be calibrated using blueberry samples with established correspondences. When the highest matching similarity score reaches the preset matching threshold, the defect change analysis module associates the corresponding first appearance information with the second appearance information to be associated.

[0081] After establishing the association, the defect change analysis module establishes the correspondence between the surface areas of the blueberries before and after washing based on the direction of the stem end, the position of the outer contour, and the position of surface features.

[0082] When the highest matching similarity does not reach the preset matching threshold, the corresponding blueberry is marked as a failed association. Successfully associated blueberries are included in the determination of the original epidermal defects exposed after cleaning, the residual attachment index R, and the newly added damage index W. Blueberries that fail to be associated are not included in the above defect change statistics.

[0083] For blueberries that fail to be associated, the grading module determines the final grade based on their fruit diameter information and second appearance information, and the determined final grade is not higher than the initial grade of the blueberry.

[0084] When the final grade cannot be reliably determined based solely on fruit diameter and second appearance information, the sorting device controls the transport of the blueberry to the manual re-inspection exit.

[0085] When there are no successfully associated blueberries in the processing sub-batch, or when the proportion of failed association blueberries exceeds the preset failure ratio, the parameter feedback module will not adjust the processing parameters according to the processing sub-batch, and will keep the cleaning intensity parameters and conveying speed parameters used in the next processing sub-batch of the same parameter matching category unchanged.

[0086] like Figure 2 As shown, the defect change analysis module identifies the epidermal defects displayed in the second appearance information, and whose corresponding surface areas are covered by removable attachments in the first appearance information, as the original epidermal defects exposed after cleaning.

[0087] By using the above identification method, we can avoid directly judging the original skin defects revealed after the attachment is removed as new damage caused during the cleaning process.

[0088] The residual deposit index R is determined according to the following formula: R = Ar / A0 × 100%.

[0089] Where A0 represents the total area of ​​removable residue on successfully associated blueberries before cleaning in the current processing sub-batch, and Ar represents the total area of ​​removable residue remaining on successfully associated blueberries after cleaning. The corresponding area of ​​unassociated blueberries is not included in A0 and Ar.

[0090] When there are successfully associated blueberries in the processing sub-batch and A0 is 0, the residual attachment index R is 0.

[0091] The newly added damage index W during cleaning is determined according to the following formula: W=(An+ΔAd) / At×100%.

[0092] Where An represents the total area of ​​the damaged area of ​​the blueberries successfully associated in the current processing sub-batch, newly identified after cleaning, and not belonging to the original epidermal defects exposed after cleaning. For epidermal defects identified before cleaning, if the area of ​​the corresponding defect after cleaning is greater than the area before cleaning, the difference is included in ΔAd; if the area of ​​the corresponding defect after cleaning is not greater than the area before cleaning, the increase in the area of ​​the corresponding defect is 0. ΔAd is the sum of the above area increases. At represents the total effective surface area of ​​the successfully associated blueberries.

[0093] The area of ​​the corresponding region of the blueberry that failed to associate is not included in An, ΔAd, and At.

[0094] When there are no successfully associated blueberries in the processing sub-batch, the residual attachment index R and the newly added damage index W after cleaning are not determined, and no feedback is given based on the processing sub-batch execution parameters.

[0095] The aforementioned area can be determined based on the number of pixels in the corresponding region of the image. When the imaging ratios of the first and second detection devices are the same or have undergone size calibration, the pixel area of ​​the corresponding region can be directly used for proportional calculation.

[0096] The grading module determines the highest quality grade of blueberries based on their fruit diameter information, secondary appearance information, original skin defects exposed after washing, and appearance re-inspection standard parameters, in descending order of grade, and sets this highest quality grade as the final grade of the blueberries.

[0097] If the actual skin condition of the blueberry after washing still meets the grading standards of its initial grade, the blueberry retains its initial grade.

[0098] When blueberries show significant pre-existing skin defects after washing, or if washing causes additional damage and they no longer meet the initial grading criteria, the final grading will be reassessed based on the actual skin condition after washing. The reassessed final grading will be lower than the initial grading.

[0099] If blueberries do not meet the minimum usable grade standards, or if they are moldy, rotten, oozing, or have open cracks, they will be rejected. The final grade mentioned above is determined after the blueberries have completed the cleaning process and a visual inspection; it does not represent the final product grade after all subsequent processing steps. If blueberries require further quick-freezing, stem removal, impurity removal, or color sorting, they can be re-graded according to product standards after the corresponding processes are completed.

[0100] The parameter feedback module uses the processing sub-batch as the feedback unit. Based on the residual attachment index R and the newly added cleaning damage index W of the current processing sub-batch, it determines the cleaning intensity parameter and conveying speed parameter to be used in the next processing sub-batch of the same parameter matching category.

[0101] The determined parameters can be adjusted based on the parameters used in the current processing sub-batch, or they can remain unchanged.

[0102] In this embodiment, the residual threshold R0 is selected as 5%, and the damage threshold W0 is selected as 1.0%. In other embodiments, the residual threshold R0 can be selected in the range of 3% to 8%, and the damage threshold W0 can be selected in the range of 0.5% to 2.0%.

[0103] The following parameter adjustment ratios are all based on the corresponding parameter values ​​of the current processing sub-batch.

[0104] like Figure 4As shown, when W is greater than W0, it indicates that the current processing parameters result in a higher rate of new damage to the blueberries. The parameter feedback module performs at least one of the following adjustments for the next processing sub-batch of the same parameter matching category: reducing at least one parameter value in the cleaning intensity parameters by 5% to 20%; increasing the conveyor speed by 5% to 20%.

[0105] When W is not greater than W0 and R is greater than R0, it indicates that the new damage from cleaning is within the allowable range, but the residual deposits are too high. The parameter feedback module performs at least one of the following adjustments for the next processing sub-batch of the same parameter matching category: increasing at least one parameter value in the cleaning intensity parameter by 5% to 15%; or decreasing the conveyor speed by 5% to 15%.

[0106] When W is not greater than W0 and R is not greater than R0, it indicates that the current processing parameters can balance the removal of attached substances and the integrity of the fruit. In this case, the cleaning intensity and conveying speed parameters used in the next processing sub-batch of the same parameter matching category should be maintained.

[0107] The parameter feedback module saves the cleaning intensity parameters and conveying speed parameters corresponding to different parameter matching categories, and updates them independently based on the corresponding feedback results.

[0108] The parameter feedback provided for the current parameter matching category does not change the processing parameters used for other parameter matching categories.

[0109] like Figure 5 As shown, the invention is used as follows: S1. The blueberries to be processed are conveyed to the first detection device through the conveying device, and the first detection device obtains the fruit diameter information and the first appearance information of the blueberries before washing.

[0110] S2. The control device determines the initial grade and main defect types of blueberries based on fruit diameter information and first appearance information, and determines the parameter matching category and appearance re-inspection standard parameters based on the initial grade and main defect types. The batch buffering device divides blueberries of the same parameter matching category into multiple processing sub-batches.

[0111] S3. The control device matches the cleaning intensity parameters and conveying speed parameters for the current processing sub-batch, and controls the cleaning device and conveying device to clean and convey the blueberries according to the matched parameters. The blueberries are conveyed in a single layer between the detection positions of the first detection device and the second detection device.

[0112] S4. The second detection device acquires the second appearance information of the blueberries after washing in the current processing sub-batch. The defect change analysis module determines the candidate matching range according to the release order of the blueberries in the current processing sub-batch, and establishes a correlation between the first appearance information and the second appearance information corresponding to the same blueberry based on the fruit diameter, outer contour, stem end features and surface texture features, while establishing the correspondence between the surface areas of the blueberries before and after washing.

[0113] The defect change analysis module identifies the original skin defects exposed after cleaning and determines the residual attachment index R and the newly added damage index W of the current processing sub-batch.

[0114] S5, the grade adjustment module determines the final grade of blueberries based on fruit diameter information, second appearance information, original skin defects exposed after washing, and appearance re-inspection standard parameters, and controls the sorting device to sort blueberries according to the final grade.

[0115] S6. The parameter feedback module adjusts at least one of the cleaning intensity parameter and conveying speed parameter used in subsequent processing sub-batches of the same parameter matching category based on the residual attachment index R and the newly added damage index W, or keeps the cleaning intensity parameter and conveying speed parameter unchanged.

[0116] S7. Apply the cleaning intensity parameters and conveying speed parameters determined in step S6 to the next processing sub-batch of the same parameter matching category. The next processing sub-batch uses the parameters determined in step S6 and performs cleaning and conveying according to the cleaning and conveying requirements in step S3. Then, steps S4 to S6 are executed, thereby forming a parameter feedback closed loop for the category.

[0117] Furthermore, when associating the first and second appearance information of the same blueberry, the candidate matching range is first determined based on the release order of the blueberry in the processing sub-batch and the preset time interval.

[0118] Then, the matching similarity is determined based on the fruit diameter, outer contour, stem end features and surface texture features. The first appearance information with the highest matching similarity and reaching the preset matching threshold is associated with the second appearance information to be associated. The surface region correspondence is established based on the stem end direction, outer contour position and surface feature position.

[0119] Furthermore, the first and second detection devices can be respectively set in the light-shielding detection area and use the same or calibrated lighting conditions to reduce the impact of changes in ambient light on the identification results of the attachment area and the epidermal defect area.

[0120] Furthermore, the residual residue index R is calculated using the processing sub-batch as the statistical unit, based on the A0 and Ar values ​​corresponding to all successfully associated blueberries in that processing sub-batch, without calculating the residual ratio for each individual blueberry.

[0121] When there are successfully associated blueberries in the processing sub-batch, and the total area A0 of the removable attachment area of ​​all successfully associated blueberries before cleaning is 0, the residual attachment index R of the processing sub-batch is determined to be 0.

[0122] When there are no successfully associated blueberries in the processing sub-batch, the residual attachment index R and the newly added damage index W after cleaning are not determined, and no feedback is given based on the processing sub-batch execution parameters.

[0123] Furthermore, new damage from cleaning includes newly appearing abrasions, epidermal breaks, cracks, and loss of bloom after cleaning, as well as the expansion of epidermal defects identified before cleaning. Existing epidermal defects that become apparent after cleaning due to the removal of adhering substances are not included in the new damage from cleaning.

[0124] Furthermore, after completing the above cleaning and sorting, subsequent processing parameters can be set according to the final grade of the blueberries. First-grade blueberries can be quickly frozen at a lower temperature and then stored at low temperature; second-grade blueberries can be stored using conventional quick-freezing and freezing; and third-grade blueberries can be stored frozen or otherwise appropriately classified and utilized depending on their integrity and degree of defects.

[0125] Example 2 This embodiment illustrates the parameter feedback process when the residual amount of attached material is too high.

[0126] Blueberries belonging to the same parameter matching category were selected to form the first and second processing sub-batches. The first processing sub-batches were cleaned using a spray pressure of 0.18 MPa, a spray flow rate of 18 L / min, and a conveying speed of 0.20 m / s.

[0127] The total area of ​​removable residue A0 before cleaning in the first processing sub-batch is 1200 mm², and the total area of ​​removable residue Ar after cleaning is 84 mm².

[0128] Based on the calculation of R = Ar / A0 × 100%, the residual adhering matter index R of the first processed sub-batch is 7.0%.

[0129] In the first processing sub-batch, the total effective area At of the blueberry surface is 50,000 mm². The total area An of the damaged areas newly identified after cleaning that are not part of the original epidermal defects exposed after cleaning is 200 mm², and the increase in area ΔAd of the epidermal defects identified before cleaning after cleaning is 150 mm².

[0130] According to the calculation based on W = (An + ΔAd) / At × 100%, the newly added damage index W for cleaning in the first processing sub-batch is 0.7%.

[0131] In this embodiment, the residual threshold R0 is 5%, and the damage threshold W0 is 1.0%. Since W is not greater than W0 and R is greater than R0, it indicates that the new damage in the first processing sub-batch is within the allowable range, but the residual residue after cleaning is too high.

[0132] The parameter feedback module uses the processing parameters of the first processing sub-batch as a benchmark, and increases the spray pressure of the second processing sub-batch by 10% to 0.198MPa, while keeping the spray flow rate and conveying speed unchanged.

[0133] Other adjustment methods include keeping the spray pressure and spray flow rate constant while reducing the conveying speed by 10%, or adjusting both the cleaning intensity parameters and the conveying speed parameters simultaneously.

[0134] After the second processing sub-batch completes cleaning, the second appearance information is re-acquired, and its residual deposit index R and newly added cleaning damage index W are determined. The residual deposit index R of the second processing sub-batch is 4.5%, and the newly added cleaning damage index W is 0.8%. Compared with the first processing sub-batch, the residual deposit index R of the second processing sub-batch decreased from 7.0% to 4.5%, and the newly added cleaning damage index W did not exceed the damage threshold W0 of 1.0%, indicating that increasing the spray pressure can reduce the deposit residue of blueberries of this parameter matching category, while keeping the newly added cleaning damage within the allowable range. Since both R and W of the second processing sub-batch are not greater than the corresponding thresholds, the cleaning intensity parameters and conveying speed parameters used for subsequent processing sub-batches of this parameter matching category are maintained.

[0135] Example 3 This embodiment illustrates the parameter feedback process when the amount of newly added damage during cleaning is too high.

[0136] Blueberries belonging to the epidermal vulnerability type were identified as belonging to the same parameter matching category, and thus formed the first processing sub-batch and the second processing sub-batch.

[0137] The first processing batch was cleaned using a spray pressure of 0.15 MPa, a spray flow rate of 16 L / min, and a conveying speed of 0.20 m / s.

[0138] The total area of ​​removable residue A0 before cleaning in the first processing sub-batch is 1000 mm², and the total area of ​​removable residue Ar after cleaning is 30 mm², with a residual residue index R of 3.0%.

[0139] In the first processing sub-batch, the total effective surface area At of blueberries was 50,000 mm². The total damaged area An, which was newly identified after cleaning and was not part of the original epidermal defects exposed after cleaning, was 450 mm². The area increase ΔAd of epidermal defects identified before cleaning after cleaning was 250 mm². The newly added damage index W after cleaning was 1.4%.

[0140] Since W is greater than the damage threshold W0 of 1.0%, it indicates that the current cleaning and transport parameters have a strong effect on this type of blueberry.

[0141] The parameter feedback module uses the processing parameters of the first processing sub-batch as a benchmark, reduces the spray pressure of the second processing sub-batch by 10% to 0.135MPa, increases the conveying speed by 10% to 0.22m / s, and keeps the spray flow rate unchanged.

[0142] The second processing sub-batch was processed using the above parameters. After cleaning, the second appearance information was re-acquired, and its residual residue index R and cleaning-induced damage index W were determined. The residual residue index R of the second processing sub-batch was 4.0%, and the cleaning-induced damage index W was 0.9%. Compared with the first processing sub-batch, the cleaning-induced damage index W of the second processing sub-batch decreased from 1.4% to 0.9%, and the residual residue index R did not exceed the residual threshold R0 of 5%, indicating that reducing the spray pressure and increasing the conveying speed could reduce the cleaning-induced damage of blueberries in this parameter-matched category, while keeping the residue within the allowable range. Since W of the second processing sub-batch was not greater than the damage threshold W0, and R was not greater than the residual threshold R0, the above parameters were continued to be used for the next processing sub-batch in this parameter-matched category.

[0143] Example 4 This embodiment illustrates the parameter maintenance process when both the residual deposit index and the newly added damage index after cleaning are within the allowable range.

[0144] The current processing sub-batch with the same parameter matching category is cleaned using a spray pressure of 0.16MPa, a spray flow rate of 17L / min, and a conveying speed of 0.21m / s.

[0145] The total area of ​​removable contaminants before cleaning in the current processing sub-batch is A0, which is 900 mm². After cleaning, the total area of ​​removable contaminants remaining is Ar, which is 36 mm². The residual contaminant index R is 4.0%.

[0146] The total effective surface area At of blueberries in the current processing sub-batch is 50,000 mm². The total damaged area An of newly identified areas after cleaning that are not part of the original epidermal defects exposed after cleaning is 250 mm². The increase in area ΔAd of epidermal defects identified before cleaning after cleaning is 150 mm². The newly added damage index W after cleaning is 0.8%.

[0147] Since R is no greater than 5% of the residual threshold R0 and W is no greater than 1.0% of the damage threshold W0, the parameter feedback module maintains the spray pressure, spray flow rate and conveying speed parameters used in the next processing sub-batch of the matching parameter category.

[0148] This embodiment illustrates that when both the removal of adhering substances and the integrity of blueberries meet the preset requirements, there is no need to continue adjusting the processing parameters, thereby avoiding frequent fluctuations in processing parameters between different processing sub-batches.

[0149] This invention distinguishes between pre-existing epidermal defects exposed after washing and newly added damage by associating the appearance information of the same blueberry before and after washing and establishing a correspondence between surface areas.

[0150] Based on the residual residue index and new damage index of the processing sub-batch, the cleaning intensity parameter and conveying speed parameter are adjusted or maintained for subsequent processing sub-batches with the same parameter matching category, and the determined parameters are used for the next processing sub-batch, so as to take into account the cleaning effect of blueberries, fruit integrity and the accuracy of grading after cleaning.

[0151] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. All modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for classifying and cleaning blueberries based on appearance defects and fruit diameter, characterized in that, A sorting and cleaning system comprising a conveying device, a first detection device, a batch buffering device, a cleaning device, a second detection device, a sorting device, and a control device is used to process blueberries. The method includes the following steps: S1. Obtain blueberry diameter information and first appearance information before cleaning through the first detection device. The first appearance information includes information on removable attachments, skin defects, and stem residue. S2. The control device determines the initial grade and main defect type of blueberries based on the fruit diameter information and the first appearance information, determines the parameter matching category and appearance re-inspection standard parameters based on the initial grade and main defect type, and controls the batch buffering device to divide blueberries of the same parameter matching category into multiple processing sub-batches. S3. The control device matches the cleaning intensity parameters and conveying speed parameters for the current processing sub-batch, and controls the cleaning device and conveying device to clean and convey the blueberries in the current processing sub-batch according to the matched parameters, and keeps the blueberries conveyed in a single layer between the detection position of the first detection device and the detection position of the second detection device. S4. Obtain the second appearance information of blueberries after washing in the current processing sub-batch through the second detection device; determine the candidate matching range according to the release order of blueberries in the current processing sub-batch; establish the association between the first appearance information and the second appearance information corresponding to the same blueberry according to the fruit diameter, outer contour, fruit stem end features and surface texture features, and establish the correspondence between the surface areas of blueberries before and after washing; identify the epidermal defects shown in the second appearance information and whose corresponding surface areas are covered by removable attachments in the first appearance information as the original epidermal defects exposed after washing; determine the residual attachment index of the current processing sub-batch according to the removable attachments remaining after washing, and determine the new damage index of the current processing sub-batch after washing according to the damage newly identified after washing that does not belong to the original epidermal defects exposed after washing and the expansion of the epidermal defects identified before washing after washing. S5. Determine the final grade of blueberries based on the fruit diameter information, second appearance information, original skin defects exposed after washing, and appearance re-inspection standard parameters, and control the sorting device to sort blueberries according to the final grade. S6. Based on the residual deposit index and the new damage index after cleaning, adjust at least one of the cleaning intensity parameter and conveying speed parameter used in subsequent processing sub-batches with the same parameter matching category, or keep the cleaning intensity parameter and conveying speed parameter unchanged. S7. Apply the cleaning intensity parameter and conveying speed parameter determined in step S6 to the next processing sub-batch of the same parameter matching category, so that the next processing sub-batch adopts the parameters determined in step S6 and is cleaned and conveyed according to the cleaning and conveying requirements in step S3, and then execute steps S4 to S6.

2. The blueberry sorting and cleaning method based on appearance defects and fruit diameter according to claim 1, characterized in that, The first and second detection devices respectively acquire appearance information of blueberries from multiple directions; In step S4, the candidate matching range is determined according to the release order of blueberries in the current processing sub-batch and the preset time interval; the matching similarity between each candidate first appearance information and the second appearance information to be associated is determined according to the fruit diameter, outer contour, fruit stem end features and surface texture features; and the first appearance information with the highest matching similarity and reaching the preset matching threshold is associated with the second appearance information to be associated.

3. The blueberry sorting and cleaning method based on appearance defects and fruit diameter according to claim 2, characterized in that, In step S4, the surface area correspondence of blueberries before and after washing is established based on the direction of the stem end, the position of the outer contour, and the position of surface features.

4. The blueberry sorting and cleaning method based on appearance defects and fruit diameter according to claim 2, characterized in that, The residual attachment index is denoted as R. The residual attachment index R is determined according to R = Ar / A0 × 100%, where A0 is the total area of ​​removable attachment area of ​​blueberries that have established the association in the current processing sub-batch before cleaning, and Ar is the total area of ​​removable attachment area of ​​blueberries that have established the association after cleaning. When there are blueberries that have established the association in the processing sub-batch and A0 is 0, the residual attachment index R is 0. The newly added damage index during cleaning is denoted as W. The newly added damage index W is determined according to W = (An + ΔAd) / At × 100%, where An is the total area of ​​the damaged area of ​​the blueberries in the current processing sub-batch that have established the association and are newly identified after cleaning and do not belong to the original epidermal defects exposed after cleaning; ΔAd is the increase in area of ​​epidermal defects identified before cleaning of the blueberries that have established the association and are now cleaned; and At is the total effective area of ​​the surface region of the blueberries that have established the association.

5. The blueberry sorting and cleaning method based on appearance defects and fruit diameter according to claim 4, characterized in that, The cleaning intensity parameter includes at least one of spray pressure and spray flow rate; The residual threshold R0 of the residual attachment index R is 3% to 8%, and the damage threshold W0 of the newly added damage index W is 0.5% to 2.0%. The following parameter adjustment ratios are all based on the corresponding parameter values ​​of the current processing sub-batch. When W is greater than W0, perform at least one of the following adjustments for the next processing sub-batch of the same parameter matching category: reduce the value of at least one of the cleaning intensity parameters by 5% to 20%; increase the conveying speed by 5% to 20%; When W is not greater than W0 and R is greater than R0, perform at least one of the following adjustments on the next processing sub-batch of the same parameter matching category: increase the value of at least one of the cleaning intensity parameters by 5% to 15%; decrease the conveying speed by 5% to 15%; When W is not greater than W0 and R is not greater than R0, the cleaning intensity parameter and conveying speed parameter used in the next processing sub-batch of the same parameter matching category remain unchanged.

6. The blueberry sorting and cleaning method based on appearance defects and fruit diameter according to claim 1, characterized in that, The main defect types include removable attachment type, epidermal vulnerable type, and stem remnant type, wherein the epidermal vulnerable type is a type that has at least one of the following characteristics: abrasion, epidermal damage, and absence of bloom; Based on the preset basic processing parameters corresponding to the initial grade of blueberries, when the blueberries belong to the type of removable attachments, at least one of the following adjustments is performed: increasing at least one parameter value among the cleaning intensity parameters; reducing the conveying speed; When blueberries are of the aforementioned skin-vulnerable type, at least one of the following adjustments shall be performed: reducing the value of at least one of the cleaning intensity parameters; increasing the conveying speed; When the blueberry belongs to the type with a damaged stem, the threshold for determining the damaged stem in the appearance re-inspection standard parameters is lowered.

7. The blueberry sorting and cleaning method based on appearance defects and fruit diameter according to claim 5, characterized in that, Using the processing sub-batch as the feedback unit, the cleaning intensity parameters and conveying speed parameters determined based on the feedback results of the current processing sub-batch are used for the next processing sub-batch of the same parameter matching category; the cleaning intensity parameters and conveying speed parameters corresponding to different parameter matching categories are saved separately and updated independently based on the feedback results of the corresponding parameter matching categories.

8. A blueberry sorting and cleaning system based on appearance defects and fruit diameter, comprising a conveying device, a detection device, a cleaning device, a sorting device, and a control device, wherein the control device is connected to the conveying device, the detection device, the cleaning device, and the sorting device respectively, characterized in that, The detection device includes a first detection device and a second detection device respectively disposed upstream and downstream of the cleaning device along the conveying direction of the conveying device. A batch buffer device is disposed between the first detection device and the cleaning device. The sorting device is disposed downstream of the second detection device. The batch buffer device is connected to the control device. The batch buffer device includes multiple buffer channels and controllable baffles disposed at the outlets of each buffer channel. The conveying device is used to convey blueberries at a single-layer interval between the detection positions of the first detection device and the second detection device. The first detection device is used to acquire blueberry diameter information and first appearance information before washing. The first appearance information includes information on removable attachments, skin defects, and stem residue. The second detection device is used to acquire second appearance information of blueberries after washing. The control device is used to determine the initial grade and main defect type of blueberries based on the fruit diameter information and the first appearance information, determine the parameter matching category and appearance re-inspection standard parameters based on the initial grade and main defect type, control the batch buffer device to divide blueberries of the same parameter matching category into multiple processing sub-batches, match the cleaning intensity parameters and conveying speed parameters for the current processing sub-batches, and control the cleaning device and conveying device to operate according to the matched parameters. The control device is also used to determine the candidate matching range according to the release order of blueberries in the batch buffer device, and to establish a correlation between the first appearance information and the second appearance information corresponding to the same blueberry according to the fruit diameter, outer contour, stem end features and surface texture features, and to establish a correspondence between the surface areas of the blueberries before and after cleaning; to identify the epidermal defects shown in the second appearance information and whose corresponding surface areas are covered by removable attachments in the first appearance information as the original epidermal defects exposed after cleaning; to determine the residual attachment index of the current processing sub-batch according to the removable attachments remaining after cleaning, and to determine the new damage index of the current processing sub-batch after cleaning according to the damage newly identified after cleaning that does not belong to the original epidermal defects exposed after cleaning and the expansion of the epidermal defects identified before cleaning after cleaning; The control device is also used to determine the final grade of blueberries based on the fruit diameter information, the second appearance information, the original skin defects exposed after washing, and the appearance re-inspection standard parameters, and to control the sorting device to sort the blueberries according to the final grade; it is also used to adjust or maintain the washing intensity parameters and conveying speed parameters used in the next processing sub-batch of the same parameter matching category based on the residual attachment index and the washing new damage index.

9. The blueberry sorting and cleaning system based on appearance defects and fruit diameter according to claim 8, characterized in that, Both the first and second detection devices include an image acquisition component for acquiring appearance information of blueberries from multiple directions; The upstream of the first detection device is provided a flattening guide section or a vibration dispersion section for forming a single-layer conveying state of blueberries. The first detection device and the second detection device are provided with mutually separated conveying channels, which are used to restrict blueberries from passing each other or changing positions across channels. The multiple cache channels are single-column cache channels corresponding to parameter matching categories. A diversion execution mechanism connected to the control device is provided between the first detection device and the batch cache device. The diversion execution mechanism is used to send blueberries into the corresponding single-column cache channel. The controllable material stop is used to release blueberries to form a processing sub-batch according to a preset quantity or preset cache time.

Images