A fruit quality grading system and method based on big data construction point cloud data

Abstract

The application provides a fruit quality grading system and method based on big data and point cloud data, and relates to the technical field of food processing, which comprises the following steps: performing external detection and internal detection on each fruit; constructing point cloud data according to the data of external detection and internal detection, and generating multi-dimensional data of each fruit; generating a unique code according to the multi-dimensional data of each fruit according to coding logic, and dividing each fruit into multiple batch numbers according to the coding difference degree of each fruit, and distributing the multi-dimensional data of each fruit according to the batch numbers; distributing and storing the fruits according to the batch numbers; setting multiple quality characteristics, comparing the quality coefficients of each batch, distinguishing the quality levels of the fruits, and generating corresponding keywords. The application uses big data processing, can form a data model of each production cycle by combining the average level of the production area and the existing cognition.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of food processing, in particular to a fruit quality grading system and method based on big data to construct point cloud data. BACKGROUND

[0002] In the prior art, fruit grading is an important process for classifying fruits into different grades according to their quality, size, appearance, and other characteristics.

[0003] The traditional method is to rely on manual fruit grading, where trained professionals use their senses such as eyesight and touch to assess the quality and appearance of fruits, and they will classify fruits into different grades such as first grade, second grade, third grade, etc. according to absolute characteristics.

[0004] However, using absolute characteristics for fruit sorting usually relies on pre-set standards or thresholds, and this method cannot fully consider the differences in characteristics of fruits from different origins and different batches, so it may lead to overly strict or overly lenient sorting results, and cannot accurately assess the actual quality of fruits from each origin and each batch, burying the characteristics of most fruits, such as some people preferring sour fruits and some people preferring sweet fruits, but this one-size-fits-all characteristic makes all fruits selected by the same standard; and using absolute characteristics for fruit sorting lacks flexibility and adaptability, affecting the accuracy and efficiency of fruit sorting.

[0005] Therefore, there is an urgent need for a fruit quality grading system based on big data to construct point cloud data to solve the technical problems of poor accuracy and efficiency of sorting in traditional methods and inability to accurately assess the actual quality of fruits from each origin and each batch, burying the characteristics of most fruits. SUMMARY

[0006] To solve the problems of the prior art, the present application provides a fruit quality grading system and method based on big data to construct point cloud data, which solves the technical problems of poor accuracy and efficiency of sorting in traditional methods and inability to accurately assess the actual quality of fruits from each origin and each batch, burying the characteristics of most fruits.

[0007] In one aspect, the present application provides a fruit quality grading system based on big data to construct point cloud data, comprising:

[0008] a data acquisition module for external detection and internal detection of each fruit;

[0009] a data construction module for constructing point cloud data based on the data of the external detection and the internal detection, and generating multi-dimensional data for each fruit;

[0010] The data storage module generates a unique code according to the multi-dimensional data of each fruit based on coding logic, and divides each fruit into batches according to the coding difference, and stores the multi-dimensional data of each fruit according to the batch number.

[0011] The sorting module sorts and stores the fruits according to the batch number.

[0012] The data processing module analyzes the point cloud data to set multiple quality characteristics, sets a weight for each quality characteristic, analyzes the multi-dimensional data of each batch of fruits according to the weight, and generates a quality coefficient.

[0013] The quality grading module compares the quality coefficient of each batch to distinguish the quality level of the fruits and generates a keyword corresponding to the batch number.

[0014] Preferably, the data acquisition module uses a three-dimensional scanner to scan the fruit and obtain the three-dimensional shape information of the fruit's exterior, and generates point cloud data containing the point coordinates, depth information and color attributes of the fruit's surface.

[0015] The data acquisition module also obtains two-dimensional grayscale images of the fruit's interior through X-ray photography of at least two vertical planes.

[0016] The data construction module processes the three-dimensional shape information of the fruit's exterior, including denoising, point cloud registration and surface reconstruction, to improve data quality and accuracy.

[0017] The data construction module also processes the two-dimensional grayscale images of the fruit's interior, including denoising, edge detection and image enhancement operations, to extract the internal contour and internal features of the fruit.

[0018] The data construction module also uses a corresponding point matching algorithm to match the internal contour and internal features of the fruit with the point coordinates and depth information of the surface based on the corresponding relationship between the two-dimensional images and the three-dimensional data, to generate multi-dimensional data for each fruit.

[0019] Preferably, the data storage module is pre-set with a home threshold, each coded data is provided with an upper floating interval and a lower floating interval according to the home threshold, when the multi-dimensional information of a fruit is inputted, a first code is generated according to the coding logic, when a second fruit is inputted, a second code is generated, the first code and the second code are compared, when the second code falls into the upper floating interval or the lower floating interval of the first code, a new batch number is not generated, when the second code does not belong to the upper floating interval and the lower floating interval, a new batch number is established, when the nth fruit is inputted, the coding with the highest similarity is screened, the nth code and the coding with the highest similarity are compared, when the nth code falls into the upper floating interval or the lower floating interval of the coding with the highest similarity, a new batch number is not generated, when the nth code does not belong to the upper floating interval and the lower floating interval, a new batch number is established, and the distributed stored data is indexed according to the batch number, so as to quickly search and query.

[0020] Preferably, the data processing module extracts the quality evaluation standard of the fruit from the historical data, sets the logical relationship between each quality feature and the point coordinates, depth information and color attribute of the fruit surface and the internal contour and internal feature, and sets the corresponding weight for the quality feature according to the logical relationship.

[0021] Preferably, the data processing module performs data normalization processing on the multi-dimensional data of each fruit in each batch number, performs weighted calculation according to the corresponding weight set for the quality feature, and calculates the quality coefficient of the fruit of each batch according to the weighted feature value, the quality coefficient is a comprehensive evaluation index, which reflects the comprehensive performance of the fruit on the quality feature according to historical experience.

[0022] Preferably, the quality grading module is pre-set with a quality threshold L, the batch Y max with the maximum quality coefficient and the batch Y min with the minimum quality coefficient are extracted, the quality floating condition of the fruit is compared and the number P of grades is calculated, and the calculation formula is:

[0023] P=floor((Y max -Y min ) / L)+1

[0024] Preferably, the quality grading module is pre-set with a group of keyword thresholds, the quality coefficient is mapped to different keywords according to the range of the quality coefficient, and the keywords are generated according to the high and low of the quality coefficient, the size of the feature value and other factors.

[0025] On the other hand, the present application also proposes a fruit quality grading method based on point cloud data of big data, comprising:

[0026] Step S1: performing external detection and internal detection on each fruit;

[0027] Step S2: Construct point cloud data based on the data from the external detection and the internal detection to generate multi-dimensional data for each fruit;

[0028] Step S3: Generate a unique code according to the coding logic based on the multi-dimensional data of each fruit, and divide the fruit into multiple batch numbers according to the coding difference of each fruit. Distribute the multi-dimensional data of each fruit according to the batch number.

[0029] Step S4: Sort and store the fruits according to the batch number;

[0030] Step S5: Analyze the point cloud data to set multiple quality features, assign weights to each quality feature, analyze the multi-dimensional data of each batch of fruit based on the weights, and generate quality coefficients.

[0031] Step S6: Compare the quality coefficient of each batch, distinguish the quality level of the fruit, and generate the keywords corresponding to the batch number.

[0032] Preferably, in step S1, each fruit undergoes external and internal testing, including:

[0033] A 3D scanner is used to scan the fruit to obtain the 3D shape information of the fruit's exterior, and point cloud data is generated, which includes the point coordinates, depth information and color attributes of the fruit's surface.

[0034] In step S2, the step of constructing point cloud data based on external and internal detection data to generate multi-dimensional data for each fruit includes:

[0035] The data acquisition module also acquires two-dimensional grayscale images of the inside of the fruit by X-ray imaging in at least two vertical planes;

[0036] The external three-dimensional shape information of the fruit is processed, including denoising, point cloud registration, and surface reconstruction, to improve data quality and accuracy;

[0037] The internal two-dimensional grayscale image of the fruit is processed, including operations such as noise reduction, edge detection and image enhancement, to extract the internal contour and internal features of the fruit;

[0038] The corresponding point matching algorithm is used to match the internal contour and internal features of the fruit with the point coordinates and depth information of the surface based on the correspondence between the two-dimensional image and the three-dimensional data, so as to generate multi-dimensional data for each fruit.

[0039] In step S3, the fruit is divided into multiple batch numbers based on the coding differences of each fruit, and the multi-dimensional data of each fruit is distributed and stored according to the batch numbers, including:

[0040] A home threshold is preset, each code is provided with an upper floating interval and a lower floating interval according to the home threshold, when multi-dimensional information of a fruit is inputted, a first code is generated according to the coding logic, when a second fruit is inputted, a second code is generated, the first code and the second code are compared, when the second code falls into the upper floating interval or the lower floating interval of the first code, a new batch number is not generated, when the second code does not belong to the upper floating interval and the lower floating interval, a new batch number is established, when an n-th fruit is inputted, the highest similarity code of the n-th fruit is screened, the n-th code and the highest similarity code are compared, when the n-th code falls into the upper floating interval or the lower floating interval of the highest similarity code, a new batch number is not generated, when the n-th code does not belong to the upper floating interval and the lower floating interval, a new batch number is established, and an index is established for the distributed storage data according to the batch number, so as to quickly search and query.

[0041] Preferably, the point cloud data is analyzed to set multiple quality characteristics, and a weight is set for each quality characteristic, including:

[0042] The historical data of the fruit quality evaluation standard is extracted, the logical relationship between each quality characteristic and the point coordinates, depth information and color attribute of the fruit surface and the internal contour and internal characteristics is set, and the corresponding weight of the quality characteristic is set according to the logical relationship;

[0043] In the step S5, the multi-dimensional data of each batch of fruits is analyzed according to the weight to generate a quality coefficient, including:

[0044] The multi-dimensional data of each fruit in each batch number is subjected to data normalization processing, and the corresponding weight is set according to the quality characteristic for weighted calculation, and the quality coefficient of each batch of fruits is calculated according to the characteristic value after weighting, the quality coefficient is a comprehensive evaluation index, which reflects the comprehensive performance of the fruit on the quality characteristic according to historical experience;

[0045] In the step S6, the quality coefficients of each batch are compared to distinguish the quality levels of the fruits, including:

[0046] A quality threshold L is preset, the batch Y max with the minimum quality coefficient is extracted min , the quality floating condition of the fruit is compared and the number P of levels is calculated, and the calculation formula is:

[0047] P = floor((Y max -Y min ) / L) + 1

[0048] In the step S6, the keyword corresponding to the batch number is generated, including:

[0049] A set of keyword threshold values are set, which are mapped to different keyword extraction according to the range of quality coefficients, and the generation of keywords is performed according to the high and low of the quality coefficients, the size of the characteristic values and other factors.

[0050] The beneficial effects of the present application are:

[0051] The sorting module of the present application can efficiently sort fruits according to multi-dimensional information, and through big data processing, the average level of the production area can be combined with existing cognition to generate a data model for each production cycle.

[0052] In addition, after big data processing according to the quality grading of the present application, the quality of fruits in each production area can be tracked and analyzed, and the data of each production cycle can be efficiently compared to determine the average level and quality fluctuation of the production area, which helps to develop quality control strategies and improve the production process, and continuously approaches the quality evaluation standard of each fruit in the historical data. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 The structure diagram of the fruit quality grading system based on big data point cloud data of the present application;

[0054] Figure 2 The flowchart of the fruit quality grading method based on big data point cloud data of the present application. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0056] On the one hand, referring to Figure 1 The present embodiment proposes a fruit quality grading system based on big data point cloud data, which includes:

[0057] The data acquisition module performs external detection and internal detection on each fruit;

[0058] The data construction module constructs point cloud data according to the data of external detection and internal detection, and generates multi-dimensional data of each fruit;

[0059] The data storage module generates a unique code according to the multi-dimensional data of each fruit according to the coding logic, and divides the multi-dimensional data of each fruit into multiple batch numbers according to the coding difference degree of each fruit, and stores the multi-dimensional data of each fruit in a distributed manner according to the batch number.

[0060] a sorting module that sorts and stores the fruits according to batch numbers;

[0061] a data processing module that analyzes the point cloud data to set multiple quality characteristics, sets weights for each quality characteristic, analyzes the multi-dimensional data of each batch of fruits according to the weights, and generates a quality coefficient;

[0062] a quality grading module that compares the quality coefficients of each batch to distinguish the quality levels of the fruits and generates keywords corresponding to the batch numbers.

[0063] It can be understood that through external detection and internal detection, data on multiple aspects of the fruits such as appearance characteristics, internal structure, and maturity can be obtained. This makes it possible to comprehensively evaluate and analyze the fruits. The collected data is constructed into point cloud data, and a unique code and batch number are generated according to the multi-dimensional data of each fruit. Through distributed storage, the data of each fruit can be efficiently managed and queried, facilitating subsequent processing and analysis. And according to the batch number, the fruits are sorted and stored, which can realize the rapid sorting and management of the fruits. This helps to improve the efficiency and accuracy of the production line. In addition, through the data processing module, the point cloud data is analyzed, multiple quality characteristics are set and weights are set, the multi-dimensional data of each batch of fruits is analyzed according to the weights, and a quality coefficient is generated. This makes it possible to objectively evaluate and compare the quality of the fruits. Finally, by comparing the quality coefficients of each batch, the quality levels of the fruits can be distinguished, and corresponding keywords can be generated for each batch number, which provides a useful reference and guidance for quality management and market positioning.

[0064] In some embodiments of the present application, the data acquisition module uses a three-dimensional scanner to scan the fruits and obtain three-dimensional shape information of the fruits outside, generating point cloud data containing point coordinates, depth information and color attributes of the fruit surface;

[0065] The data acquisition module also acquires two-dimensional grayscale images of the fruit inside through X-ray shooting of at least two vertical planes;

[0066] The data construction module processes the external three-dimensional shape information of the fruits, including denoising, point cloud registration and surface reconstruction, to improve the data quality and accuracy;

[0067] The data construction module also processes the internal two-dimensional grayscale images of the fruits, including denoising, edge detection and image enhancement operations, to extract the internal contour and internal features of the fruits.

[0068] The data construction module also uses a corresponding point matching algorithm to match the internal contour and internal features of the fruits with the point coordinates and depth information of the surface according to the corresponding relationship between the two-dimensional images and the three-dimensional data, generating multi-dimensional data for each fruit.

[0069] In some embodiments of the present application, a home threshold is preset in the data storage module, each coded data is provided with an upper floating interval and a lower floating interval according to the home threshold, when the multi-dimensional information of a fruit is input, a first code is generated according to the coding logic, when a second fruit is input, a second code is generated, the first code and the second code are compared, when the second code falls within the upper floating interval or the lower floating interval of the first code, a new batch number is not generated, when the second code does not belong to the upper floating interval and the lower floating interval, a new batch number is established, when the nth fruit is input, the coding with the highest degree of similarity is screened, the nth code and the coding with the highest degree of similarity are compared, when the nth code falls within the upper floating interval or the lower floating interval of the coding with the highest degree of similarity, a new batch number is not generated, when the nth code does not belong to the upper floating interval and the lower floating interval, a new batch number is established, and an index is established for the distributed stored data according to the batch number, so as to quickly retrieve and query.

[0070] In some embodiments of the present application, the data processing module extracts the quality evaluation standard of the fruit from the historical data, sets the logical relationship between each quality characteristic and the point coordinates, depth information and color attribute of the fruit surface and the internal contour and internal characteristics, and sets the corresponding weight for the quality characteristic according to the logical relationship.

[0071] In some embodiments of the present application, the data processing module performs data normalization processing on the multi-dimensional data of each fruit in each batch number, and performs weighted calculation according to the corresponding weight set for the quality characteristic, and calculates the quality coefficient of the fruit of each batch according to the weighted characteristic value. The quality coefficient is a comprehensive evaluation index, which reflects the comprehensive performance of the fruit on the quality characteristics according to historical experience.

[0072] In some embodiments of the present application, the quality grading module is provided with a quality threshold L, and the batch Y max with the maximum quality coefficient and the batch Y min with the minimum quality coefficient are extracted, the quality floating condition of the fruit is compared and the number of grades P is calculated, and the calculation formula is:

[0073] P = floor((Y max -Y min ) / L) + 1

[0074] In some embodiments of the present application, the quality grading module is provided with a set of keyword thresholds, the quality coefficient is mapped to different keywords according to the range of the quality coefficient, and the keywords are generated according to the high and low of the quality coefficient, the size of the characteristic value and other factors.

[0075] On the other hand, referring to Figure 2 , the present embodiment also proposes a fruit quality grading method based on point cloud data of big data, which comprises:

[0076] Step S1: external detection and internal detection are performed on each fruit;

[0077] Step S2: point cloud data is constructed according to the data of external detection and internal detection, and multi-dimensional data of each fruit is generated;

[0078] Step S3: a unique code is generated according to the multi-dimensional data of each fruit according to the encoding logic, and each fruit is divided into multiple batch numbers according to the coding difference, and the multi-dimensional data of each fruit is distributed according to the batch number;

[0079] Step S4: the fruits are shunted and stored according to the batch number;

[0080] Step S5: multiple quality characteristics are set by analyzing the point cloud data, and the weight of each quality characteristic is set, and the multi-dimensional data of each batch of fruits is analyzed according to the weight to generate a quality coefficient;

[0081] Step S6: compare the quality coefficient of each batch, distinguish the quality level of the fruit, and generate the corresponding keyword of the batch number.

[0082] In some embodiments of the present application, in step S1, external detection and internal detection are performed on each fruit, including:

[0083] The three-dimensional scanner is used to scan the fruit to obtain the three-dimensional shape information of the fruit outside, and generate point cloud data containing point coordinates, depth information and color attributes of the fruit surface;

[0084] In step S2, the point cloud data is constructed according to the data of external detection and internal detection, and the multi-dimensional data of each fruit is generated, including:

[0085] The data acquisition module also acquires two-dimensional gray images of the fruit inside through X-ray shooting of at least two vertical planes;

[0086] The external three-dimensional shape information of the fruit is processed, including denoising, point cloud registration and surface reconstruction, to improve the data quality and accuracy;

[0087] The internal two-dimensional gray image of the fruit is processed, including denoising, edge detection and image enhancement operations, to extract the internal contour and internal features of the fruit;

[0088] The internal contour and internal features of the fruit are matched with the point coordinates, depth information of the surface using a corresponding point matching algorithm according to the corresponding relationship between the two-dimensional image and the three-dimensional data, to generate multi-dimensional data of each fruit;

[0089] In step S3, the fruits are divided into batches according to the coding difference, and the multi-dimensional data of each fruit is stored in a distributed manner according to the batch number, including:

[0090] The preset home threshold is used to set the upper floating interval and the lower floating interval for each code, when the multi-dimensional information of a fruit is input, a first code is generated according to the coding logic, when the second fruit is input, a second code is generated, the first code and the second code are compared, when the second code falls within the upper floating interval or the lower floating interval of the first code, a new batch number is not generated, when the second code does not belong to the upper floating interval and the lower floating interval, a new batch number is established, when the nth fruit is input, the code with the highest degree of similarity is selected, the nth code and the code with the highest degree of similarity are compared, when the nth code falls within the upper floating interval or the lower floating interval of the code with the highest degree of similarity, a new batch number is not generated, when the nth code does not belong to the upper floating interval and the lower floating interval, a new batch number is established, and an index is established for the distributed stored data according to the batch number, so as to quickly search and query.

[0091] In some embodiments of the present application, the point cloud data is analyzed to set multiple quality characteristics, and a weight is set for each quality characteristic, including:

[0092] The quality evaluation standard of the fruit is extracted from the historical data, the logical relationship between each quality characteristic and the point coordinates, depth information and color attribute of the fruit surface, and the internal contour and internal characteristics is set, and the corresponding weight is set for the quality characteristic according to the logical relationship;

[0093] In step S5, the multi-dimensional data of each batch of fruits is analyzed according to the weight to generate a quality coefficient, including:

[0094] The multi-dimensional data of each fruit in each batch number is subjected to data normalization processing, and the corresponding weight is set according to the quality characteristic for weighted calculation, and the quality coefficient of each batch of fruits is calculated according to the weighted characteristic value, the quality coefficient is a comprehensive evaluation index, which reflects the comprehensive performance of the fruit on the quality characteristic according to the historical experience;

[0095] In step S6, the quality coefficients of each batch are compared to distinguish the quality level of the fruit, including:

[0096] A quality threshold L is preset, the batch Y max with the maximum quality coefficient and the batch Y min with the minimum quality coefficient are extracted, the quality floating condition of the fruit is compared and the number P of levels is calculated, and the calculation formula is:

[0097] P = floor((Y max -Y min ) / L) + 1

[0098] In step S6, the keywords corresponding to the batch number are generated, including:

[0099] There is a set of keyword threshold values, which are mapped to different keyword extraction according to the range of quality coefficients, and the generation of keywords is performed according to the high and low of the quality coefficient, the size of the characteristic value and other factors.

[0100] The fruit quality grading system and method based on big data to construct point cloud data according to the embodiment have the following advantages:

[0101] The sorting module of the embodiment can efficiently sort fruits according to multi-dimensional information, and through big data processing, the average level of the production area can be combined with the existing cognition to generate a data model for each production cycle.

[0102] In addition, after the quality grading according to the embodiment is processed by big data, the quality of the fruit of each production area can be tracked and analyzed, the data of each production cycle can be efficiently compared, and the average level and quality fluctuation of the production area can be determined, which helps to formulate quality control strategies and improve the production process, and continuously approaches the fruit quality evaluation standard in the historical data.

[0103] The present application can also have other various embodiments, and those skilled in the art can make various corresponding changes and modifications according to the present application without departing from the spirit and essence of the present application, but these corresponding changes and modifications should all belong to the protection scope of the claims attached to the present application.

Claims

1. A fruit quality grading system based on point cloud data constructed from big data, characterized in that, include: The data acquisition module performs external and internal testing on each fruit. The data construction module constructs point cloud data based on the data from the external and internal detections, generating multi-dimensional data for each fruit. The data storage module generates a unique code based on the multi-dimensional data of each fruit according to the coding logic, and divides the data into multiple batch numbers according to the coding difference of each fruit, and stores the multi-dimensional data of each fruit in a distributed manner according to the batch number. The sorting module sorts and stores the fruits according to the batch number; The data processing module analyzes the point cloud data to set multiple quality features, assigns weights to each quality feature, analyzes the multi-dimensional data of each batch of fruit based on the weights, and generates a quality coefficient. The quality grading module compares the quality coefficient of each batch, distinguishes the quality level of the fruit, and generates keywords corresponding to the batch number. The data storage module has a preset attribution threshold. Each code has an upper and lower floating range based on the attribution threshold. When multi-dimensional information of a fruit is entered, a first code is generated according to the coding logic. When a second fruit is entered, a second code is generated. The first code and the second code are compared. If the second code falls within the upper or lower floating range of the first code, no new batch number is generated. If the second code does not belong to the upper or lower floating range, a new batch number is created. When the nth fruit is entered, the code with the highest similarity is selected. The nth code is compared with the closest code. If the nth code falls within the upper or lower floating range of the closest code, no new batch number is generated. If the nth code does not belong to the upper or lower floating range, a new batch number is created. An index is built based on the batch number for distributed storage data to facilitate fast retrieval and querying.

2. The fruit quality grading system based on big data and point cloud data as described in claim 1, characterized in that, The data acquisition module uses a 3D scanner to scan the fruit to obtain the 3D shape information of the fruit's exterior, generating point cloud data that includes the point coordinates, depth information, and color attributes of the fruit's surface. The data acquisition module also acquires two-dimensional grayscale images of the inside of the fruit by X-ray imaging in at least two vertical planes; The data construction module processes the external three-dimensional shape information of the fruit, including denoising, point cloud registration, and surface reconstruction, to improve data quality and accuracy. The data construction module also processes the internal two-dimensional grayscale image of the fruit, including operations such as noise reduction, edge detection and image enhancement, to extract the internal contour and internal features of the fruit; The data construction module also uses a corresponding point matching algorithm to match the internal contours and features of the fruit with the point coordinates and depth information of the surface based on the correspondence between the two-dimensional image and the three-dimensional data, thereby generating multi-dimensional data for each fruit.

3. The fruit quality grading system based on point cloud data constructed from big data as described in claim 1, characterized in that, The data processing module extracts historical data to evaluate the quality of fruits, sets logical relationships between each quality feature and the point coordinates, depth information, and color attributes on the fruit surface, as well as with the internal contours and internal features, and assigns corresponding weights to the quality features based on these logical relationships.

4. The fruit quality grading system based on big data and point cloud data as described in claim 3, characterized in that, The data processing module normalizes the multi-dimensional data of each fruit within each batch number, performs weighted calculations based on the quality characteristics, and calculates the quality coefficient of each batch of fruit based on the weighted feature values. The quality coefficient is a comprehensive evaluation index that reflects the overall performance of the fruit in terms of quality characteristics based on historical experience.

5. The fruit quality grading system based on big data and point cloud data as described in claim 4, characterized in that, The quality grading module is preset with a quality threshold L, and extracts the batch Y with the highest quality coefficient. max And the batch with the lowest quality coefficient Y min Compare the quality fluctuations of the fruit and calculate the grade number P. The calculation formula is as follows:

6. The fruit quality grading system based on big data and point cloud data as described in claim 5, characterized in that, The quality grading module sets a set of keyword thresholds, maps them to different keyword extractions according to the range of quality coefficients, and generates keywords based on factors such as the quality coefficient and the size of the feature value.

7. A method for fruit quality grading based on point cloud data constructed from big data, applied to the fruit quality grading system based on point cloud data constructed from big data as described in any one of claims 1-6, characterized in that, include: Step S1: Perform external and internal testing on each fruit; Step S2: Construct point cloud data based on the data from the external detection and the internal detection to generate multi-dimensional data for each fruit; Step S3: Generate a unique code according to the coding logic based on the multi-dimensional data of each fruit, and divide the fruit into multiple batch numbers according to the coding difference of each fruit. Distribute the multi-dimensional data of each fruit according to the batch number. Step S4: Sort and store the fruits according to the batch number; Step S5: Analyze the point cloud data to set multiple quality features, assign weights to each quality feature, analyze the multi-dimensional data of each batch of fruit based on the weights, and generate quality coefficients. Step S6: Compare the quality coefficient of each batch, distinguish the quality level of the fruit, and generate the keywords corresponding to the batch number.

8. The fruit quality grading method based on big data point cloud data as described in claim 7, characterized in that, In step S1, each fruit undergoes external and internal testing, including: A 3D scanner is used to scan the fruit to obtain the 3D shape information of the fruit's exterior, and point cloud data is generated, which includes the point coordinates, depth information and color attributes of the fruit's surface. In step S2, point cloud data is constructed based on external and internal detection data to generate multi-dimensional data for each fruit, including: The data acquisition module also acquires two-dimensional grayscale images of the inside of the fruit by X-ray imaging in at least two vertical planes; The external three-dimensional shape information of the fruit is processed, including denoising, point cloud registration, and surface reconstruction, to improve data quality and accuracy; The internal two-dimensional grayscale image of the fruit is processed, including operations such as noise reduction, edge detection and image enhancement, to extract the internal contour and internal features of the fruit; The corresponding point matching algorithm is used to match the internal contour and internal features of the fruit with the point coordinates and depth information of the surface based on the correspondence between the two-dimensional image and the three-dimensional data, so as to generate multi-dimensional data for each fruit. In step S3, the fruit is divided into multiple batch numbers based on the coding differences of each fruit, and the multi-dimensional data of each fruit is distributed and stored according to the batch numbers, including: A preset attribution threshold is set. Each code has an upper and lower floating range based on the attribution threshold. When multi-dimensional information of a fruit is entered, a first code is generated according to the coding logic. When a second fruit is entered, a second code is generated. The first code and the second code are compared. If the second code falls within the upper or lower floating range of the first code, no new batch number is generated. If the second code does not belong to the upper or lower floating range, a new batch number is created. When the nth fruit is entered, the code with the highest similarity is selected. The nth code is compared with the closest code. If the nth code falls within the upper or lower floating range of the closest code, no new batch number is generated. If the nth code does not belong to the upper or lower floating range, a new batch number is created. An index is built based on the batch number as distributed storage data for fast retrieval and querying.

9. The fruit quality grading method based on point cloud data constructed from big data according to claim 7, characterized in that, The point cloud data is analyzed by setting multiple quality features, and a weight is assigned to each quality feature, including: Extract historical data to evaluate the quality of fruit, set logical relationships between each quality feature and the point coordinates, depth information, and color attributes on the fruit surface, as well as with the internal contour and internal features, and assign corresponding weights to the quality features based on the logical relationships. In step S5, multi-dimensional data of each batch of fruit is analyzed based on weights to generate a quality coefficient, including: The multi-dimensional data of each fruit in each batch number is normalized, and weighted calculation is performed according to the quality characteristics. The quality coefficient of each batch of fruit is calculated based on the weighted feature values. The quality coefficient is a comprehensive evaluation index that reflects the overall performance of the fruit in terms of quality characteristics based on historical experience. In step S6, comparing the quality coefficients of each batch to distinguish the fruit quality level includes: Given a preset quality threshold L, the batch Y with the highest quality coefficient is extracted. max And the batch with the lowest quality coefficient Y min Compare the quality fluctuations of the fruit and calculate the grade number P. The calculation formula is as follows: In step S6, the keywords corresponding to the batch number are generated, including: A set of keyword thresholds is set, and the thresholds are mapped to different keywords based on the range of quality coefficients. Keywords are generated based on factors such as the quality coefficient and the size of the feature value.

Images