Blockchain-based intelligent management method for rice germplasm resources

Abstract

This invention discloses a blockchain-based intelligent management method for rice germplasm resources, relating to the fields of blockchain data storage and rice germplasm resource management. The method acquires initial morphological image sequences and initial physiological parameter sets of rice germplasm resources, extracts external contour features and internal component features, and cross-maps these features to generate a resource identification code. The resource identification code is written into the current blockchain storage block, and a first verification field is calculated based on the block hash value of the current storage block and the resource identification code. Upon release, the current morphological image sequence and current physiological parameter set are collected, and a resource code to be verified is generated by cross-mapping according to the same features. Release permissions are verified based on the resource code to be verified and the first verification field. By fusing morphological and physiological features into a single feature identification code and using blockchain storage verification, the problem of easy duplication and confusion of single feature identifiers is reduced, enhancing the tamper-proof and traceability of germplasm resource release permission verification.

Description

Technical Field

[0001] This invention relates to the fields of blockchain data storage technology and rice germplasm resource management technology, specifically a blockchain-based intelligent management method for rice germplasm resources. Background Technology

[0002] In the preservation and utilization of rice germplasm resources, current management methods typically rely on a combination of manual registration, labeling, barcodes or QR codes, and centralized databases. During the entry process, the germplasm name, origin, appearance description, and some physiological indicators are recorded. During the exit process, authorization is confirmed by manually verifying label information or querying database records. Some management systems perform single external contour recognition on images of rice seeds or plants, or detect physiological parameters such as protein content, starch content, and moisture content and compare them with the entry records.

[0003] In existing solutions, there is a lack of an inseparable binding relationship between the label and the individual germplasm. After the label is copied, worn, or replaced, the centralized database records struggle to automatically detect anomalies. Single external contour recognition has limited effectiveness in distinguishing between rice germplasms with similar appearances, and single physiological parameter comparison cannot reflect differences in external morphology. Using these two methods independently makes it difficult to form a stable resource identification code. Germplasm files and outbound permission records in the centralized database are at risk of being tampered with, accidentally deleted, or having their permissions bypassed. The outbound verification process relies on single-point records and cannot independently verify the identification code formed upon entry and the code to be verified upon exit. Rice germplasm resources are typically entered and exited with a long interval, and the conditions for morphological image acquisition and physiological parameter measurement differ. Existing methods lack a mechanism for restoring and comparing the fused feature code from the entry point during outbound verification, making the outbound verification results susceptible to fluctuations in data at a single moment.

[0004] To address the aforementioned issues, a management method is needed that can integrate the external morphological characteristics and internal component characteristics of rice germplasm resources to form a stable identity identifier, and independently verify the stored data and identity identifier upon release from storage, in order to avoid the unreliability of identity and permissions caused by the instability of single feature identification and the tampering of centralized data. Summary of the Invention

[0005] To address the current issues in rice germplasm resource management, such as unstable identification of single morphological or physiological characteristics, easy replacement or tampering of identifiers and database records, and reliance on centralized records for outbound permission verification, this invention provides a blockchain-based intelligent management method for rice germplasm resources. This method generates resource identifier codes through cross-mapping of external contour features and internal component features, and uses blockchain block hashes and resource identifier codes to calculate verification fields for outbound permission verification. This reduces the risk of germplasm resource identifiers being copied or replaced, and enhances the reliability of outbound permission verification.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides a blockchain-based intelligent management method for rice germplasm resources, comprising: responding to a germplasm resource entry request, acquiring an initial morphological image sequence and an initial physiological parameter set of the rice germplasm resource; extracting external contour features of the germplasm resource based on the initial morphological image sequence, and extracting internal component features of the germplasm resource based on the initial physiological parameter set; performing feature cross-mapping between the external contour features and the internal component features to generate a resource identifier code for the germplasm resource; writing the resource identifier code into the current evidence block of the blockchain, and calculating a first verification field based on the block hash value of the current evidence block and the resource identifier code; when the germplasm resource is released from the blockchain, collecting the current morphological image sequence and the current physiological parameter set of the germplasm resource, and generating a resource code to be verified according to the same feature cross-mapping method; verifying the release permission of the germplasm resource based on the resource code to be verified and the first verification field. This method generates a unique resource identifier code by cross-mapping the morphological characteristics and internal physiological components of rice germplasm resources, and uses the hash information of the blockchain notarization block to generate a verification field. The consistency verification of the resource code to be verified is carried out in the outbound process, which can realize intelligent management and anti-counterfeiting traceability of rice germplasm resources in the whole process of warehousing, notarization and outbound, and improve the accuracy and security of resource management.

[0007] Preferably, extracting the external contour features of germplasm resources based on the initial morphological image sequence includes: performing frame-by-frame binarization on the initial morphological image sequence to obtain a set of germplasm region boundary pixels in each frame; calculating the geometric center of the set of germplasm region boundary pixels in each frame and aligning the geometric centers of each frame to a unified coordinate system; statistically analyzing the superposition frequency distribution of the germplasm region boundary pixel sets in each aligned frame in the unified coordinate system; and extracting the set of boundary pixel coordinates with frequencies higher than a preset frequency threshold as the external contour features based on the superposition frequency distribution. By aligning the geometric centers of multiple frames and filtering the superposition frequency, contour deviations caused by pose changes or local noise in a single frame image can be reduced, improving the robustness and consistency of external contour feature extraction.

[0008] Preferably, the step of extracting the internal component features of the germplasm resource based on the initial physiological parameter set includes: obtaining the protein content, starch content, and moisture content of the germplasm resource from the initial physiological parameter set; mapping the protein content, starch content, and moisture content to preset protein quantification intervals, starch quantification intervals, and moisture quantification intervals, respectively, to obtain protein quantification values, starch quantification values, and moisture quantification values; concatenating the protein quantification values, starch quantification values, and moisture quantification values ​​into an internal component feature vector according to a preset arrangement order; converting each component in the internal component feature vector into binary code according to a preset bit width, and concatenating the binary code order of each component into the internal component feature. This allows the internal quality parameters of rice germplasm resources to be converted into a unified binary feature expression, facilitating cross-mapping with external contour features and ensuring the repeatability and stability of the resource identification coding generation process.

[0009] Preferably, the step of performing feature cross-mapping between the external contour features and the internal component features to generate a resource identification code for the germplasm resource includes: sorting the external contour features according to the horizontal coordinate values ​​of the contour pixels to obtain a contour pixel sequence; calculating the Euclidean distance between every two adjacent pixels in the contour pixel sequence to obtain a distance sequence; determining the two pixels corresponding to the maximum value in the distance sequence as a first anchor point and a second anchor point; determining a baseline segment using the first anchor point and the second anchor point; projecting the internal component features onto the baseline segment along a direction perpendicular to the baseline segment to obtain a projection component; weighting and fusing the length value of the baseline segment and the projection component according to a preset weighting coefficient, and quantizing the fusion result as the resource identification code. By fusing the geometric anchor points of the external contour and the projection of the internal component features onto the baseline segment, the generated resource identification code simultaneously associates the apparent morphology and intrinsic quality of the rice germplasm resource, enhancing the uniqueness and identification capability of the resource identification code.

[0010] Preferably, calculating the first verification field based on the block hash value of the current evidence storage block and the resource identifier code includes: extracting the block hash value from the block header of the current evidence storage block; truncating the block hash value to a fixed byte length to obtain a hash truncation segment; interleaving the hash truncation segment and the resource identifier code in an alternating byte arrangement to obtain an interleaved sequence; performing cyclic redundancy check calculation on the interleaved sequence to obtain the first verification field; and storing the resource identifier code and the first verification field together in the transaction record of the current evidence storage block. By interleaving the block hash value and the resource identifier code and performing cyclic redundancy check, the first verification field can be bound to the resource identifier code and its block, preventing the resource identifier code from being replaced or tampered with, and improving the credibility of the evidence storage information.

[0011] Preferably, the step of collecting the current morphological image sequence and current physiological parameter set of the germplasm resource when it is taken out of storage, and generating the resource code to be verified according to the same feature cross-mapping method, includes: in response to the germplasm resource take-out request, acquiring the current morphological image sequence and current physiological parameter set of the germplasm resource at the current moment; extracting the current contour features based on the current morphological image sequence, and extracting the current component features based on the current physiological parameter set; and generating the resource code to be verified by applying the current contour features and the current component features to the same feature cross-mapping method. Since the generation method of the resource code to be verified is consistent with the generation method of the resource identification code when it is put into storage, the actual germplasm resource characteristics at the time of take-out can be effectively compared with the evidence characteristics at the time of storage, thereby improving the reliability of the verification results.

[0012] Preferably, the step of verifying the release permission of the germplasm resource based on the resource code to be verified and the first verification field includes: reading the resource identifier code and the first verification field corresponding to the germplasm resource from the transaction records of the current evidence block; performing a bitwise XOR operation between the resource code to be verified and the read resource identifier code to obtain a difference value; if the difference value is zero, performing a cyclic redundancy check on the resource code to be verified and the first verification field; if the check passes, it is determined that the germplasm resource has release permission. By using a dual verification method of comparing resource code differences and verifying the verification field, release is only allowed if the morphological characteristics, compositional characteristics, and blockchain evidence verification information are all consistent, which can effectively prevent germplasm resources from being mixed, replaced, or released without authorization during storage.

[0013] As a technical solution of the present invention, when the hash truncation segment and the resource identifier code are interleaved in a byte-alternating manner, the first byte of the hash truncation segment is used as the first byte of the interleaved sequence. This method fixes the starting position of the interleaved sequence, facilitating the generation of the verification field and ensuring consistency during subsequent review.

[0014] As a technical solution of the present invention, when generating the code of the resource to be verified, if the number of frames in the current morphological image sequence is insufficient, the number of frames is supplemented to be the same as the number of frames in the initial morphological image sequence through interpolation. This can avoid inconsistencies in feature extraction caused by insufficient frames collected during outbound processing, and ensure the stability of the outbound verification process.

[0015] As a technical solution of the present invention, the bitwise XOR operation adopts a byte-by-byte XOR method, and before the XOR, the resource code to be verified and the resource identifier code are padded to the same length of bytes. By unifying the byte length and performing byte-by-byte XOR comparison, the accuracy and consistency of the encoding comparison process can be improved.

[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0017] This method generates resource identification codes by cross-mapping external contour features and internal component features of rice germplasm resources. External contour features are derived from an initial morphological image sequence, while internal component features are derived from physiological parameters such as protein content, starch content, and moisture content. The two pixels with the largest Euclidean distance in the contour pixel sequence are selected as the first and second anchor points. A baseline segment is defined using these two anchor points. Internal component features are projected onto the baseline segment along a direction perpendicular to it. The length of the baseline segment and the projected components are then weighted and fused to quantize the resource identification code. This code generation depends on both external morphology and internal quality factors. Consistent coding is difficult to achieve when morphology or component is unilaterally imitated or replaced. Germplasm with similar appearance but different components can form different codes, as can germplasm with similar components but different morphologies. The baseline segment is selected based on the maximum distance between contour pixels, reducing the impact of image rotation and translation on code generation. Projecting internal component features onto the morphological baseline segment achieves spatial binding between morphology and components, reducing information fragmentation caused by independent coding of multi-dimensional physiological parameters. After the resource identifier code is written into the current evidence storage block of the blockchain, the block hash value is extracted from the block header of the current evidence storage block. A hash segment is obtained by truncating the hash segment and the resource identifier code in an alternating byte arrangement to form an interleaved sequence. Cyclic redundancy check (CRUD) is then performed on the interleaved sequence to obtain the first verification field. The resource identifier code and the first verification field are then associated and stored in the transaction record of the current evidence storage block. During the release process, a resource identifier code to be verified is generated by cross-mapping based on the same characteristics. The already-certified resource identifier code and the first verification field are read from the transaction record. First, a bitwise XOR operation is performed between the resource identifier code to be verified and the already-certified resource identifier code. If the difference is zero, a CRUD check is performed again. If the check passes, the germplasm resource is determined to have release authority. The verification field is generated by the block hash and the resource identifier code, which binds the resource identifier code in the transaction record to the block it belongs to. Replacing the resource identifier code alone or modifying the block hash alone will fail the subsequent review. The outbound verification does not rely on a single point record in a centralized database, but is completed by comparing and verifying the code to be verified with the blockchain evidence field. The outbound permission verification process is tamper-proof and traceable, and can detect cases of identifier copying, replacement and inconsistency in transaction records. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a flowchart of a blockchain-based intelligent management method for rice germplasm resources.

[0020] Figure 2 This is a flowchart of the method for generating identification codes for rice germplasm resources;

[0021] Figure 3 This is a flowchart of the blockchain-based rice germplasm resource release authorization verification process. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] See Figure 1 This invention provides a blockchain-based intelligent management method for rice germplasm resources. In response to a germplasm resource entry request, the initial morphological image sequence and initial physiological parameter set of the rice germplasm resource are acquired. External contour features of the germplasm resource are extracted from the initial morphological image sequence, and internal component features are extracted from the initial physiological parameter set. The external contour features and internal component features are cross-mapped to generate a resource identifier code for the germplasm resource. The resource identifier code is written into the current evidence block of the blockchain, and a first verification field is calculated based on the block hash value of the current evidence block and the resource identifier code. When the germplasm resource is released from the blockchain, the current morphological image sequence and current physiological parameter set of the germplasm resource are collected, and a resource code to be verified is generated using the same feature cross-mapping method. The release permission of the germplasm resource is verified based on the resource code to be verified and the first verification field.

[0024] In practice, the initial morphological image sequence is binarized frame by frame. Each color image in the initial morphological image sequence is converted into a grayscale image, and the grayscale value at each coordinate position in the grayscale image is denoted as... ,in This represents the x-coordinate number of the column in a grayscale image. This represents the y-coordinate index of the row containing the pixel in the grayscale image. The grayscale image is binarized using a binarization segmentation threshold, and the binarization process is performed according to the following formula:

[0025]

[0026] in, Represents the binary image in image coordinates The pixel value at that location, when the value is 1, represents the image coordinates. The pixel at that location belongs to the germplasm region; a value of 0 indicates an image coordinate. The pixel at that location does not belong to the germplasm region; Represents the grayscale image in image coordinates The grayscale value at that location; This represents the binarization segmentation threshold, ranging from 0 to 255, and is automatically calculated on the grayscale image using the maximum inter-class variance (MOL) method. The MOL method works as follows: iterates through each grayscale level from 0 to 255, classifying pixels with grayscale values ​​less than the current grayscale level as the background class, and classifying pixels with grayscale values ​​greater than or equal to the current grayscale level as the germplasm class. The MOL variance between the background and germplasm classes is calculated, and the grayscale level with the largest MOL variance is selected as the binarization segmentation threshold. .

[0027] After frame-by-frame binarization, a binary image corresponding to each frame in the initial morphological image sequence is obtained. Boundary pixel detection is performed on the germplasm regions with a pixel value of 1 in each frame of the binary image, resulting in a set of germplasm region boundary pixels for each frame. The boundary pixel detection is performed as follows: traversing each pixel position with a pixel value of 1 in the binary image, if at least one pixel position with a pixel value of 0 exists in the four neighboring regions of the current pixel position, then the current pixel position is marked as a germplasm region boundary pixel; if the pixel values ​​of all four pixel positions in the four neighboring regions of the current pixel position are 1, then the current pixel position is not marked as a germplasm region boundary pixel. The four neighboring regions refer to the left, right, upper, and lower pixel positions adjacent to the current pixel position. All pixel positions marked as germplasm region boundary pixels in each frame of the binary image constitute the set of germplasm region boundary pixels for the corresponding frame image.

[0028] The geometric center is calculated for the set of boundary pixels of the germplasm region in each frame of the image. The geometric center is calculated as follows: obtain the x-coordinate and y-coordinate values ​​of all boundary pixels in the set of boundary pixels of the germplasm region, calculate the average of the x-coordinate values ​​of all boundary pixels as the x-coordinate of the geometric center, calculate the average of the y-coordinate values ​​of all boundary pixels as the y-coordinate of the geometric center, and combine the x-coordinate and y-coordinate of the geometric center to form the geometric center coordinates.

[0029] The geometric centers of each frame image are aligned to a unified coordinate system. The origin of the unified coordinate system is set as the average of the coordinates of the geometric centers of all frames. Specifically, the average of the x-coordinates of the geometric centers of all frames is used as the x-coordinate of the origin of the unified coordinate system, and the average of the y-coordinates of the geometric centers of all frames is used as the y-coordinate of the origin of the unified coordinate system. Each boundary pixel coordinate in the boundary pixel set of the germplasm region of each frame image is translated. The translation method is as follows: subtract the x-coordinate of the corresponding frame image's geometric center from the x-coordinate of the boundary pixel coordinate, and add the x-coordinate of the origin of the unified coordinate system to obtain the aligned boundary pixel x-coordinate; subtract the y-coordinate of the corresponding frame image's geometric center from the y-coordinate of the boundary pixel coordinate, and add the y-coordinate of the origin of the unified coordinate system to obtain the aligned boundary pixel y-coordinate. The aligned boundary pixel x-coordinate and aligned boundary pixel y-coordinate constitute the aligned boundary pixel coordinates, and all aligned boundary pixel coordinates of the corresponding frame image constitute the aligned germplasm region boundary pixel set.

[0030] In a unified coordinate system, the superimposed frequency distribution of the boundary pixel set of the germplasm region in each frame of the aligned image is statistically analyzed. Specifically, a frequency matrix is ​​established, with its rows and columns covering the possible coordinate ranges of all aligned boundary pixel coordinates in the unified coordinate system. Each element in the frequency matrix is ​​initially set to 0. The aligned boundary pixel set of the germplasm region in each frame is traversed. For each aligned boundary pixel coordinate read, the value of the corresponding element in the frequency matrix is ​​incremented by 1. After the traversal, each element in the frequency matrix represents the cumulative frequency of the corresponding coordinate position appearing as a boundary pixel of the germplasm region across all frames of the image, forming a superimposed frequency distribution.

[0031] External contour features are extracted based on the overlay frequency distribution. A preset frequency threshold is set to 60% of the total number of frames involved in the overlay, and the calculated result is rounded up. The total number of frames involved in the overlay is the total number of frames in the initial morphological image sequence. If the element value corresponding to a coordinate position in the frequency matrix is ​​higher than the preset frequency threshold, that coordinate position is added to the external contour features; if the element value corresponding to a coordinate position in the frequency matrix is ​​lower than or equal to the preset frequency threshold, that coordinate position is not added to the external contour features. The external contour features consist of all added coordinate positions, representing a set of boundary pixel coordinates.

[0032] In specific implementation, please refer to Figure 2The protein, starch, and moisture contents of rice germplasm resources were obtained from the initial physiological parameter set. Protein content represents the percentage of protein mass in the total mass of the rice germplasm resources; starch content represents the percentage of starch mass; and moisture content represents the percentage of moisture mass. These values ​​were obtained through physicochemical analysis of the rice germplasm resource samples using the storage detection equipment and were stored as part of the initial physiological parameter set.

[0033] Protein content is mapped to a preset protein quantification range to obtain the protein quantification value. The preset protein quantification range is set to an integer range of 0 to 255, and the original value range of protein content is 0% to 100%. The protein quantification value is calculated as follows: divide the protein content by 100%, multiply by 255, and round the result to the nearest integer. If the protein content is 0%, the protein quantification value is 0; if the protein content is 100%, the protein quantification value is 255. The protein quantification range is set to 0 to 255 because the protein quantification value is represented using 8-bit binary encoding. 8-bit binary encoding can represent the integer range of 0 to 255, and the protein quantification range matches the width of the 8-bit binary encoding.

[0034] The starch content is mapped to a preset starch quantification range to obtain the starch quantification value. The preset starch quantification range is set to an integer range of 0 to 255, and the original value range of starch content is 0% to 100%. The starch quantification value is calculated as follows: divide the starch content by 100%, multiply by 255, and round the result to the nearest integer. If the starch content is 0%, the starch quantification value is 0; if the starch content is 100%, the starch quantification value is 255. The starch quantification range is set to 0 to 255 because the starch quantification value is represented using 8-bit binary encoding. 8-bit binary encoding can represent the integer range of 0 to 255, and the starch quantification range matches the width of the 8-bit binary encoding.

[0035] The moisture content is mapped to a preset moisture quantization range to obtain a moisture quantization value. The preset moisture quantization range is set to an integer range of 0 to 255, with the original moisture content ranging from 0% to 100%. The moisture quantization value is calculated by dividing the moisture content by 100%, multiplying by 255, and rounding the result to the nearest integer. If the moisture content is 0%, the moisture quantization value is 0; if the moisture content is 100%, the moisture quantization value is 255. The moisture quantization range is set to 0 to 255 because the moisture quantization value is represented using 8-bit binary encoding. 8-bit binary encoding can represent the integer range of 0 to 255, and the moisture quantization range matches the width of the 8-bit binary encoding.

[0036] The protein quantification value, starch quantification value, and moisture quantification value are concatenated into an internal component feature vector according to a preset order. The preset order is as follows: protein quantification value as the first component of the internal component feature vector, starch quantification value as the second component, and moisture quantification value as the third component. The internal component feature vector is a three-dimensional vector, with the first component corresponding to the protein quantification value, the second component corresponding to the starch quantification value, and the third component corresponding to the moisture quantification value.

[0037] Each component in the internal component feature vector is converted into binary code according to a preset bit width of 8 bits. Protein quantification values ​​are converted to 8-bit binary codes, starch quantification values ​​are converted to 8-bit binary codes, and moisture quantification values ​​are converted to 8-bit binary codes. The conversion method involves converting the protein, starch, and moisture quantification values ​​into binary form separately. If the length of the converted binary code is less than 8 bits, 0 to 8 bits are padded to the left of the binary code; if the length of the converted binary code is equal to 8 bits, the binary code remains unchanged. Protein, starch, and moisture quantification values ​​are all within the integer range of 0 to 255, and the length of the converted binary code will not exceed 8 bits.

[0038] The binary codes of each component are sequentially concatenated to form the internal component feature. The sequential concatenation method involves piecing together the 8-bit binary codes of the protein quantification value, starch quantification value, and moisture quantification value in the order of protein quantification value, starch quantification value, and moisture quantification value, resulting in a 24-bit binary code that serves as the internal component feature. The first 8 bits of the internal component feature correspond to the protein quantification value, the middle 8 bits correspond to the starch quantification value, and the last 8 bits correspond to the moisture quantification value.

[0039] The external contour features are sorted according to the x-coordinate values ​​of the contour pixels to obtain a contour pixel sequence. The external contour features are represented as a set of boundary pixel coordinates, each consisting of an x-coordinate and a y-coordinate. The sorting method is to sort all boundary pixel coordinates in the external contour features from smallest to largest x-coordinate; if multiple boundary pixel coordinates have the same x-coordinate value, they are then sorted from smallest to largest y-coordinate value. After sorting, a contour pixel sequence is formed, where each element is a boundary pixel coordinate.

[0040] Calculate the Euclidean distance between every two adjacent pixels in the contour pixel sequence to obtain a distance sequence. Adjacent pixels in the contour pixel sequence refer to the boundary pixel coordinates of two adjacent positions in the sequence. For the _th ... The coordinates of the first boundary pixel and the first The coordinates of the boundary pixel are calculated using the Euclidean distance method: calculate the... The x-coordinate value of the boundary pixel coordinates and the first The square of the difference between the x-coordinate values ​​of the boundary pixel coordinates is used to calculate the th... The ordinate value of the boundary pixel coordinates and the first The square of the difference between the ordinate values ​​of the boundary pixel coordinates is taken by adding the two squared values ​​and then taking the square root to obtain the first squared value. The coordinates of the first boundary pixel and the first The Euclidean distance between the boundary pixel coordinates. Following the order of adjacent pixels in the contour pixel sequence, the Euclidean distances between all adjacent pixels are calculated sequentially to form a distance sequence. The distance sequence contains the first... The element represents the nth element in the contour pixel sequence. The coordinates of the first boundary pixel and the first The Euclidean distance between the boundary pixel coordinates.

[0041] The two pixels corresponding to the maximum value in the distance sequence are designated as the first and second anchor points. The maximum value in the distance sequence represents the Euclidean distance between the two farthest adjacent boundary pixel coordinates in the contour pixel sequence. The two pixels corresponding to the maximum value in the distance sequence are the pair of adjacent boundary pixel coordinates that generate the maximum Euclidean distance in the contour pixel sequence. The first pixel of the two pixels corresponding to the maximum value in the distance sequence is designated as the first anchor point, and the second pixel is designated as the second anchor point. Both the first and second anchor points are boundary pixel coordinates in the external contour feature.

[0042] A baseline segment is defined using the first and second anchor points. The baseline segment is a straight line connecting the first and second anchor points, with the first endpoint being the first anchor point and the second endpoint being the second anchor point. The length of the baseline segment is calculated as follows: the square of the difference between the x-coordinate of the first anchor point and the x-coordinate of the second anchor point is calculated, and the square of the difference between the y-coordinate of the first anchor point and the y-coordinate of the second anchor point is also calculated. The two squared values ​​are then added together and the square root is taken to obtain the length of the baseline segment.

[0043] The internal component features are projected onto the baseline segment along a direction perpendicular to the baseline segment to obtain the projected components. The projection method is as follows: within the plane containing the first and second anchor points, the internal component features are treated as a binary encoded string, and each binary value in the internal component features is sequentially mapped to a projected length unit. The length of the internal component feature is 24 bits, and each binary value is either 0 or 1. When projecting the internal component features onto the baseline segment, the projection direction is perpendicular to the baseline segment, and each binary value of the internal component feature is projected onto the baseline segment along a direction perpendicular to the baseline segment. After projection, the projected position of each binary value on the baseline segment is obtained. The magnitude of the projected component is equal to the sum of the projected length units corresponding to all binary values ​​of 1 in the internal component feature. The projected length unit corresponding to each binary value of 1 is set to one-twenty-fourth of the length of the baseline segment. The projected component is calculated by counting the number of binary values ​​of 1 in the internal component feature and multiplying the number of binary values ​​of 1 by one-twenty-fourth of the length of the baseline segment to obtain the projected component.

[0044] The length value of the baseline segment and the projected components are weighted and fused according to a preset weighting coefficient, and the fusion result is quantized into a resource identifier code. The weighted fusion is performed according to the following formula:

[0045]

[0046] in, Indicates the fusion result. This represents the weighting coefficient for the baseline segment length, with a value ranging from 0 to 1. Set to 0.6, based on the main structural information of the external contour features reflected by the length of the baseline segment, and assign a large weight value of 0.6; Indicates the length of the baseline segment; This represents the weighting coefficient for the projected components, with a value ranging from 0 to 1. The weight is set to 0.4, based on the fact that the projected components reflect the cross-correlation between internal component features and external contour features, and is assigned a weight value of 0.4. and The sum is 1; Indicates the projected component.

[0047] The fusion result is quantized into a resource identifier code. The quantization method involves converting the fusion result into binary code according to a preset encoding bit length (64 bits). The fusion result is a real number; multiplying it by a quantization coefficient and rounding down yields the quantized integer value. The quantization coefficient is set to... The reciprocal of , i.e., the quantization coefficient, is The fusion result is divided by the quantization coefficient to obtain a quantized integer value. This quantized integer value is then converted into a 64-bit binary code to obtain the resource identifier code. The resource identifier code is a 64-bit binary code that can uniquely represent the identification information of rice germplasm resources after cross-mapping of external contour features and internal component features.

[0048] In practice, the block hash value is extracted from the block header of the current evidence-preserving block. The current evidence-preserving block is the most recently written block on the blockchain. The block header of the current evidence-preserving block contains multiple fields, and the block hash value field records the hash value obtained by hashing the block header data of the current evidence-preserving block. The extraction method involves reading the block header data of the current evidence-preserving block, locating the block hash value field from the block header data, and obtaining the corresponding byte sequence as the block hash value. When using the SHA-256 hash algorithm, the block hash value is 32 bytes.

[0049] The block hash value is truncated into a hash truncation segment with a fixed byte length. The fixed byte length is set to 8 bytes. This is because the resource identifier is encoded in 64-bit binary code, and 64 bits equals 8 bytes. When alternating bytes between the hash truncation segment and the resource identifier, they must have the same byte length to complete the byte-by-byte interleaving. The truncation method involves taking 8 consecutive bytes starting from the first byte of the block hash value to obtain the hash truncation segment. The hash truncation segment consists of bytes 1 through 8 of the block hash value.

[0050] The hash truncation and resource identifier code are interleaved in a byte-alternating manner to obtain an interleaved sequence. The hash truncation is 8 bytes long, and the resource identifier code is 8 bytes long. The interleaving combination method is as follows: the first byte of the interleaved sequence uses the first byte of the hash truncation segment; the second byte of the interleaved sequence uses the first byte of the resource identifier encoding; the third byte of the interleaved sequence uses the second byte of the hash truncation segment; the fourth byte of the interleaved sequence uses the second byte of the resource identifier encoding; the fifth byte of the interleaved sequence uses the third byte of the hash truncation segment; the sixth byte of the interleaved sequence uses the third byte of the resource identifier encoding; the seventh byte of the interleaved sequence uses the fourth byte of the hash truncation segment; the eighth byte of the interleaved sequence uses the fourth byte of the resource identifier encoding; the ninth byte of the interleaved sequence uses the fifth byte of the hash truncation segment; the tenth byte of the interleaved sequence uses the fifth byte of the resource identifier encoding; the eleventh byte of the interleaved sequence uses the sixth byte of the hash truncation segment; the twelfth byte of the interleaved sequence uses the sixth byte of the resource identifier encoding; the thirteenth byte of the interleaved sequence uses the seventh byte of the hash truncation segment; the fourteenth byte of the interleaved sequence uses the seventh byte of the resource identifier encoding; the fifteenth byte of the interleaved sequence uses the eighth byte of the hash truncation segment; and the sixteenth byte of the interleaved sequence uses the eighth byte of the resource identifier encoding. The interleaved sequence uses the first byte of the hash truncation as the first byte of the interleaved sequence, and the total length of the interleaved sequence is 16 bytes.

[0051] In some embodiments, the byte order of the hash truncation segment and the byte order of the resource identifier encoding remain unchanged. When interleaving, starting from the first byte of the hash truncation segment, one byte is taken alternately from the hash truncation segment and the resource identifier encoding until both byte sequences are taken, resulting in an interleaved sequence.

[0052] Cyclic redundancy check (CRC) calculations are performed on the interleaved sequence to obtain the first check field. The CRC-32 cyclic redundancy check algorithm is used, and its generator polynomial is: ,in This represents the formal variables in the generator polynomial. The cyclic redundancy check (CRC) calculation process is as follows: The 16 bytes of the interleaved sequence are arranged sequentially into 128 bits of binary data, which serves as the data to be checked; 32 zero bits are appended to the end of the data to be checked, forming extended data; a modulo-2 division operation is performed on the extended data using the CRC-32 generator polynomial, with an XOR operation performed during the modulo-2 division. The 32-bit remainder obtained after the operation is the CRC result, which is used as the first check field. The first check field is 32 bits of binary data, corresponding to 4 bytes.

[0053] The resource identifier code and the first verification field are associated and stored in the transaction record of the current evidence storage block. Specifically, a storage record is created in the transaction record of the current evidence storage block, with the resource identifier code as the resource identifier code field and the first verification field as the first verification field. The resource identifier code field and the first verification field are stored together in an associated manner. After the storage record is written into the transaction record of the current evidence storage block, the transaction record of the current evidence storage block simultaneously contains both the resource identifier code and the first verification field, and the resource identifier code and the first verification field are associated through the same transaction record.

[0054] In practice, in response to a germplasm resource release request, the system acquires the current morphological image sequence and current physiological parameter set of the rice germplasm resource to be released at the current moment. The current morphological image sequence is obtained by image acquisition equipment deployed in the germplasm resource release channel, which continuously captures images of the rice germplasm resource from multiple angles. The frames in the current morphological image sequence are arranged in chronological order of capture time. The current physiological parameter set is obtained by physiological detection equipment deployed in the germplasm resource release channel, which performs physiochemical analysis on the rice germplasm resource sample. The current physiological parameter set includes the protein content, starch content, and moisture content of the rice germplasm resource at the current moment. The total number of frames in the current morphological image sequence is denoted as . Obtain the total number of frames in the initial morphological image sequence, denoted as . .

[0055] In some embodiments, if the total number of frames in the current morphological image sequence Less than the total number of frames in the initial morphological image sequence The number of frames in the current shape image sequence is padded to match the number of frames in the initial shape image sequence using interpolation. If the total number of frames in the current shape image sequence is... Greater than or equal to the total number of frames in the initial morphological image sequence No interpolation or frame replacement is performed; the current morphological image sequence is used directly for subsequent processing. If the total number of frames in the current morphological image sequence... Greater than the total number of frames in the initial morphological image sequence Extract images from the current image sequence at equal intervals according to the shooting time sequence. The frame image yields a current morphological image sequence with the same number of frames as the initial morphological image sequence. If the total number of frames in the current morphological image sequence is... Less than the total number of frames in the initial morphological image sequence Calculate the difference in the number of frames that need to be supplemented. The calculation method for making up the frame count difference is as follows: .

[0056] The interpolation frame replacement method involves generating an inserted frame image between every two adjacent frames in the current morphological image sequence. The original frame images in the current morphological image sequence are arranged in chronological order of capture time, with the number of adjacent frame pairs being [number missing]. The frame difference that needs to be supplemented. The frames are allocated to all adjacent frame pairs, with the allocation method being to fill in the frame number difference. Divide by the number of adjacent frame pairs The average number of inserted frames and the remainder between each adjacent frame pair are obtained. The average number of inserted frames is then distributed to each adjacent frame pair, and the remainder is distributed sequentially to the first few adjacent frame pairs according to their order of arrangement. Each adjacent frame pair can have at most one more inserted frame than the average number of inserted frames. Each inserted frame image is generated through linear interpolation. The pixel value of the inserted frame image is calculated by weighting the pixel values ​​at the same pixel coordinate position in the preceding and following frames of its adjacent frame pair. The linear interpolation is performed using the following formula:

[0057]

[0058] in, This indicates that the pixel column number in the inserted frame image is Pixel row number is Inserted frame number is The pixel value at the pixel position; This indicates that the pixel column number in the previous frame image is Pixel row number is The pixel value at the pixel position; This indicates that the pixel column number in the next frame image is Pixel row number is The pixel value at the pixel position; This represents the total number of segments of the frame interval inserted between adjacent frame pairs, and is taken as the number of frames inserted between adjacent frame pairs plus 1. This indicates the position index of the currently inserted frame within the adjacent frame pair, with a value ranging from 1 to... integers, when When the value is 1, the inserted frame image is closer to the previous frame image. equal The inserted frame image is closer to the next frame image; This indicates the frame number of the inserted frame image in the current morphological image sequence after padding; This indicates the frame number of the previous frame in the current morphological image sequence before completion; This indicates the frame number of the next frame in the current morphological image sequence before the supplementation.

[0059] After the interpolation and frame interpolation process is completed, the current morphological image sequence with the interpolated frame count is obtained. The total number of frames in the current morphological image sequence with the interpolated frame count is equal to the total number of frames in the initial morphological image sequence. same.

[0060] The current contour features are extracted based on the supplemented current morphological image sequence. The specific execution method for extracting the current contour features is as follows: The supplemented current morphological image sequence is binarized frame by frame to obtain the set of pixels representing the current germplasm region boundary in each frame; the geometric center of the set of pixels representing the current germplasm region boundary in each frame is calculated, and the geometric centers of each frame are aligned to a unified coordinate system; in the unified coordinate system, the superposition frequency distribution of the set of pixels representing the current germplasm region boundary in each aligned frame is statistically analyzed; based on the superposition frequency distribution, the set of boundary pixel coordinates with frequencies higher than a preset frequency threshold is extracted as the current contour features. The above processing procedure is the same as the process for extracting external contour features from the initial morphological image sequence, and the preset frequency threshold is the same as that used when extracting external contour features from the initial morphological image sequence.

[0061] Extracting current component features based on the current physiological parameter set. The specific execution method for extracting current component features is as follows: Obtain the protein, starch, and moisture content of the rice germplasm resources to be released from the current physiological parameter set at the current moment; map the current protein, starch, and moisture content to preset protein, starch, and moisture quantification intervals, respectively, to obtain the current protein, starch, and moisture quantification values; concatenate the current protein, starch, and moisture quantification values ​​in a preset order to form the current internal component feature vector; convert each component in the current internal component feature vector into binary code according to a preset bit width, and concatenate the binary codes of each component in sequence to form the current component feature. The above processing is the same as the process for extracting internal component features from the initial physiological parameter set.

[0062] The current contour features and current component features are mapped together using the same feature cross-mapping method to generate the resource code to be verified. The specific execution method for generating the resource code to be verified is as follows: The current contour features are sorted according to the horizontal coordinate values ​​of the contour pixels to obtain the current contour pixel sequence; the Euclidean distance between every two adjacent pixels in the current contour pixel sequence is calculated to obtain the current distance sequence; the two pixels corresponding to the maximum value in the current distance sequence are determined as the current first anchor point and the current second anchor point; the current baseline segment is determined using the current first anchor point and the current second anchor point; the current component features are projected onto the current baseline segment along a direction perpendicular to the current baseline segment to obtain the current projection component; the length value of the current baseline segment and the current projection component are weighted and fused according to a preset weighting coefficient, and the fusion result is quantized as the resource code to be verified. The above processing is the same as the process for generating the resource identifier code during initial database entry, and the preset weighting coefficients are the same as those used during initial database entry.

[0063] In specific implementation, please refer to Figure 3 The process involves retrieving the resource identifier code and the first verification field corresponding to the rice germplasm resource to be released from the transaction records of the current evidence storage block. The current evidence storage block is the block where the resource identifier code and the first verification field were written when the rice germplasm resource was put into storage. The reading method is as follows: Locate the current evidence storage block on the blockchain based on the germplasm number or resource identifier code of the rice germplasm resource; search for the storage record containing the resource identifier code and the first verification field from the transaction records of the current evidence storage block; read the byte sequence corresponding to the resource identifier code field from the storage record as the resource identifier code; and read the byte sequence corresponding to the first verification field field as the first verification field.

[0064] The resource code to be verified is XORed with the read resource identifier code using a bitwise XOR operation to obtain the difference value. The bitwise XOR operation is performed byte-by-byte, and the resource code to be verified and the resource identifier code are padded to the same length in bytes before the XOR operation. Specifically, the length in bytes of the resource code to be verified is obtained, denoted as... Get the length of the resource identifier encoding in bytes, denoted as .like Less than Then, zero bytes are padded to the left of the high-order bits of the resource code to be verified, and the number of zero bytes padded is [number missing]. After supplementation, the byte length of the resource code to be verified is the same as the byte length of the resource identifier code; if Greater than Then, zero bytes are padded to the left of the high-order bits of the resource identifier encoding, and the number of zero bytes padded is [number missing]. After supplementation, the byte length of the resource identifier encoding is the same as the byte length of the resource encoding to be verified; if equal If not, no supplementary operation is performed. The padded resource code to be verified and the padded resource identifier code are XORed byte-by-byte in byte order to obtain the difference value. The byte-by-byte XOR operation is as follows: The first byte of the padded resource code to be verified is XORed with the first byte of the padded resource identifier code to obtain the first byte of the difference value; the second byte of the padded resource code to be verified is XORed with the second byte of the padded resource identifier code to obtain the second byte of the difference value; this process is repeated until all bytes have undergone XOR operation, and all the difference bytes are then combined in order to form the difference value. The difference value is obtained using the following formula:

[0065]

[0066] in, The difference value is the byte sequence obtained by XORing the padded resource code to be verified with the padded resource identifier code byte by byte. This indicates the padded resource code to be verified, which is the resource code to be verified or the byte sequence of the resource code to be verified after padding the high-order bits with zeros on the left. This indicates the padded resource identifier code, which is either the resource identifier code or the byte sequence after padding the high-order bits with zeros on the left. This represents the bitwise XOR operator, for... and The XOR operation is performed on the binary bits at the same position. The result is 0 when the two binary bits are the same and 1 when the two binary bits are different.

[0067] In some embodiments, after obtaining the difference value, it is determined whether the difference value is zero. The determination method is as follows: check each binary bit of each byte in the difference value byte by byte. If all binary bits in the difference value are 0, the difference value is determined to be zero; if at least one binary bit in the difference value is 1, the difference value is determined to be non-zero.

[0068] If the difference is zero, the resource code to be verified is compared with the first check field using a cyclic redundancy check (CRC) algorithm. The CRC verification uses the same CRC-32 cyclic redundancy check algorithm as the one used to generate the first check field, and the generator polynomial is... ,in This represents the formal variable in the generator polynomial. The execution method for cyclic redundancy check (CRC) verification is as follows: The byte sequence of the resource code to be verified is concatenated with the byte sequence of the first check field, with the byte sequence of the resource code to be verified first, followed by the byte sequence of the first check field, forming a check data segment. This check data segment is used as the data to be verified, and 32 zero bits are appended to the end of the data to be verified, forming extended check data. A modulo-2 division operation is performed on the extended check data using the CRC-32 generator polynomial. An XOR operation is performed during the modulo-2 division, and the 32-bit remainder is obtained after the operation, which is recorded as the verification remainder. If all 32 bits of the verification remainder are 0, the CRC verification is considered successful; if at least one bit in the 32 bits of the verification remainder is 1, the CRC verification is considered unsuccessful.

[0069] If the difference value is zero and the cyclic redundancy check passes, the rice germplasm resource to be released is determined to have release authority. If the difference value is not zero, the cyclic redundancy check is not performed, and the rice germplasm resource to be released is not determined to have release authority. If the difference value is zero but the cyclic redundancy check fails, the rice germplasm resource to be released is not determined to have release authority.

[0070] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A blockchain-based intelligent management method for rice germplasm resources, characterized in that: include: In response to the request to add germplasm resources to the database, the initial morphological image sequence and initial physiological parameter set of rice germplasm resources are obtained; The external contour features of the germplasm resources are extracted based on the initial morphological image sequence, and the internal component features of the germplasm resources are extracted based on the initial physiological parameter set. The external contour features and the internal component features are cross-mapped to generate the resource identification code for the germplasm resource; Write the resource identifier code into the current evidence storage block of the blockchain, and calculate the first verification field based on the block hash value of the current evidence storage block and the resource identifier code; When germplasm resources are released from storage, the current morphological image sequence and current physiological parameter set of the germplasm resources are collected, and the code of the resource to be verified is generated according to the same feature cross-mapping method. The release permission of the germplasm resource is verified based on the resource code to be verified and the first verification field.

2. The intelligent management method for rice germplasm resources based on blockchain according to claim 1, characterized in that, Extracting the external contour features of germplasm resources based on the initial morphological image sequence includes: The initial morphological image sequence is binarized frame by frame to obtain the set of germplasm region boundary pixels in each frame image; Calculate the geometric center of the set of boundary pixels of the germplasm region in each frame of the image, and align the geometric centers of each frame of the image to a unified coordinate system; In the unified coordinate system, the superposition frequency distribution of the boundary pixel set of the germplasm region in each frame image after alignment is statistically analyzed; Based on the superimposed frequency distribution, the set of boundary pixel coordinates with frequencies higher than a preset frequency threshold is extracted as the external contour feature.

3. The intelligent management method for rice germplasm resources based on blockchain according to claim 2, characterized in that, The step of extracting internal component characteristics of germplasm resources based on the initial physiological parameter set includes: The protein content, starch content, and moisture content of the germplasm resource were obtained from the initial physiological parameter set. The protein content, starch content, and moisture content are mapped to preset protein quantification intervals, starch quantification intervals, and moisture quantification intervals, respectively, to obtain protein quantification values, starch quantification values, and moisture quantification values. The protein quantification value, starch quantification value, and moisture quantification value are concatenated into an internal component feature vector according to a preset arrangement order; Each component in the internal component feature vector is converted into binary code according to a preset bit width, and the binary codes of each component are concatenated in sequence to form the internal component feature.

4. The intelligent management method for rice germplasm resources based on blockchain according to claim 3, characterized in that, The step of performing feature cross-mapping between the external contour features and the internal component features to generate the resource identifier code for the germplasm resource includes: The external contour features are sorted according to the horizontal coordinate values ​​of the contour pixels to obtain the contour pixel sequence; Calculate the Euclidean distance between every two adjacent pixels in the contour pixel sequence to obtain a distance sequence; Determine the two pixels corresponding to the maximum value in the distance sequence as the first anchor point and the second anchor point; A baseline segment is determined using the first anchor point and the second anchor point. The internal component features are then projected onto the baseline segment along a direction perpendicular to the baseline segment to obtain the projected components. The length value of the baseline segment and the projection component are weighted and fused according to a preset weighting coefficient, and the fusion result is quantized into the resource identifier code.

5. The intelligent management method for rice germplasm resources based on blockchain according to claim 4, characterized in that, The first verification field is calculated based on the block hash value of the current evidence storage block and the resource identifier code, including: Extract the block hash value from the block header of the current evidence-stored block; The block hash value is truncated to a fixed byte length to obtain a hash truncation segment; The hash segment and the resource identifier code are interleaved and combined in an alternating byte order to obtain an interleaved sequence; Cyclic redundancy check (CRC) calculation is performed on the interleaved sequence to obtain the first CRC field; The resource identifier code and the first verification field are associated and stored in the transaction record of the current evidence storage block.

6. The intelligent management method for rice germplasm resources based on blockchain according to claim 5, characterized in that, When germplasm resources are released from storage, the current morphological image sequence and current physiological parameter set of the germplasm resource are collected, and a code for the resource to be verified is generated according to the same feature cross-mapping method, including: In response to a germplasm resource retrieval request, obtain the current morphological image sequence and current physiological parameter set of the germplasm resource at the current moment; Extract the current contour features based on the current morphological image sequence, and extract the current component features based on the current physiological parameter set; The current contour features and the current component features are mapped together using the same feature cross-mapping method to generate the resource code to be verified.

7. The intelligent management method for rice germplasm resources based on blockchain according to claim 6, characterized in that, The step of verifying the release permission of the germplasm resource based on the resource code to be verified and the first verification field includes: Read the resource identifier code and the first verification field corresponding to the germplasm resource from the transaction records of the current evidence storage block; Perform a bitwise XOR operation between the resource code to be verified and the read resource identifier code to obtain the difference value; If the difference value is zero, then the resource code to be verified is subjected to cyclic redundancy check verification with the first verification field; If the review is successful, the germplasm resource is deemed to have the authority to be released from the warehouse.

8. The intelligent management method for rice germplasm resources based on blockchain according to claim 5, characterized in that, When the hash truncation segment and the resource identifier code are interleaved in a byte-alternating manner, the first byte of the hash truncation segment is used as the first byte of the interleaving sequence.

9. The intelligent management method for rice germplasm resources based on blockchain according to claim 6, characterized in that, When generating the resource code to be verified, if the number of frames in the current morphological image sequence is insufficient, the number of frames is supplemented to be the same as the number of frames in the initial morphological image sequence by interpolation.

10. The intelligent management method for rice germplasm resources based on blockchain according to claim 7, characterized in that, The bitwise XOR operation adopts a byte-by-byte XOR method, and before the XOR, the resource code to be verified and the resource identifier code are padded to the same length of bytes.

Images