CN-122022829-A - Quartz sand quality traceability management system based on block chain

CN122022829ACN 122022829 ACN122022829 ACN 122022829ACN-122022829-A

Abstract

The invention relates to the technical field of blockchains and discloses a blockchain-based quartz sand quality tracing management system which comprises a robust power law regression module, a dehumidification brightness field calculation module, a quartz sand quality prediction module, an input promise construction module, a blockchain verification module and a blockchain tracing module, wherein the robust power law regression module is used for regression to obtain a wet-induced heteroscedure power law slope, the dehumidification brightness field calculation module is used for calculating to obtain a dehumidification brightness field, the quartz sand quality prediction module is used for obtaining predicted values of iron content and whiteness of quartz sand through a prediction model, the input promise construction module is used for constructing an input promise and writing the input promise into a blockchain, the blockchain verification module is used for proving through an intelligent contract verification range, and the blockchain tracing module is used for carrying out minimum disclosure inquiry on the blockchain. According to the method, the influence of humidity on imaging is described through the power law slope of the wet-induced heteroscedastic, the power inverse mapping and the layer-by-layer characteristic modulation are combined, so that the wet-induced heteroscedastic interference is effectively counteracted, the accuracy of quartz sand iron content and whiteness prediction is improved, meanwhile, the non-falsifiable characteristic of a block chain is relied on, and the quality data quantization, the accurate storage and the safe tracing of the quartz sand are realized in the whole process.

Inventors

LI YAJIE
ZHANG YAN

Assignees

北京亚泽石英材料有限公司

Dates

Publication Date: 20260512
Application Date: 20251205

Claims (10)

1. Quartz sand quality traceability management system based on block chain, which is characterized by comprising: the robust power law regression module is used for obtaining the background removing brightness of each image in the quartz sand image sequence through morphological opening operation, calculating band-pass texture energy and block-level brightness average value according to the background removing brightness and sand grain size, and obtaining a wet-induced heteroscedastic power law slope through robust power law regression; the dehumidification brightness field calculation module is used for carrying out power inverse mapping on the standardized background brightness according to the power law slope of the wet-induced heteroscedastic power law to obtain a dehumidification brightness field; The quartz sand quality prediction module is used for inputting a dehumidification brightness field into a prediction model, taking a wet-induced heteroscedastic power law slope as a layer-by-layer characteristic linear modulation quantity, and outputting predicted values of iron content and whiteness of quartz sand; the input promise construction module is used for carrying out hash operation on the image sequence, the wet heteroscedastic power law slope and the prediction model parameters to form an input promise, and writing the input promise into a blockchain in combination with the predicted values of iron content and whiteness of quartz sand; The block chain verification module is used for establishing promise containing random blinding factors for iron content, generating range evidence and verifying the range evidence through a block chain intelligent contract, and when verification is passed, converting the batch quality state from an initial state to a qualified state, and otherwise, converting the batch quality state to a reworking state; And the blockchain tracing module is used for generating vector promises for predicted values of the wet-induced heteroscedastic power law slope, the iron content and the whiteness of quartz sand at each station and each moment, registering the vector promises on a blockchain, providing point-by-point unsealing evidence during audit and carrying out minimum disclosure inquiry on the blockchain.
2. The quartz sand quality traceability management system based on the blockchain as defined in claim 1, wherein the linear brightness of each image in the quartz sand image sequence is firstly subjected to morphological open operation through a preset structural element radius to obtain an image background; Determining a first filtering scale, a second filtering scale and a block side length according to the pixelation amount of the sand scale, wherein the block side length is equal to four times of the pixelation amount of the sand scale, the first filtering scale is equal to the pixelation amount of the sand scale, and the second filtering scale is equal to twice of the pixelation amount of the sand scale; dividing the background removing brightness of each image into non-overlapping square blocks according to the side length of each block, wherein the number of pixels of each square block is fixed; Calculating the difference value of two filters corresponding to the two filtering scales, convoluting the difference value with the background brightness, squaring the convolution result, and then taking the average value in the square to obtain the band-pass texture energy of each square; And calculating the average value of the background brightness in the blocks, and obtaining the block-level brightness average value of each block.
3. The blockchain-based quartz sand quality traceability management system of claim 1, wherein a numerical stability constant is set, and a processed block-level luminance average is defined as a maximum between the block-level luminance average and the numerical stability constant; obtaining regression intercept and wet-induced variance power law slope by a robust power law regression method taking minimized absolute residual as a cost function, wherein the cost function is calculated by calculating the product between the wet-induced variance power law slope and the logarithmic value of the processed block-level brightness mean value, and then calculating the sum of the logarithmic value of minimized band-pass texture energy and the absolute value of the difference between the regression intercept and the product.
4. The blockchain-based quartz sand quality traceability management system according to claim 1, wherein a minimum value and a maximum value in the background removal brightness are obtained, the minimum value is subtracted from the background removal brightness of each pixel position, the sum of a numerical stability constant and a difference between the maximum value and the minimum value is divided to obtain a standardized background removal brightness, the standardized background removal brightness is used as a base, the inverse of a power law slope of a wet variance is used as an index, and a power operation is performed to obtain a dehumidification brightness field.
5. The quartz sand quality traceability management system based on the blockchain is characterized in that a prediction model is formed by sequentially connecting a preset number of convolution layers, each convolution layer performs convolution operation through a convolution kernel preset by the layer and a feature map output by the previous layer, a middle feature map is obtained by inputting a nonlinear activation function after a convolution result is overlapped with a preset bias item of the layer, a first modulation coefficient and a second modulation coefficient are built according to a power law slope of a wet-induced differential, the first modulation coefficient is built by a first trainable scalar and a second trainable scalar corresponding to the layer in a linear combination mode, the second modulation coefficient is built by a third trainable scalar and a fourth trainable scalar corresponding to the layer in a linear combination mode, and the feature map output by the current layer is obtained by overlapping the second modulation coefficient after multiplying the middle feature map with the first modulation coefficient and is used as the feature map input by the next convolution layer.
6. The quartz sand quality traceability management system based on the blockchain according to claim 5, wherein global average pooling is performed on the feature map output by the last convolution layer, and the average value of feature maps corresponding to all pixel positions of the image is calculated to obtain an aggregate feature vector; And multiplying the transpose of the second weight parameter with the aggregation feature vector, and then superposing a second offset parameter to obtain the predicted value of the whiteness of the quartz sand.
7. The quartz sand quality traceability management system based on the blockchain as in claim 1, wherein each image in the quartz sand image sequence is converted into binary data, the binary data corresponding to all the images are spliced in sequence according to the image acquisition sequence to obtain a first byte string, and hash operation is performed on the first byte string through the SHA256 algorithm to obtain an image sequence input promise; Converting the wet-induced heteroscedastic power law slope into binary data according to a fixed length rule, and carrying out hash operation on the binary data through an SHA256 algorithm to obtain a wet-induced heteroscedastic power law slope input promise; Converting each parameter in the prediction model into binary data according to a fixed length rule, sequentially splicing binary data of all the parameters according to a preset sequence to obtain a second byte string, and carrying out hash operation on the second byte string through an SHA256 algorithm to obtain a prediction model parameter input promise; Converting predicted values of iron content and whiteness of quartz sand into binary data according to fixed length rules, inputting an image sequence into a promise, inputting a wet variance power law slope into a promise, inputting a predicted model parameter into a promise according to a preset sequence, splicing the encoded iron content predicted values and the encoded whiteness predicted values in sequence to obtain a blockchain transaction load, executing a blockchain writing operation, submitting the transaction load to a blockchain, and generating a corresponding transaction identifier.
8. The quartz sand quality traceability management system based on the blockchain according to claim 1, wherein two public group generating elements are preset, a random blinding factor is uniformly selected from an integer set of large prime order groups, based on the two group generating elements, the iron content predicted value is used as an exponent of a first group generating element to execute power operation, the random blinding factor is used as an exponent of a second group generating element to execute power operation, and the results of the power operation are multiplied to obtain an iron content promise; Setting a service threshold value, and generating a range evidence for proving that the iron content predicted value is between 0 and the service threshold value through a section evidence construction algorithm based on a secret summation protocol; submitting the iron content promise, the range proof and the business threshold value to a blockchain intelligent contract, and the intelligent contract runs the range proof verification algorithm and outputs a Boolean type verification result; if the verification result is passed, the batch quality state is updated from the initial state to the qualified state; if the verification result is not passed, the batch quality state is updated from the initial state to the reworking state, and the updated batch quality state, the iron content promise and the range demonstration are output.
9. The quartz sand quality traceability management system based on the blockchain according to claim 1, wherein for each station and each moment, a wet-induced heteroscedastic power law slope, a quartz sand iron content predicted value and a quartz sand whiteness predicted value corresponding to the moment of the station are obtained, and three data are combined according to a fixed sequence to form a data triplet of the moment of the station; And setting a positive integer constant of the time capacity corresponding to each station, and obtaining a position mapping value corresponding to each data triplet in a calculation mode of multiplying the constant by the station number and adding the time value.
10. The blockchain-based quartz sand quality traceability management system of claim 9, wherein a unique polynomial is defined by lagrangian interpolation such that the function value of the polynomial at each location mapping value is equal to the finite field element obtained by mapping the corresponding data triplet; performing a blockchain write operation to submit the vector commitment to the blockchain to generate a corresponding serial transaction identifier; The method comprises the steps of providing a target station and a target moment to be audited, obtaining a target position mapping value according to a position mapping calculation mode, generating a data combination comprising corresponding finite field elements and point-wise unsealing evidence according to a constructed polynomial and the target position mapping value through a point-wise unsealing algorithm, submitting vector promises, the target position mapping value, the finite field elements and the point-wise unsealing evidence to a block chain, executing a verification algorithm by the block chain and outputting a verification result, and restoring to obtain a wet heterogenous power law slope, a quartz sand iron content predicted value and a quartz sand whiteness predicted value of the target station corresponding to the target moment if the verification result is passing through a decoding function which is in dual with digital packing.

Description

Quartz sand quality traceability management system based on block chain Technical Field The invention relates to the technical field of blockchain, in particular to a quartz sand quality traceability management system based on a blockchain. Background The quality detection of the existing quartz sand mostly depends on manual visual or common image recognition technology, and trace moisture in the environment can be adsorbed in the quartz sand production process, so that the phenomenon of wet heterovaria occurs in the imaging process, namely, excessive sharpening of bright texture and suppressed dark texture are caused, the influence cannot be counteracted by common image recognition, and the deviation of the predicted value of the iron content and the whiteness is larger, so that the accuracy requirement of the high-end industry on quality detection is difficult to meet. The existing tracing mode mainly adopts a centralized database to store quality data, the data is easy to falsify artificially and lacks an effective verification mechanism, a part of schemes introduce blockchain uplink verification, but most of schemes directly disclose the full quantity of data to cause commercial privacy leakage such as production process details, if the quality data at a specific station is required to be queried at a specific moment, the full quantity of on-chain data is required to be traversed, the query efficiency is low, accurate tracing cannot be realized, and therefore, the problems are caused that a producer is difficult to quickly locate quality abnormal links, an auditor cannot efficiently verify data authenticity and seriously influence quartz sand production quality control and market trust. Disclosure of Invention The invention provides a quartz sand quality traceability management system based on a blockchain, which solves the technical problems in the background technology. The invention provides a quartz sand quality traceability management system based on a blockchain, which comprises the following components: the robust power law regression module is used for obtaining the background removing brightness of each image in the quartz sand image sequence through morphological opening operation, calculating band-pass texture energy and block-level brightness average value according to the background removing brightness and sand grain size, and obtaining a wet-induced heteroscedastic power law slope through robust power law regression; the dehumidification brightness field calculation module is used for carrying out power inverse mapping on the standardized background brightness according to the power law slope of the wet-induced heteroscedastic power law to obtain a dehumidification brightness field; The quartz sand quality prediction module is used for inputting a dehumidification brightness field into a prediction model, taking a wet-induced heteroscedastic power law slope as a layer-by-layer characteristic linear modulation quantity, and outputting predicted values of iron content and whiteness of quartz sand; the input promise construction module is used for carrying out hash operation on the image sequence, the wet heteroscedastic power law slope and the prediction model parameters to form an input promise, and writing the input promise into a blockchain in combination with the predicted values of iron content and whiteness of quartz sand; The block chain verification module is used for establishing promise containing random blinding factors for iron content, generating range evidence and verifying the range evidence through a block chain intelligent contract, and when verification is passed, converting the batch quality state from an initial state to a qualified state, and otherwise, converting the batch quality state to a reworking state; And the blockchain tracing module is used for generating vector promises for predicted values of the wet-induced heteroscedastic power law slope, the iron content and the whiteness of quartz sand at each station and each moment, registering the vector promises on a blockchain, providing point-by-point unsealing evidence during audit and carrying out minimum disclosure inquiry on the blockchain. Further, firstly carrying out morphological opening operation on the linear brightness of each image in the quartz sand image sequence through the radius of a preset structural element to obtain an image background, and then taking the maximum value between the result of subtracting the image background from the linear brightness and 0 to obtain background-removing brightness; Determining a first filtering scale, a second filtering scale and a block side length according to the pixelation amount of the sand scale, wherein the block side length is equal to four times of the pixelation amount of the sand scale, the first filtering scale is equal to the pixelation amount of the sand scale, and the second filtering scale is equal to twice of the pixelation amount of the sand scal