Intelligent traceability method and system of electronic scale and electronic scale

Abstract

The invention provides an intelligent traceability method and system for an electronic scale and the electronic scale, and the method comprises the steps: collecting product weight data obtained through the measurement of the electronic scale, and obtaining a weight data sequence arranged according to a time sequence; performing abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence marked to be abnormal; performing trend analysis on the weight data sequence marked abnormally to obtain a weight trend curve; and performing traceability judgment according to the weight trend curve in combination with a preset product standard weight range to obtain a traceability judgment result. According to the method and the device, the product weight data measured by the electronic scale are collected, abnormal fluctuation detection and trend analysis are performed, and then traceability judgment is performed, so that the defect that the current electronic scale cannot accurately trace whether the product is abnormal or not according to the product weight information is overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to an intelligent traceability method and system for an electronic scale, and an electronic scale. Background Art

[0002] With the increasing complexity of modern commodity supply chains and the continuous improvement of consumers' attention to product quality and origin, product traceability technology plays a key role in ensuring product quality and safety and enhancing consumer trust. Among many product monitoring links, the electronic scale, as an important tool for obtaining product weight information, is widely used.

[0003] Most traditional electronic scales only have basic weight measurement functions and can only instantaneously display the weight value of the measured product. For the processing of product weight data, it is usually just a simple single record or short-term storage of a small amount of data, and it is difficult to form a systematic and continuous weight data sequence. This makes traditional electronic scales appear inadequate when faced with scenarios that require long-term tracking of product weight change trends.

[0004] When it comes to product traceability, due to the inability to deeply analyze and effectively correlate weight data, it is difficult to accurately trace whether a product is abnormal through weight information.

[0005] In summary, there are deficiencies in the existing electronic scale technology in terms of intelligent traceability, and it is difficult to meet the urgent needs of the current market for product quality control and accurate traceability. Summary of the Invention

[0006] The main purpose of the present invention is to provide an intelligent traceability method and system for an electronic scale, and an electronic scale, aiming to overcome the defect that the current electronic scale cannot accurately trace whether a product is abnormal through product weight information.

[0007] To achieve the above purpose, the present invention provides an intelligent traceability method for an electronic scale, including the following steps:

[0008] Collect the product weight data measured by the electronic scale to obtain a weight data sequence arranged in chronological order;

[0009] Perform abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence marked with abnormalities;

[0010] Perform trend analysis on the weight data sequence marked with abnormalities to obtain a weight trend curve;

[0011] According to the weight trend curve, combined with the preset product standard weight range, perform traceability judgment to obtain a traceability judgment result.

[0012] Further, the performing abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence marked with abnormalities includes:

[0013] Based on the statistical analysis algorithm, calculate the change rate between adjacent data in the weight data sequence. When the change rate exceeds the preset threshold, mark it as an abnormal point to obtain the weight data sequence with marked abnormalities.

[0014] Further, the abnormal fluctuation detection of the weight data sequence to obtain the weight data sequence with marked abnormalities includes:

[0015] Quantize the data in the weight data sequence and map it to the quantum state space, and calculate the transition probability between quantum states;

[0016] When the deviation value of the transition probability from the normal range exceeds the preset threshold, mark it as an abnormal point to obtain the weight data sequence with marked abnormalities.

[0017] Further, according to the weight trend curve, combined with the preset product standard weight range, perform traceability judgment to obtain the traceability judgment result, including:

[0018] Use the numerical differentiation method to calculate the curvature value of the weight trend curve at each sampling point to obtain the curvature sequence;

[0019] Obtain the product standard weight range of products from normal sources and calculate the corresponding standard weight trend curve; calculate the curvature of the standard weight trend curve at each sampling point to obtain the standard curvature sequence;

[0020] Adopt the kernel density estimation method to calculate the similarity between the curvature sequence and the standard curvature sequence; among them, use the kernel density estimation method to calculate the probability density function of the curvature sequence, and calculate the divergence between the probability density function and the standard curvature sequence to evaluate the similarity between the two;

[0021] If the similarity is less than the threshold, it is determined that there is an abnormality in the source of the product measured by the electronic scale; if the similarity is not less than the threshold, it is determined that the source of the product measured by the electronic scale is normal.

[0022] Further, after obtaining the traceability judgment result, it includes:

[0023] Based on the traceability judgment result, match the corresponding undirected graph structure template;

[0024] Overlay the weight trend curve onto the undirected graph structure template to divide the undirected graph structure template into multiple regions; among them, multiple key points of the weight trend curve coincide with the specified positions of the undirected graph structure template;

[0025] According to the positional relationship and area relationship of each region, assign a corresponding number to each region and configure serial numbers for the nodes in each region;

[0026] Add the data in the weight data sequence to the undirected graph structure template in the order of the numbers of each region and the sequence numbers of each node, and obtain an undirected graph of weight data for local storage.

[0027] Further, after obtaining the undirected graph of weight data for local storage, it includes:

[0028] Generate a management key based on the undirected graph of weight data and the weight trend curve;

[0029] Encrypt the weight data sequence based on the management key and transmit it to the management terminal.

[0030] Further, generating a management key based on the undirected graph of weight data and the weight trend curve includes:

[0031] Calculate the eigenvalues and eigenvectors of the Laplacian matrix of the undirected graph of weight data to obtain a topological feature vector containing the connectivity of the graph and node distribution information;

[0032] Calculate the fractal dimension of the weight trend curve, and extract the turning point and extreme point features of the weight trend curve, and combine them into a geometric feature vector;

[0033] Based on the correlation function of vector space mapping, map the topological feature vector and the geometric feature vector into the same high-dimensional vector space, calculate the relative position relationship in the high-dimensional vector space, and obtain a correlation quantization value;

[0034] Perform a chaotic mapping on the correlation quantization value to generate a chaotic sequence;

[0035] Select a subsequence with a preset length from the chaotic sequence according to a preset rule, and perform encoding conversion on the subsequence to obtain the management key.

[0036] Further, generating a management key based on the undirected graph of weight data and the weight trend curve includes:

[0037] Calculate the degree centrality, betweenness centrality, and closeness centrality of each node in the undirected graph of weight data to obtain a set of node centrality feature vectors;

[0038] Calculate the slope change rate, curvature change rate, positions of peaks and valleys, and amplitudes of the weight trend curve to obtain a curve geometric feature vector;

[0039] Take the centroid of the undirected graph of weight data as the origin, and establish a coordinate system with the horizontal and vertical directions as the coordinate axes, and map the weight trend curve into the coordinate system;

[0040] For each key point of the weight trend curve, calculate its Euclidean distance to all nodes in the undirected graph of weight data to obtain a distance matrix;

[0041] Normalize the distance matrix to obtain a normalized distance matrix;

[0042] Fuse the node centrality feature vector set, the curve geometric feature vector, and the normalized distance matrix to obtain a fused feature vector.

[0043] Perform a hashing process on the fused feature vector to obtain an initial hash value; perform an iterative obfuscation process on the initial hash value to obtain the management key.

[0044] The present invention also provides an intelligent traceability system for an electronic scale, including:

[0045] An acquisition unit for acquiring product weight data measured by the electronic scale to obtain a weight data sequence arranged in chronological order;

[0046] A detection unit for detecting abnormal fluctuations in the weight data sequence to obtain a weight data sequence with marked abnormalities;

[0047] An analysis unit for performing trend analysis on the weight data sequence with marked abnormalities to obtain a weight trend curve;

[0048] A traceability unit for making a traceability judgment based on the weight trend curve and in combination with a preset product standard weight range to obtain a traceability judgment result.

[0049] The present invention also provides an electronic scale, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of the method described in any one of the above are implemented.

[0050] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described in any one of the above are implemented.

[0051] The intelligent traceability method, system, and electronic scale provided by the present invention include: collecting the product weight data measured by the electronic scale to obtain a weight data sequence arranged in chronological order; performing abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence with marked abnormalities; performing trend analysis on the weight data sequence with marked abnormalities to obtain a weight trend curve; and performing traceability judgment based on the weight trend curve in combination with a preset product standard weight range to obtain a traceability judgment result. In the present invention, by collecting the product weight data measured by the electronic scale, performing abnormal fluctuation detection and trend analysis, and then performing traceability judgment, the defect that the current electronic scale cannot accurately trace whether the product is abnormal through the product weight information is overcome. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a schematic diagram of the steps of the intelligent traceability method of the electronic scale in an embodiment of the present invention;

[0053] Figure 2 is a structural block diagram of the intelligent traceability system of the electronic scale in an embodiment of the present invention;

[0054] Figure 3 is a schematic structural block diagram of the electronic scale in an embodiment of the present invention.

[0055] The implementation, functional features, and advantages of the present invention will be further described in conjunction with the embodiments with reference to the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0057] Referring to Figure 1 , an embodiment of the present invention provides an intelligent traceability method for an electronic scale, including the following steps:

[0058] Step S1: Collect the product weight data measured by the electronic scale to obtain a weight data sequence arranged in chronological order;

[0059] Step S2: Perform abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence with marked abnormalities;

[0060] Step S3: Perform trend analysis on the weight data sequence with marked abnormalities to obtain a weight trend curve;

[0061] Step S4: Perform traceability judgment based on the weight trend curve in combination with a preset product standard weight range to obtain a traceability judgment result.

[0062] In this embodiment, as described in step S1 above, during various links such as product production, transportation, and sales, the electronic scale measures the weight of the product in real time. Through a dedicated data acquisition module, the above weight data is continuously obtained and recorded in chronological order. This process is like building a timeline, arranging the weight data corresponding to each measurement moment in an orderly manner to form a weight data sequence. This sequence not only contains the weight information of the product at different time points but also provides the original data basis for subsequent analysis. For example, in the supply chain of agricultural products, from the initial weighing after picking in the field, to multiple spot-check weighings during transportation, and then to the final weighing at the sales terminal, these data are collected in sequence, laying the foundation for comprehensively understanding the weight change process of agricultural products.

[0063] As described in step S2 above: Data mining and analysis algorithms are adopted, such as the 3σ criterion based on statistics and the ARIMA model in time series analysis. These algorithms analyze the weight data sequence, calculate statistical features such as the mean and standard deviation of the data, and time series features such as the autocorrelation between data. When a data point deviates from the normal range by more than a set threshold (automatically determined by the algorithm according to the data characteristics or manually set in combination with product characteristics), it is marked as an abnormal point, and then a weight data sequence with marked abnormalities is obtained. For example, in the production process of pharmaceuticals, if during the weighing link before packaging of a certain batch of pharmaceuticals, individual weight data deviates too much from the overall data, it can be detected in time through abnormal fluctuation detection, preventing unqualified products from flowing into the next link.

[0064] As described in step S3 above, for the weight data sequence with marked abnormalities, trend analysis methods are used, such as the moving average method and the exponential smoothing method. The moving average method smooths data fluctuations and highlights the long-term trend of the data by calculating the average value of data within a certain time window. The exponential smoothing method assigns different weights according to the time distance of the data, paying more attention to the influence of recent data on the trend. Through these methods, the weight data sequence is transformed into a weight trend curve. This curve intuitively shows the change trend of the product weight over time, helping users quickly understand the overall trend of the product weight. For example, during the aging test of electronic products, the subtle change trend of the product weight during the aging process can be clearly seen through the weight trend curve to judge whether the product quality is stable.

[0065] As described in step S4 above, based on the weight trend curve obtained in step S3, in combination with a pre-set standard weight range for the product (this range can be determined based on various factors such as the production process standards of the product, historical data statistical analysis, industry norms, etc.). By comparing the weight trend curve with the standard weight range, it is judged whether the product weight fluctuates within the normal range. If the weight trend curve is completely within the standard weight range, it can be preliminarily determined that the product source is normal; if part or all of the curve exceeds the standard range, further analyze information such as the time, frequency of the abnormal points and their association with the product supply chain links, so as to infer which link the product may have problems, achieve traceability judgment, and obtain the traceability judgment result. For example, in the food processing industry, if the weight trend curve of a certain brand of food exceeds the standard range, through traceability judgment, it can be traced whether there is a problem with the weight control in the raw material procurement link or the weight control in the production and processing process is inaccurate.

[0066] In one embodiment, the detecting abnormal fluctuations in the weight data sequence to obtain a marked abnormal weight data sequence includes:

[0067] Based on a statistical analysis algorithm, calculate the change rate between adjacent data in the weight data sequence, and mark it as an abnormal point when the change rate exceeds a preset threshold, to obtain a marked abnormal weight data sequence.

[0068] In this embodiment, the above statistical analysis algorithm can deeply explore the change situation between adjacent data in the weight data sequence. Simply put, it is to compare every two adjacent data to see how much the latter data has changed compared to the former data. For example, in a sequence recording the weights of fruits, first an apple is weighed at 200 grams, and then the next apple is weighed at 210 grams. The statistical analysis algorithm can calculate the degree of change in the weights of these two apples. It will compare every pair of adjacent data in the entire weight data sequence in this way, and obtain a series of values representing the degree of data change.

[0069] The preset threshold is used to determine whether the change in data is normal. Its setting is not arbitrary and many factors need to be considered. First, it is the characteristics of the product itself and the standard requirements of the industry. For different products, their normal ranges of weight change are different. For example, for precision electronic components, it is normal for their weight change to be very small; while for some building materials, a slight weight change may also be normal. Secondly, a large amount of historical weight data of similar products in the past will also be referred to. Through the analysis of this historical data, we can know what range the degree of data change is approximately under normal circumstances. Then, based on this information, a suitable value is determined. If it exceeds this value, it indicates that the data change may not be normal. Moreover, this threshold is not fixed. If the production process of the product is improved or the requirements for the product weight become higher, this threshold will be adjusted again.

[0070] After calculating the degree of change between each pair of adjacent data, these degrees of change are compared with the previously set preset threshold. Once it is found that a certain degree of change exceeds the threshold, it indicates that the corresponding data point is very likely to be abnormal. At this time, a special "mark" will be made for this data point. This mark can be in many forms, such as adding an asterisk next to the data, or recording the position and relevant situation of this data point in a special record form.

[0071] After completing such comparison and marking operations for each pair of adjacent data in the entire weight data sequence, we obtain a new weight data sequence in which the data points with abnormal fluctuations have been clearly marked. This new sequence is like a list with "problem prompts", which can quickly tell which data may be problematic and provides an important basis for further analyzing the trend of product weight change and finding the source of the problem later.

[0072] In one embodiment, the detecting abnormal fluctuations in the weight data sequence to obtain a weight data sequence with marked abnormalities includes:

[0073] Quantize and map the data in the weight data sequence to the quantum state space, and calculate the transition probability between quantum states;

[0074] When the deviation value of the transition probability from the normal range exceeds the preset threshold, mark it as an abnormal point to obtain a weight data sequence with marked abnormalities.

[0075] In this embodiment, traditional weight data is classical numerical information, while the quantum state space is an abstract space constructed based on the concepts of quantum mechanics. Quantizing and mapping the weight data into the quantum state space means transforming the original classical weight data into information represented by quantum states. Quantum states have properties such as superposition and entanglement, enabling data to be represented and processed in a completely new way. This mapping process can utilize the concept of qubits (quantum bits) to encode the weight data into different state combinations of the quantum state.

[0076] First, a suitable mapping rule needs to be determined. For example, according to the value range of the weight data, it can be divided into several intervals, with each interval corresponding to a specific quantum state in the quantum state space. For a continuous sequence of weight data, each data is transformed into the corresponding quantum state according to this mapping rule. It's like finding a "position" for each weight data in the quantum state space. By mapping the weight data into the quantum state space, the properties of quantum mechanics can be utilized to process data, providing more information and more powerful processing capabilities for subsequent analysis. The superposition property of quantum states allows multiple states to be processed simultaneously, improving the efficiency and accuracy of data processing.

[0077] In the quantum state space, quantum states are not fixed; there is a possibility of mutual conversion between them, and this conversion possibility is represented by the transition probability. The calculation of the transition probability is based on relevant theories and models in quantum mechanics, such as the Schrödinger equation in quantum mechanics. These theories describe the evolution of quantum states over time and the laws of mutual conversion between different quantum states. According to the characteristics and mutual relationships of each quantum state in the quantum state space, algorithms in quantum mechanics are used to calculate the transition probability between adjacent or related quantum states. This requires considering factors such as the energy and phase of the quantum state. For example, by analyzing and calculating the wave function of the quantum state, the probability values of transitions between different quantum states are obtained. The transition probability reflects the dynamic changes of the weight data in the quantum state space. Under normal circumstances, the transition probability between the quantum states corresponding to the weight data should be within a relatively stable range. By calculating the transition probability, small changes and potential abnormal fluctuations in the weight data can be captured.

[0078] The above normal range is obtained based on a large amount of historical data statistical analysis. By collecting and analyzing the quantum state transition probabilities corresponding to the weight data under normal past conditions, a reasonable probability interval is determined, and this interval is the normal range. For example, after quantizing the weight data of multiple batches of products and calculating the transition probabilities, it is found that in most cases, the transition probabilities are between 0.1 and 0.3, so this interval can be used as the normal range. The preset threshold is a critical value used to judge whether the transition probability is abnormal. Its determination needs to comprehensively consider multiple factors, such as the characteristics of the product, the accuracy requirements of the measurement, etc. If the threshold is set too small, it may lead to too many misjudgments; if the threshold is set too large, some real abnormal situations may be missed. Usually, a suitable preset threshold is determined through experiments and experience. For example, the preset threshold is set to 0.05, that is, when the deviation value of the transition probability from the normal range exceeds 0.05, it is considered that there is an abnormality.

[0079] Compare each calculated transition probability with the normal range and calculate its deviation value. If the deviation value exceeds the preset threshold, mark the weight data point corresponding to the transition probability as an abnormal point. The marking method can be various. For example, add a special identifier in the data sequence, or record the relevant information of this data point in a dedicated abnormal record form.

[0080] In one embodiment, according to the weight trend curve, combined with the preset product standard weight range, perform traceability judgment to obtain a traceability judgment result, including:

[0081] Use the numerical differentiation method to calculate the curvature values of the weight trend curve at each sampling point to obtain a curvature sequence;

[0082] Obtain the product standard weight range of products from normal sources and calculate the corresponding standard weight trend curve; calculate the curvature of the standard weight trend curve at each sampling point to obtain a standard curvature sequence;

[0083] Adopt the kernel density estimation method to calculate the similarity between the curvature sequence and the standard curvature sequence; among them, use the kernel density estimation method to calculate the probability density function of the curvature sequence, and calculate the divergence between the probability density function and the standard curvature sequence to evaluate the similarity between the two;

[0084] If the similarity is less than the threshold, it is determined that the source of the product measured by the electronic scale is abnormal; if the similarity is not less than the threshold, it is determined that the source of the product measured by the electronic scale is normal.

[0085] In this embodiment, numerical differentiation is a method for approximately calculating derivatives through discrete data. When dealing with the weight trend curve, since the curve is composed of a series of discrete sampling points, traditional calculus methods cannot be directly used to calculate the curvature. The numerical differentiation method utilizes the data of these sampling points and approximates the calculation of the first derivative and second derivative at each point of the curve through a specific algorithm, and then obtains the curvature value.

[0086] The above-mentioned curvature reflects the degree of bending of the curve. For the weight trend curve, at each sampling point, according to the first derivative and second derivative obtained by numerical differentiation, substituting them into the curvature calculation formula can calculate the curvature at that point. By performing such calculations on each sampling point of the weight trend curve, a series of curvature values can be obtained.

[0087] Arranging the calculated curvature values of each sampling point in the order of the sampling points forms a curvature sequence. This curvature sequence can more finely reflect the bending changes of the weight trend curve. Compared with the original weight trend curve, the curvature sequence provides more in-depth information about the shape characteristics of the curve and provides a richer basis for subsequent traceability judgment.

[0088] The product standard weight range is determined based on factors such as the historical data of products from normal sources, production process requirements, and industry standards. By collecting a large amount of weight data of products from normal production processes with reliable quality and conducting statistical analysis, a reasonable weight fluctuation range is determined, and this range is the product standard weight range. For example, for a certain brand of canned beverage, based on multiple production inspection data, its standard weight range is determined to be between 330 grams and 335 grams. According to the obtained product standard weight range, combined with the weight change laws of normal products at different stages of production, transportation, sales, etc., a standard weight trend curve is simulated. This curve represents the change trend of product weight over time or other relevant factors under normal circumstances. For example, in the food production process, from raw material procurement to finished product leaving the factory, the product weight will change to a certain extent as the processing process progresses. By analyzing and modeling these normal changes, the standard weight trend curve can be obtained. Using the same numerical differentiation method as above, the curvature values are calculated at each sampling point of the standard weight trend curve. Arranging these curvature values in the order of the sampling points, a standard curvature sequence is obtained. The standard curvature sequence represents the bending characteristics of the normal product weight trend curve and is a reference standard for subsequent judgment of whether the product to be tested is normal.

[0089] Furthermore, the above-mentioned kernel density estimation is a non-parametric statistical method used to estimate the probability density function of a random variable. For the curvature sequence, since its data distribution may be complex and does not satisfy a specific parametric distribution form, the kernel density estimation method can be used to more flexibly estimate its probability density function. This method places a kernel function (such as a Gaussian kernel function) at each data point and performs a weighted sum of these kernel functions to obtain the probability density function of the entire curvature sequence.

[0090] Take the curvature sequence to be detected and the standard curvature sequence as inputs respectively, and use the kernel density estimation method to calculate their respective probability density functions. The probability density function describes the distribution probability of curvature values at different values, and through this function, the statistical characteristics of the curvature sequence can be more comprehensively understood.

[0091] The above-mentioned divergence is an index to measure the degree of difference between two probability distributions. Common divergences include Kullback-Leibler (KL) divergence, etc. By calculating the divergence between the probability density function of the curvature sequence to be detected and the probability density function of the standard curvature sequence, the degree of difference between the two can be quantified. The smaller the divergence value, the more similar the distributions of the two sequences; the larger the divergence value, the more obvious the difference. Therefore, the divergence can be used as an effective index to evaluate the similarity between the curvature sequence and the standard curvature sequence.

[0092] The above-mentioned threshold is a critical value determined based on a large amount of experimental data and practical application experience. By calculating the similarity between the curvature sequences of known normal and abnormal products and the standard curvature sequence, and analyzing the distribution of divergence values in different situations, a suitable threshold is determined. For example, through multiple experiments, it is found that when the divergence value is less than 0.1, the product is from a normal source; when the divergence value is greater than 0.1, the product is more likely to have an abnormal source, then the threshold can be set to 0.1.

[0093] Compare the calculated similarity (i.e., the divergence value) with the set threshold. If the similarity is less than the threshold, it means that the curvature sequence to be detected is significantly different from the standard curvature sequence, then it is determined that the source of the product measured by the electronic scale is abnormal; if the similarity is not less than the threshold, it means that the two are relatively similar, and it is determined that the source of the product measured by the electronic scale is normal. In this way, the traceability judgment of the product source is realized by using the similarity analysis of the curvature sequence.

[0094] In one embodiment, after obtaining the traceability judgment result, it includes:

[0095] Based on the traceability judgment result, match the corresponding undirected graph structure template;

[0096] Superimpose the weight trend curve onto the undirected graph structure template to divide the undirected graph structure template into multiple regions; wherein, multiple key points of the weight trend curve coincide with designated positions of the undirected graph structure template;

[0097] According to the positional relationship and area relationship of each region, assign corresponding numbers to each region, and configure serial numbers for the nodes in each region;

[0098] In the order of the numbers of each region and the serial numbers of each node, add the data in the weight data sequence to the undirected graph structure template in sequence to obtain an undirected graph of weight data for local storage.

[0099] In this embodiment, the above-mentioned undirected graph structure template is a pre-designed graphic framework with a specific topological structure. Different templates can represent different product traceability scenarios or data organization methods. It is like a general container that can accommodate and display weight data-related information according to certain rules, providing a structured basis for subsequent data processing and analysis.

[0100] The traceability judgment result contains information such as whether the product source is normal and the possible abnormal links. According to this information, the most matching template will be selected from multiple pre-stored undirected graph structure template libraries. For example, if the traceability judgment result shows that there may be an abnormality in the production link of the product, an undirected graph structure template that highlights the relevant nodes and connection relationships in the production link will be matched; if it is judged that the product source is normal, a more general template that comprehensively displays the entire supply chain process will be selected.

[0101] Place the previously obtained weight trend curve on the selected undirected graph structure template, so that multiple key points of the weight trend curve coincide with the designated positions of the undirected graph structure template. The above key points can be turning points, extreme value points, etc. of the weight trend curve that have special significance, and the designated positions are the positions preset in the undirected graph structure template corresponding to the product supply chain links or important data nodes. Through this superposition, a close association is established between the weight trend curve and the undirected graph structure template.

[0102] The existence of the weight trend curve will divide the undirected graph structure template, splitting it into multiple different regions. The above regions are formed based on the relative positional relationship between the weight trend curve and the undirected graph structure. For example, in an undirected graph structure template representing the process from raw material procurement to finished product sales of a product, the weight trend curve may pass through or surround certain nodes and connections at different supply chain stages, thus dividing the entire graph into different regions representing the raw material procurement stage, the production and processing stage, the transportation stage, etc. Each region corresponds to a specific link or time period in the product supply chain, providing a basis for subsequent data classification and organization.

[0103] The positional relationship of the regions reflects their relative positions in the undirected graph structure template, such as up and down, left and right, front and back, etc.; the area relationship reflects the scope covered by each region. Based on these relationships, a unique number is assigned to each region. The order of numbering can be determined according to rules such as from left to right, from top to bottom, etc., which helps subsequent orderly access and processing of data. For example, in an undirected graph divided into multiple regions, the region in the upper left corner may be numbered 1, and the region adjacent to its right is numbered 2, and so on.

[0104] Inside each region, serial numbers are configured for the nodes therein. The node serial numbers can be arranged according to the importance, logical order, etc. of the nodes within the region. For example, in a region representing the production and processing stage, according to the sequence of the production process, the raw material input node is numbered 1, the first processing procedure node is numbered 2, and so on. By numbering and configuring serial numbers for the regions and nodes respectively, a clear data indexing system is established, facilitating subsequent accurate positioning and organization of weight data.

[0105] According to the numbers and serial numbers assigned to the regions and nodes previously, the data in the weight data sequence are sequentially added to the corresponding positions in the undirected graph structure template. For example, start from the region numbered 1, and according to the serial number order of the nodes in this region, add the corresponding data in the weight data sequence to the nodes; after completing the region numbered 1, then process the region numbered 2, and so on. In this way, the weight data is in one-to-one correspondence with each node and region of the undirected graph, forming an undirected graph containing weight data information, that is, the weight data undirected graph.

[0106] Locally storing the undirected graph of weight data means storing the processed graphical data containing rich traceability information in a local device or storage system. Local storage facilitates subsequent queries, analyses, and comparisons of this data at any time. For example, when it is necessary to further delve into the relationship between the weight change trend of a certain product and the supply chain links, the undirected graph of weight data can be directly retrieved from the local for detailed analysis. At the same time, local storage also provides a certain degree of guarantee for data security and privacy, avoiding possible loss or leakage of data during transmission.

[0107] In one embodiment, after obtaining the undirected graph of weight data for local storage, it includes:

[0108] Generating a management key based on the undirected graph of weight data and the weight trend curve;

[0109] Encrypting the weight data sequence based on the management key and transmitting it to the management terminal.

[0110] In one embodiment, generating a management key based on the undirected graph of weight data and the weight trend curve includes:

[0111] Calculating the eigenvalues and eigenvectors of the Laplacian matrix of the undirected graph of weight data to obtain a topological feature vector containing the connectivity of the graph and node distribution information;

[0112] Calculating the fractal dimension of the weight trend curve, and extracting the turning point and extreme point features of the weight trend curve, and combining them into a geometric feature vector;

[0113] Based on the correlation function of vector space mapping, mapping the topological feature vector and the geometric feature vector into the same high-dimensional vector space, calculating the relative position relationship in the high-dimensional vector space, and obtaining a correlation quantization value;

[0114] Performing a chaotic mapping on the correlation quantization value to generate a chaotic sequence;

[0115] Selecting a subsequence of a preset length from the chaotic sequence according to a preset rule, and performing encoding conversion on the subsequence to obtain the management key.

[0116] In this embodiment, for the undirected graph of weight data, its Laplacian matrix is first constructed. The Laplacian matrix is an important tool for describing the structural properties of a graph, and it is related to the adjacency matrix and degree matrix of the graph. By calculating the degrees of nodes in the undirected graph and the connection relationships between nodes, the Laplacian matrix is obtained. For example, for an undirected graph with n nodes, the elements of its Laplacian matrix can be determined according to whether nodes i and j are adjacent and their degrees.

[0117] Next, calculate the eigenvalues and eigenvectors of the Laplacian matrix. This is a common operation in linear algebra, by solving the eigen-equation \(Lx = \lambda x\), where \(L\) is the Laplacian matrix, \(\lambda\) is the eigenvalue, and \(x\) is the corresponding eigenvector. The obtained eigenvalues and eigenvectors contain important information such as the connectivity of the graph and the node distribution. For example, smaller eigenvalues correspond to larger connected components of the graph, while eigenvectors describe the distribution of nodes in different connected components. Organize and combine the calculated eigenvalues and eigenvectors to form a topological feature vector. This vector can be used as a feature representation of the undirected graph of weight data, which condenses the key topological information of the graph and provides a basis for subsequent feature fusion with the weight trend curve.

[0118] The above fractal dimension is an important indicator to describe the complexity of the weight trend curve. Calculate the fractal dimension of the weight trend curve through a preset algorithm, such as the box dimension algorithm. This algorithm covers the curve with boxes of different scales, calculates the relationship between the number of boxes and the scale, and thus obtains the fractal dimension. The fractal dimension reflects the self-similarity and roughness of the curve. A higher fractal dimension indicates that the curve is more complex, with more details and variations. The turning point is the point where the slope of the weight trend curve changes, and the extreme point is the maximum or minimum point of the curve. Determine the positions and values of the turning points and extreme points by performing numerical analysis on the curve, such as calculating the first derivative and second derivative of the curve. The point where the first derivative is zero may be an extreme point, and the point where the second derivative is zero may be a turning point. These points reflect the key characteristics of the change of weight over time or other variables. Combine the characteristics of the fractal dimension, turning points, and extreme points to form a geometric feature vector. This vector describes the characteristics of the weight trend curve from a geometric perspective, including the overall complexity of the curve and the information of key change points, providing a basis for associating with the topological features of the undirected graph.

[0119] Then, using the correlation function of vector space mapping, the topological feature vector and the geometric feature vector are respectively mapped into a high-dimensional vector space. This mapping process can be achieved through specific mathematical functions, such as kernel functions. The kernel function maps the vectors in the low-dimensional space into the high-dimensional space, making the non-linear relationships that are difficult to handle in the low-dimensional space become easier to analyze in the high-dimensional space. By selecting an appropriate kernel function, the topological feature vector and the geometric feature vector can be reasonably represented in the high-dimensional space. In the high-dimensional vector space, calculate the relative position relationship between the topological feature vector and the geometric feature vector. This can be achieved by calculating the distance, angle, or other similarity metrics between them. For example, the Euclidean distance can be used to calculate the distance between two vectors, and the smaller the distance, the more similar the two vectors are; or the cosine similarity can be used to calculate the cosine value of the angle between two vectors, and the closer the cosine value is to 1, the more similar the directions of the two vectors are. Through these calculations, an association quantization value is obtained, which is used to measure the degree of association between the weight data undirected graph and the weight trend curve.

[0120] Furthermore, chaotic mapping is a method of generating a chaotic sequence by non-linearly transforming the input value. Select an appropriate chaotic mapping function, such as the Logistic mapping or the Henon mapping, take the association quantization value as the input, and generate a chaotic sequence through multiple iterative calculations. The chaotic sequence is sensitive to the initial conditions, seemingly random but actually deterministic. Even a slight change in the association quantization value will generate a significantly different chaotic sequence after chaotic mapping, which increases the randomness and security of the management key. The generated chaotic sequence has good pseudo-randomness and ergodicity. Pseudo-randomness means that the sequence statistically exhibits the characteristics of a random sequence and it is difficult to predict its future values; ergodicity ensures that the sequence can evenly traverse all possible values within a certain range, making the generated management key have a high degree of diversity and complexity.

[0121] Finally, according to the preset rules, select a subsequence of a preset length from the chaotic sequence. The preset rules can be to select elements at a certain interval, or to select elements that meet the requirements according to specific conditions to form a subsequence. The length and selection rules of the subsequence can be adjusted according to actual needs and security requirements. Perform encoding conversion on the selected subsequence to obtain the management key. The encoding conversion can be carried out in various ways, such as mapping each element in the subsequence to a specific character set or number set, and then combining these characters or numbers into a string or number sequence as the management key. More complex encoding algorithms can also be used, such as taking the subsequence as the input and generating a management key with a fixed length through a cryptographic hash function or other encoding functions. The management key generated in this way has high security and uniqueness, and can effectively protect the security of the weight data sequence.

[0122] In one embodiment, generating a management key based on the undirected graph of weight data and the weight trend curve includes:

[0123] Calculating the degree centrality, betweenness centrality, and closeness centrality of each node in the undirected graph of weight data to obtain a set of node centrality feature vectors;

[0124] Calculating the slope change rate, curvature change rate, positions of peaks and valleys, and amplitude of the weight trend curve to obtain a curve geometric feature vector;

[0125] Taking the centroid of the undirected graph of weight data as the origin and the horizontal and vertical directions as coordinate axes to establish a coordinate system, and mapping the weight trend curve into the coordinate system;

[0126] For each key point of the weight trend curve, calculating its Euclidean distance to all nodes in the undirected graph of weight data to obtain a distance matrix;

[0127] Normalizing the distance matrix to obtain a normalized distance matrix;

[0128] Fusing the set of node centrality feature vectors, the curve geometric feature vector, and the normalized distance matrix to obtain a fused feature vector.

[0129] Performing a hashing process on the fused feature vector to obtain an initial hash value; performing an iterative confusion process on the initial hash value to obtain the management key.

[0130] In this embodiment, the above degree centrality measures the number of connections of a node in a graph. In the undirected graph of weight data, the degree centrality of a node is the number of edges directly connected to that node. The higher the degree of a node, the more central it is in the graph and the more direct interactions it has with other nodes. By traversing each node in the graph and counting the number of connected edges, the degree centrality value of each node can be obtained.

[0131] The above betweenness centrality reflects the degree to which a node serves as a bridge for the shortest paths between other nodes in a graph. When calculating, it is necessary to find the shortest paths between all pairs of nodes in the graph and then count the number of times each node appears on these shortest paths. The higher the betweenness centrality of a node, the more crucial its mediating role is in information transmission or data flow.

[0132] The above closeness centrality measures the average distance of a node to all other nodes in a graph. The smaller the average distance, the higher the closeness centrality of the node, which means that the node can interact with other nodes in the graph more quickly. By calculating the shortest path lengths from a node to all other nodes and taking the average, the closeness centrality of the node can be obtained.

[0133] Combine the three values of degree centrality, betweenness centrality, and closeness centrality of each node into a vector. The vectors corresponding to all nodes in the graph form the set of node centrality eigenvectors. This set describes the importance and status of nodes in an undirected graph from different perspectives.

[0134] The above slope change rate reflects the change in the rising or falling speed of the weight trend curve. By numerically differentiating the curve, calculating the slope between adjacent points on the curve, and further calculating the change rate of these slopes. A larger slope change rate indicates that the inclination of the curve changes rapidly, which may mean that the change speed of the weight is unstable.

[0135] The above curvature change rate describes the change in the degree of curve bending. Curvature measures the degree to which a curve deviates from a straight line, and the curvature can be obtained by calculating the second derivative of the curve. The curvature change rate is the change of curvature with respect to the position of the curve. It can reflect the bending mode and complexity of the weight trend curve.

[0136] The above peak value is the maximum point on the curve, and the valley value is the minimum point. By traversing the data points of the curve, comparing the sizes of adjacent points, the peak and valley values in the curve are found. At the same time, record the positions (i.e., the corresponding abscissas) and amplitudes (i.e., the corresponding ordinate values) where these peak and valley values appear.

[0137] Combine information such as the slope change rate, curvature change rate, positions and amplitudes of peak and valley values into a vector. This vector is the curve geometric feature vector, which comprehensively describes the characteristics of the weight trend curve from a geometric perspective.

[0138] The above centroid is the geometric center of the graph. For an undirected graph of weight data, the position of the centroid can be determined by calculating the average value of all node coordinates. This centroid position will be used as the origin of the new coordinate system.

[0139] Taking the centroid as the origin, determine the horizontal and vertical directions as coordinate axes on the plane where the graph is located to construct a new coordinate system. The purpose of doing this is to place the weight trend curve and the undirected graph in the same coordinate system for convenient subsequent distance calculation. Transform the coordinates of each point on the weight trend curve according to the rules of the new coordinate system so that it is in the same coordinate system as the undirected graph. In this way, the points on the curve and the nodes in the graph have a unified coordinate representation.

[0140] The key points of the above weight trend curve usually include points with special significance such as peak points, valley points, turning points, etc. These points can reflect the main characteristics and changes of the curve. The above Euclidean distance refers to the straight-line distance between two points on a plane. For each key point of the curve, calculate its Euclidean distance to all nodes in the undirected graph. Use the Euclidean distance formula for calculation. Arrange all the calculated distance values into a matrix, where the rows of the matrix correspond to the key points of the curve and the columns correspond to the nodes in the undirected graph. This distance matrix reflects the spatial relationship between the key points of the curve and the nodes of the undirected graph.

[0141] The element values in the above distance matrix may have different orders of magnitude. To eliminate the influence of this order-of-magnitude difference on subsequent processing, it is necessary to normalize the distance matrix. Normalization can make the element values in the matrix within a unified range, which is convenient for feature fusion and comparison. A common normalization method is to subtract the minimum value in the matrix from each element in the matrix, and then divide by the difference between the maximum value and the minimum value in the matrix. In this way, the element values in the matrix are mapped into the interval [0,1] to obtain the normalized distance matrix.

[0142] The above set of node centrality feature vectors describes the characteristics of the nodes in the undirected graph, the curve geometric feature vector describes the geometric features of the weight trend curve, and the normalized distance matrix reflects the spatial relationship between the curve and the graph. Fusing these three parts of features can comprehensively utilize the information they contain to obtain a more comprehensive and representative feature vector. The splicing method can be adopted. Arrange the set of node centrality feature vectors in order to form a long vector, then expand the curve geometric feature vector and the normalized distance matrix into one-dimensional vectors, and then connect them in sequence to form a fused feature vector.

[0143] The above hash processing is the process of converting input data of any length (here it is the fused feature vector) into a hash value of a fixed length. Use a secure hash function, such as SHA-256, to process the fused feature vector to obtain an initial hash value. The hash function has one-wayness and collision resistance, that is, it is very difficult to reverse the original input data from the hash value, and it is almost impossible for different input data to obtain the same hash value. To further enhance the security and randomness of the key, iterative confusion processing is performed on the initial hash value. Iterative confusion processing can be carried out by performing multiple hash operations or combining other encryption algorithms. For example, take the initial hash value as the input and perform hash operations again, repeating this process multiple times. The result obtained after iterative confusion processing is the final management key, which can be used for security operations such as encrypting weight data.

[0144] In one embodiment, generating a management key based on the weight data undirected graph and the weight trend curve includes:

[0145] Extract the node connection relationship features from the undirected graph of weight data and transform them into an adjacency matrix;

[0146] Extract the key turning points from the weight trend curve to form a coordinate matrix;

[0147] Add the adjacency matrix and the coordinate matrix element by element to obtain a fusion matrix, and perform singular value decomposition on the fusion matrix to obtain a singular value sequence;

[0148] Construct a spiral curve with the singular value sequence as a parameter, and calculate the intersection position of the spiral curve and the weight trend curve;

[0149] Obtain the nodes at the intersection positions of the corresponding weight trend curves from the undirected graph of weight data as target nodes; obtain the attribute information of each target node, and use an encryption algorithm to generate the management key.

[0150] In this embodiment, the undirected graph of weight data is a relational network containing multiple nodes, and the nodes are connected by edges. The above adjacency matrix is a way to record the node connection relationship in the graph in tabular form. For this undirected graph, create a table with the same size as the number of nodes. If there is a connection between two nodes, mark "connected" (which can be represented by 1) at the corresponding position in the table; if there is no connection, mark "not connected" (represented by 0). In this way, the connection relationship information of the nodes in the undirected graph is organized into this table, and this table is the adjacency matrix.

[0151] The above weight trend curve depicts the change of the product weight over time or other factors. On this curve, there are some particularly important points, such as the points where the curve suddenly changes the rising or falling direction (inflection points), and the points where the curve reaches the maximum or minimum weight (extreme points). These points can help better understand how the weight changes. By analyzing the curve, determine the positions of these key turning points. After finding the key turning points, record the position information of each point, that is, its values on the horizontal axis (such as representing time) and the vertical axis (representing weight). Then organize the coordinate information of these points into a table, with each row corresponding to the horizontal and vertical coordinates of a key turning point. This table is the coordinate matrix, which concentrates the key information of the weight trend curve.

[0152] The adjacency matrix and the coordinate matrix are two different information tables. Since their sizes may be different, first adjust them to the same size. Then, add the numbers at the corresponding positions of these two tables to obtain a new table. This new table is the fusion matrix, which mixes the node connection information of the undirected graph and the key turning point information of the weight trend curve.

[0153] For this new table of the fusion matrix, we use a special analysis method to decompose it into several parts. After decomposition, some special numerical values will be obtained, and these values can reflect the important characteristics of the fusion matrix. Arrange these numerical values in descending order to form a sequence, and this sequence is the singular value sequence. The singular value sequence is like a set of key information extracted from the fusion matrix.

[0154] These numerical values in the singular value sequence can be used as parameters to determine a spiral curve. Select a shape of the spiral curve (such as a shape like a mosquito coil that coils around), and then adjust the characteristics such as the size and density of this spiral curve according to the numerical values in the singular value sequence. In this way, a spiral curve is constructed. Place the constructed spiral curve and the original weight trend curve in the same coordinate system, just like placing two paintings on the same piece of paper. Then find the places where these two curves intersect, that is, the points where they coincide. Through mathematical calculation methods, determine the specific positions of these intersection points, and the positions of these intersection points contain very important information that can help us further analyze.

[0155] According to the intersection positions found previously, return to the undirected graph of weight data (with the weight trend curve superimposed in the graph), and find the nodes corresponding to these intersection positions. These nodes are the target nodes needed, and each target node has its own some characteristics and information, such as its number, its connection situation with other nodes, and some attributes related to the product weight it represents, etc. Collect all this information about the target nodes. Finally, use an encryption method, take the attribute information of the collected target nodes as input, and after a series of encryption processes, generate the above-mentioned management key.

[0156] Refer to Figure 2 , in another embodiment of the present invention, an intelligent traceability system for an electronic scale is further provided, including:

[0157] A collection unit, configured to collect the product weight data measured by the electronic scale to obtain a weight data sequence arranged in chronological order;

[0158] A detection unit, configured to perform abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence marked with abnormalities;

[0159] An analysis unit, configured to perform trend analysis on the weight data sequence marked with abnormalities to obtain a weight trend curve;

[0160] A traceability unit, configured to perform traceability judgment based on the weight trend curve in combination with a preset product standard weight range to obtain a traceability judgment result.

[0161] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment and will not be elaborated here.

[0162] Refer to Figure 3 , an electronic scale is further provided in an embodiment of the present invention. The internal structure of the electronic scale may be as Figure 3 shown. The electronic scale includes a processor, a memory, a display screen, an input device, a network interface, and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the electronic scale includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the electronic scale is used to store the corresponding data in this embodiment. The network interface of the electronic scale is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0163] Those skilled in the art can understand that Figure 3 the structure shown in

[0164] is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the electronic scale to which the solution of the present invention is applied.

[0165] In summary, an intelligent traceability method, system, and electronic scale provided in an embodiment of the present invention include: collecting product weight data measured by the electronic scale to obtain a weight data sequence arranged in chronological order; performing abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence marked with abnormalities; performing trend analysis on the weight data sequence marked with abnormalities to obtain a weight trend curve; and making a traceability judgment based on the weight trend curve in combination with a preset product standard weight range to obtain a traceability judgment result. In the present invention, by collecting product weight data measured by the electronic scale, performing abnormal fluctuation detection and trend analysis, and then making a traceability judgment, the defect that the current electronic scale cannot accurately trace whether a product is abnormal through product weight information is overcome.

[0166] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium provided by the present invention and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0167] It should be noted that in this document, the terms "include", "comprise", or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, apparatus, article, or method that includes a series of elements includes not only those elements but also other elements not explicitly listed, or also includes elements inherent to such a process, apparatus, article, or method. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of another identical element in the process, apparatus, article, or method that includes the element.

[0168] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structural or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. An intelligent tracing method for an electronic scale, characterized in that: The following steps are involved: The product weight data measured by the electronic scale is collected to obtain a weight data sequence arranged in chronological order; Performing abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence marked with abnormality; Perform trend analysis on the weight data sequence with abnormal markings to obtain a weight trend curve; According to the weight trend curve, combined with the preset product standard weight range, a traceability judgment is performed to obtain a traceability judgment result.

2. The intelligent tracing method for electronic scales according to claim 1, characterized in that: The performing abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence marked with abnormality includes: Based on the statistical analysis algorithm, the change rate between adjacent data in the weight data sequence is calculated, and when the change rate exceeds a preset threshold, it is marked as an abnormal point to obtain a weight data sequence marked with abnormality.

3. The intelligent traceability method for electronic scales according to claim 1, characterized in that: The performing abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence marked with abnormality includes: Quantizing and mapping the data in the weight data sequence into a quantum state space, and calculating the transition probability between quantum states; When the deviation value of the transition probability from the normal range exceeds a preset threshold, it is marked as an abnormal point, and a weight data sequence marked with abnormality is obtained.

4. The intelligent traceability method for electronic scales according to claim 1, characterized in that: According to the weight trend curve, combined with the preset product standard weight range, a traceability judgment is performed to obtain a traceability judgment result, including: Using a numerical differentiation method to calculate the curvature value of the weight trend curve at each sampling point, to obtain a curvature sequence; Obtain the product standard weight range of normal source products and calculate the corresponding standard weight trend curve; calculate the curvature of the standard weight trend curve at each sampling point to obtain a standard curvature sequence; A kernel density estimation method is used to calculate the similarity between the curvature sequence and the standard curvature sequence; wherein the kernel density estimation method is used to calculate the probability density function of the curvature sequence, and the divergence between the probability density function and the standard curvature sequence is calculated to evaluate the similarity between the two; If the similarity is less than the threshold, it is determined that the source of the product measured by the electronic scale is abnormal; if the similarity is not less than the threshold, it is determined that the source of the product measured by the electronic scale is normal.

5. The intelligent traceability method for electronic scales according to claim 1, characterized in that: After obtaining the traceability judgment result, the following steps are included: Based on the tracing judgment result, matching the corresponding undirected graph structure template; The weight trend curve is superimposed on the undirected graph structure template to divide the undirected graph structure template into a plurality of regions; wherein a plurality of key points of the weight trend curve coincide with designated positions of the undirected graph structure template; According to the position and area relationship of each area, a corresponding number is assigned to each area, and a serial number is configured for the nodes in each area; According to the numbering order of each area and the serial number order of each node, the data in the weight data sequence is sequentially added to the undirected graph structure template to obtain an undirected graph of weight data for local storage.

6. The intelligent tracing method for electronic scales according to claim 5, characterized in that: After obtaining the undirected graph of weight data for local storage, it includes: generating a management key based on the weight data undirected graph and the weight trend curve; The weight data sequence is encrypted based on the management key and transmitted to a management terminal.

7. The intelligent tracing method for electronic scales according to claim 6, characterized in that: Based on the weight data undirected graph and the weight trend curve, generating a management key includes: Calculating the eigenvalues ​​and eigenvectors of the Laplacian matrix of the weight data undirected graph to obtain a topological eigenvector containing the connectivity and node distribution information of the graph; Calculating the fractal dimension of the weight trend curve, and extracting the turning point and extreme point features of the weight trend curve, and combining them into a geometric feature vector; Based on the correlation function of vector space mapping, the topological feature vector and the geometric feature vector are mapped to the same high-dimensional vector space, and the relative position relationship in the high-dimensional vector space is calculated to obtain the correlation quantization value; Performing chaotic mapping on the associated quantized values ​​to generate a chaotic sequence; A subsequence of a preset length is selected from the chaotic sequence according to a preset rule, and the subsequence is converted into a code to obtain the management key.

8. The intelligent traceability method for electronic scales according to claim 6, characterized in that: Based on the weight data undirected graph and the weight trend curve, generating a management key includes: Calculate the degree centrality, betweenness centrality and closeness centrality of each node in the weight data undirected graph to obtain a node centrality feature vector set; Calculating the slope change rate, curvature change rate, peak and valley positions, and amplitude of the weight trend curve to obtain a geometric characteristic vector of the curve; Taking the centroid of the weight data undirected graph as the origin and the horizontal and vertical directions as the coordinate axes, a coordinate system is established, and the weight trend curve is mapped into the coordinate system; For each key point of the weight trend curve, calculate the Euclidean distance from the key point to all nodes in the weight data undirected graph to obtain a distance matrix; Normalizing the distance matrix to obtain a normalized distance matrix; The node centrality feature vector set, the curve geometry feature vector and the normalized distance matrix are fused to obtain a fused feature vector. The fused feature vector is hashed to obtain an initial hash value; the initial hash value is iteratively obfuscated to obtain the management key.

9. An intelligent traceability system for electronic scales, characterized in that: include: A collection unit, used to collect product weight data obtained by electronic scale measurement to obtain a weight data sequence arranged in chronological order; A detection unit, used for performing abnormal fluctuation detection on the weight data sequence to obtain a weight data sequence marked with abnormality; An analysis unit, used for performing trend analysis on the weight data sequence marked with abnormality to obtain a weight trend curve; The traceability unit is used to perform traceability judgment according to the weight trend curve and a preset product standard weight range to obtain a traceability judgment result.

10. An electronic scale, comprising a memory and a processor, wherein a computer program is stored in the memory, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images