An agricultural detection laboratory whole-process digital traceability management method

Abstract

This invention discloses a method for full-process digital traceability management in agricultural testing laboratories, comprising: sample registration to generate identification information, extracting disturbance features to form warehousing benchmark features; sample transfer to each testing node, obtaining current disturbance features, comparing with warehousing benchmark features to generate consistency results; constructing a process time sequence reconstructor, performing primitive denoising and calibration processing, and reconstructing the time sequence to generate a process time sequence; inputting an improved TimeMixer model to calculate rhythm stability and generate authenticity judgment; monitoring reagent quality or liquid level changes, constructing a reagent consumption time sequence, comparing with theoretical consumption to generate consistency results; and collecting sample consistency results, operational authenticity results, and reagent consistency results to generate traceability records. This invention achieves authenticity verification and digital traceability management of the entire testing process in agricultural testing laboratories through sample physical disturbance feature identification, process time sequence reconstruction, and reagent consumption time sequence analysis.

Description

Technical Field

[0001] This invention relates to the field of digital agricultural testing technology, and in particular to a method for full-process digital traceability management of agricultural testing laboratories. Background Technology

[0002] With increasingly stringent requirements for agricultural product quality and safety supervision, agricultural testing laboratories are playing an increasingly important role in agricultural product quality testing, food safety monitoring, and agricultural environmental monitoring. To improve management efficiency and data traceability in the testing process, existing agricultural testing laboratories typically use laboratory information management systems to digitally manage sample registration, testing process records, and test results, forming a data recording chain and achieving information-based management of the sample testing process. In practical applications, corresponding operation records are generated in the system for each stage of sample reception, sorting, pretreatment, and instrument testing. These records constitute the main traceability basis for the laboratory testing process.

[0003] However, existing traceability management methods in agricultural testing laboratories primarily rely on system-recorded information for process tracking, i.e., using operation logs, equipment records, or test result records to determine whether the testing process was executed correctly. While this traceability method can achieve process recording to a certain extent, it essentially depends on data input from manual operation or automatically recorded equipment, making it difficult to directly verify the true state of samples during the testing process. For example, during sample transfer, there may be situations such as sample replacement, sample mixing, or sample reuse, and traditional system records struggle to identify these physical anomalies. Furthermore, if operators deviate from the actual testing process but still generate complete system records during experimental operations, existing traceability methods also struggle to identify such abnormal behavior.

[0004] In terms of reagent usage management, existing technologies typically only record reagent inventory in a simple manner, lacking correlation analysis between actual reagent consumption and the number of samples tested. This results in the system still generating complete test records even if testing steps are not actually performed or reagent usage is abnormal, thus affecting the authenticity and reliability of traceability data.

[0005] Therefore, how to provide a digital traceability management method for the entire process of agricultural testing laboratories is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose a digital traceability management method for the entire process of agricultural testing laboratories. This invention utilizes techniques such as sample physical disturbance feature identification, process time sequence reconstruction, and improved TimeMixer model analysis to jointly analyze and comprehensively determine the sample flow process, experimental operation behavior, and reagent consumption process, thereby achieving authenticity verification and digital traceability management of the entire testing process in agricultural testing laboratories. By performing consistency comparison of sample physical disturbance features, performing rhythmic rearrangement analysis of the experimental process time sequence, and performing correlation verification of the reagent consumption time sequence, this invention effectively solves the problem that existing traceability methods for agricultural testing laboratories rely solely on system records, making it difficult to verify the authenticity of testing behavior. It boasts advantages such as high traceability reliability, strong verifiability of the testing process, and a high degree of intelligent experimental management.

[0007] A method for full-process digital traceability management of agricultural testing laboratories according to an embodiment of the present invention includes: The samples entering the agricultural testing laboratory are registered to generate sample identification information, mechanical excitation is applied and response signals are collected, and the physical disturbance characteristics of the samples are extracted to generate initial disturbance characteristic data as the reference characteristics for sample storage. When a sample is transferred to the sampling node, pre-processing node, or on-machine testing node, the sample entry reference features are retrieved, the disturbance features of the current sample are obtained and the consistency is compared, sample consistency judgment information is generated, and the sample transfer node record is updated. A process timing reconstructor is constructed, which receives experimental operation logs, equipment operation records, detection process step definition information and sample transfer node records, performs primitive denoising, calibration, and beat reconstruction processing on operation events, and generates experimental process timing sequences. The experimental process time sequence is input into the improved TimeMixer model to perform process rhythm rearrangement, generate process rhythm vector and calculate process rhythm stability, and generate experimental operation behavior authenticity judgment information. During the testing process, the quality or liquid level of key reagent containers is monitored to obtain change data during reagent use and construct a reagent consumption time series. Matching analysis is performed based on the number of test samples and the preset reagent consumption per sample to generate reagent consumption consistency judgment information. By comprehensively analyzing the information on sample consistency, the authenticity of experimental operations, and the consistency of reagent consumption, a digital traceability record of the entire process of agricultural testing laboratories is generated.

[0008] Optionally, the sample identification information includes sample number, sample source information, sampling time, sampling location, sample type, test items, and sample receipt time.

[0009] Optionally, the step of extracting physical disturbance features of the sample to generate initial disturbance feature data includes: The response signal acquired by the vibration acquisition device is preprocessed to remove noise and obtain a stable response signal; Calculate the vibration response frequency, amplitude change, and vibration attenuation rate based on the stable response signal; The vibration response frequency, amplitude change, and vibration attenuation rate are combined and encoded to generate initial disturbance characteristic data.

[0010] Optionally, generating sample consistency determination information and updating sample transfer node records includes: When a sample is transferred to a sampling node, pre-processing node, or on-machine testing node, the sample entry reference feature corresponding to the current sample is retrieved, and the node type information of the current transfer node is read. Apply the same mechanical excitation as when the sample was put into storage to the current test sample, collect the response signal of the current test sample, and perform noise reduction, segmentation and time alignment processing on the response signal to obtain the current node perturbation response data; The current node disturbance features are extracted based on the current node disturbance response data. The current node disturbance features include response main frequency features, response amplitude features, attenuation change features, and stage transition features. The current node disturbance features are arranged in the order of response occurrence to form a current node disturbance feature group. Based on the node type information of the current circulation node, perform node sensitivity difference calculation processing on the sample entry benchmark features and the current node disturbance feature group: According to the order of response occurrence, the sample entry reference characteristics and the current node disturbance characteristics are divided into the oscillation stage, the transition stage and the stabilization stage. The changes in the main frequency, amplitude, attenuation and the change in the stage transition position are compared for each stage to generate the difference results for each stage. Based on the node type of the current circulation node, the differences in each stage are sorted by node association to obtain the feature difference value that characterizes the degree of change in the sample state of the current node. Sample consistency determination information is generated based on the feature difference value and written into the sample transfer node record. When the feature difference value is within the preset consistency range, it is determined that the current test sample is consistent with the sample entry benchmark feature. When the feature difference value exceeds the preset consistency range, it is determined that the current test sample has an abnormal sample status.

[0011] Optionally, the generation of the experimental procedure time sequence includes: A process timing reconstructor is constructed, which consists of an operation primitive generation layer, an anchor point calibration layer, and a semantic beat generation layer. The operation primitive generation layer receives experimental operation logs, equipment operation records, detection process step definition information, and sample transfer node records, and converts the input data into standardized operation primitives. The operation primitive generation layer performs primitive denoising processing on the standardized operation primitives. The primitive denoising processing includes merging duplicate reported primitives of the same type, removing primitives with missing key fields, and suppressing high-frequency jitter primitives of the same execution carrier within the same time window, to obtain a clean primitive set. Anchor primitives are extracted from the set of cleanup primitives by the anchor calibration layer. The anchor primitives include sample scanning primitives, sealing or opening primitives, instrument method loading primitives, instrument operation start primitives, instrument operation end primitives, and result writing primitives. Based on the anchor primitives, a unified calibration process is performed on the time stamps from different sources. The unified calibration process includes using the time stamp of the sample scanning primitive as the reference time and aligning and correcting the time stamps of other primitives for the same sample to obtain a set of calibration primitives. The semantic beat generation layer performs beat merging on the calibration primitive set: The primitives of the same sample are segmented by anchor primitives as boundaries. The primitives in each segment are merged according to the event type identifier to generate a beat segment containing the start anchor, end anchor, segment duration interval and primitive category distribution in the segment. Consistency disambiguation is performed on adjacent beat segments. When the anchor point order of adjacent beat segments conflicts, the order is corrected by prioritizing the instrument start primitive and the instrument end primitive. When a beat segment is missing an end anchor point, it is truncated by a preset maximum duration interval and marked as an unclosed segment. The beat segments are aggregated according to the sample identification information to output the experimental process time sequence. The experimental process time sequence consists of beat segments arranged in sequence, and each beat segment includes the transition node identifier, the start anchor type, the end anchor type, the duration interval, and the primitive category distribution information.

[0012] Optionally, the step of generating the process rhythm vector and calculating the rhythm stability to generate experimental operation behavior authenticity determination information includes: An improved TimeMixer model is constructed, which consists of a rhythm input shaping module, an inter-segment mixing calibration module, and a stability output constraint module, wherein: The rhythm input shaping module receives the experimental procedure time sequence and divides it into multiple rhythm segments according to the operation nodes. It extracts the node interval time, node duration and node sequence encoding of each rhythm segment to form a rhythm input sequence. The inter-segment hybrid calibration module performs inter-segment hybrid calibration processing on the rhythm input sequence. The inter-segment hybrid calibration processing includes aligning adjacent rhythm segments with the start anchor point to the end anchor point as the boundary, and performing proportional calibration on the interval time and duration of adjacent rhythm segments to eliminate time scale differences from different recording sources. The stability output constraint module performs stability output constraint processing on the rhythm input sequence after inter-segment hybrid calibration to generate a process rhythm vector sequence. Calculate the rhythm change between adjacent operation nodes based on the process rhythm vector sequence: The process rhythm vectors corresponding to two adjacent nodes are aligned with their components. The interval time component, duration component, and sequence code component are compared one by one to obtain the component difference results. The component difference results are then weighted and summarized in the order of interval time priority, duration second, and sequence code last to obtain the rhythm change of adjacent operation nodes. Rhythm stability is calculated based on the rhythm variation of all adjacent operation nodes. The rhythm change of each adjacent node is truncated to limit the impact of a single abnormal node on the overall judgment. The truncated rhythm change is then aggregated to obtain the process rhythm stability. The aggregation includes calculating the center value and dispersion of the rhythm change separately, and combining the center value and dispersion to form the process rhythm stability index. The process rhythm stability index is compared with the preset stability judgment rule. When the process rhythm stability index meets the preset stability condition, a rhythm stability label is generated. When the process rhythm stability index does not meet the preset stability condition, a rhythm abnormality label is generated. Information on the authenticity of experimental operations is generated based on rhythm stability markers or rhythm abnormality markers.

[0013] Optionally, the generation of reagent consumption consistency determination information includes: Continuous monitoring of key reagent containers is conducted to obtain reagent change data recorded in chronological order. The reagent change data includes mass change data or liquid level change data of the reagent containers, and the corresponding collection time identifier is recorded for each reagent change data. A reagent change record set is constructed based on reagent change data. The difference between reagent changes in adjacent time records is calculated. The difference between the current record value and the previous record value of the reagent container is determined as the reagent consumption in the time interval. The reagent consumption record sequence is obtained by arranging the records in the order of collection time. A reagent consumption time series generation module is constructed, and node binding processing is performed on the reagent consumption record sequence. The node binding processing includes: determining the detection time interval corresponding to each test sample according to the experimental process time sequence, assigning the reagent consumption record within the detection time interval to the corresponding test sample, and generating a reagent consumption time series set that corresponds one-to-one with the test sample. The theoretical reagent consumption range is determined based on the number of samples tested and the preset reagent consumption per sample. The theoretical reagent consumption range is the reagent consumption interval obtained by multiplying the number of samples tested by the standard reagent consumption per sample. The actual total reagent consumption in the reagent consumption time series set is compared with the theoretical reagent consumption range. When the actual total reagent consumption is within the theoretical reagent consumption range, reagent consumption consistency judgment information is generated. When the actual total reagent consumption exceeds the theoretical reagent consumption range, reagent consumption anomaly judgment information is generated.

[0014] Optionally, the generation of a fully digital traceability record for the agricultural testing laboratory includes: Obtain information on sample consistency, authenticity of experimental procedures, and consistency of reagent consumption, and aggregate and associate this information according to the sample identification information. For each test sample, the sample consistency judgment information, the experimental operation behavior authenticity judgment information, and the reagent consumption consistency judgment information are combined to form the traceability judgment combination information. Perform a credibility assessment on the combined information for tracing and determining the source: Based on the results of sample consistency determination, experimental operation behavior authenticity determination, and reagent consumption consistency determination, corresponding credibility evaluation levels are assigned respectively. The credibility evaluation levels are then comprehensively weighted according to the preset weight rules to obtain the traceability credibility result of the tested sample. The traceability credibility result is compared with the preset credibility threshold. When the traceability credibility result reaches the preset credibility threshold, the test sample test process is deemed credible. When the traceability credibility result does not reach the preset credibility threshold, the test sample test process is deemed abnormal. Based on the reliability determination results of the aforementioned testing process, a full-process digital traceability record for the agricultural testing laboratory is generated and stored in association with sample identification information.

[0015] The beneficial effects of this invention are: Compared with existing agricultural testing laboratories that rely solely on laboratory information management systems for traceability, this invention extracts physical disturbance characteristics of samples upon their entry into the laboratory and establishes baseline characteristics for sample storage. These disturbance characteristics are then retrieved again and compared for consistency at key stages such as sample sorting, pretreatment, and on-machine testing. This enables continuous verification of the physical state of the tested samples, effectively avoiding issues such as sample replacement, sample confusion, or reuse. It improves the verifiability of samples throughout the entire testing process, elevating traceability management from simple information recording to a dual verification model combining information recording and physical characteristic verification.

[0016] This invention constructs a process timing reconstructor to uniformly process experimental operation logs, equipment operation records, detection process step definition information, and sample transfer node records. It converts operation data from different sources into standardized operation events and performs denoising, time calibration, and beat segment reconstruction to generate an experimental process timing sequence that reflects the true execution rhythm of the experimental process. Furthermore, it analyzes and calculates the stability of the experimental process rhythm through an improved TimeMixer model, judging the authenticity of experimental operation behavior from the perspective of process rhythm. This allows the system to identify abnormal execution of the detection process even in the presence of abnormal operations or falsified operation records, thereby improving the verifiability of the experimental operation process.

[0017] This invention also constructs a reagent consumption time series by monitoring the quality or liquid level of key reagent containers, and performs correlation analysis between reagent consumption and the number of test samples and the experimental procedure sequence. It assists in verifying the testing process from the reagent usage level, forming a multi-dimensional joint verification mechanism that integrates sample status verification, experimental operation behavior analysis, and reagent consumption consistency judgment. This effectively improves the authenticity and reliability of the full-process digital traceability management of agricultural testing laboratories, enhances the credibility of experimental data, and provides more reliable technical support for the standardized management and quality and safety supervision of agricultural testing laboratories. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a full-process digital traceability management method for agricultural testing laboratories proposed in this invention; Figure 2 This is a schematic diagram of the process flow analysis of the improved TimeMixer model for a digital traceability management method for the entire process of agricultural testing laboratories proposed in this invention. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0020] refer to Figure 1 and Figure 2 A method for full-process digital traceability management of agricultural testing laboratories, comprising: The samples entering the agricultural testing laboratory are registered to generate sample identification information, mechanical excitation is applied and response signals are collected, and the physical disturbance characteristics of the samples are extracted to generate initial disturbance characteristic data as the reference characteristics for sample storage. When a sample is transferred to the sampling node, pre-processing node, or on-machine testing node, the sample entry reference features are retrieved, the disturbance features of the current sample are obtained and the consistency is compared, sample consistency judgment information is generated, and the sample transfer node record is updated. A process timing reconstructor is constructed, which receives experimental operation logs, equipment operation records, detection process step definition information and sample transfer node records, performs primitive denoising, calibration, and beat reconstruction processing on operation events, and generates experimental process timing sequences. The experimental process time sequence is input into the improved TimeMixer model to perform process rhythm rearrangement, generate process rhythm vector and calculate process rhythm stability, and generate experimental operation behavior authenticity judgment information. During the testing process, the quality or liquid level of key reagent containers is monitored to obtain change data during reagent use and construct a reagent consumption time series. Matching analysis is performed based on the number of test samples and the preset reagent consumption per sample to generate reagent consumption consistency judgment information. By comprehensively analyzing the information on sample consistency, the authenticity of experimental operations, and the consistency of reagent consumption, a digital traceability record of the entire process of agricultural testing laboratories is generated.

[0021] In this embodiment, the sample identification information includes sample number, sample source information, sampling time, sampling location, sample type, testing items, and sample receiving time.

[0022] In this embodiment, the step of extracting physical disturbance features of the sample to generate initial disturbance feature data includes: The response signal acquired by the vibration acquisition device is preprocessed to remove noise and obtain a stable response signal; Calculate the vibration response frequency, amplitude change, and vibration attenuation rate based on the stable response signal; The vibration response frequency, amplitude change, and vibration attenuation rate are combined and encoded to generate initial disturbance characteristic data.

[0023] In this embodiment, generating sample consistency determination information and updating sample transfer node records includes: When a sample is transferred to a sampling node, pre-processing node, or on-machine testing node, the sample entry reference feature corresponding to the current sample is retrieved, and the node type information of the current transfer node is read. Apply the same mechanical excitation as when the sample was put into storage to the current test sample, collect the response signal of the current test sample, and perform noise reduction, segmentation and time alignment processing on the response signal to obtain the current node perturbation response data; The current node disturbance features are extracted based on the current node disturbance response data. The current node disturbance features include response main frequency features, response amplitude features, attenuation change features, and stage transition features. The current node disturbance features are arranged in the order of response occurrence to form a current node disturbance feature group. Based on the node type information of the current circulation node, perform node sensitivity difference calculation processing on the sample entry benchmark features and the current node disturbance feature group: Based on the order of response, the sample's baseline characteristics and the current node's disturbance characteristics are divided into three stages: oscillation initiation, transition, and stabilization. The changes in dominant frequency, amplitude, attenuation, and transition position are compared for each stage to generate differences between them. The specific division into oscillation initiation, transition, and stabilization stages is as follows: When the response signal changes from an initial static state to a continuously rising state, and the rate of change of amplitude is continuously greater than the preset oscillation threshold, the corresponding characteristic component is divided into the oscillation stage. When the amplitude change rate of the response signal changes from a continuous increase to a fluctuating decrease, and the main frequency change rate is still greater than the preset stable threshold, the corresponding feature component is classified as a transition phase. When the rate of change of amplitude and the rate of change of main frequency of the response signal both continuously fall within the preset stable range, the corresponding characteristic components are classified as stable phases. The sample entry reference features are classified into corresponding categories with the feature components in the same stage of the current node disturbance feature group to form the corresponding division results of the oscillation stage, transition stage and stable stage. Based on the node type of the current circulation node, the differences in each stage are sorted by node association to obtain the feature difference value that characterizes the degree of change in the sample state of the current node. Sample consistency determination information is generated based on feature difference values ​​and written into the sample transfer node record. When the feature difference value is within a preset consistency range, the current tested sample is determined to be consistent with the sample entry benchmark characteristics. When the feature difference value exceeds the preset consistency range, the current tested sample is determined to have an abnormal sample status. Specifically, the sample consistency determination information is generated based on feature difference values ​​as follows: The system acquires the characteristic difference values ​​of the current test sample and classifies the degree of difference according to the position of the characteristic difference values ​​in the preset consistency interval. When the characteristic difference value is inside the consistency interval, it is determined to be of the consistency level; when the characteristic difference value is close to the boundary of the consistency interval, it is determined to be of the deviation level; and when the characteristic difference value exceeds the consistency interval, it is determined to be of the abnormal level. Based on the determined difference level, a corresponding status label is generated, where the consistency level corresponds to the generation of a sample consistency label, the deviation level corresponds to the generation of a sample deviation label, and the anomaly level corresponds to the generation of a sample anomaly label. The status identifier is combined with the sample identifier information of the currently tested sample and the corresponding testing node information to form sample consistency determination information.

[0024] In this embodiment, generating the experimental procedure time sequence includes: A process timing reconstructor is constructed, which consists of an operation primitive generation layer, an anchor point calibration layer, and a semantic beat generation layer. The operation primitive generation layer receives experimental operation logs, equipment operation records, detection process step definition information, and sample transfer node records, and converts the input data into standardized operation primitives. The operation primitive generation layer performs primitive denoising processing on the standardized operation primitives. The primitive denoising processing includes merging duplicate reported primitives of the same type, removing primitives with missing key fields, and suppressing high-frequency jitter primitives of the same execution carrier within the same time window, to obtain a clean primitive set. Anchor primitives are extracted from the set of cleanup primitives by the anchor calibration layer. The anchor primitives include sample scanning primitives, sealing or opening primitives, instrument method loading primitives, instrument operation start primitives, instrument operation end primitives, and result writing primitives. Based on the anchor primitives, a unified calibration process is performed on the time stamps from different sources. The unified calibration process includes using the time stamp of the sample scanning primitive as the reference time and aligning and correcting the time stamps of other primitives for the same sample to obtain a set of calibration primitives. The semantic beat generation layer performs beat merging on the calibration primitive set: The primitives of the same sample are segmented by anchor primitives as boundaries. The primitives in each segment are merged according to the event type identifier to generate a beat segment containing the start anchor, end anchor, segment duration interval and primitive category distribution in the segment. Consistency disambiguation is performed on adjacent beat segments. When the anchor point order of adjacent beat segments conflicts, the order is corrected by prioritizing the instrument start primitive and the instrument end primitive. When a beat segment is missing an end anchor point, it is truncated by a preset maximum duration interval and marked as an unclosed segment. The beat segments are aggregated according to the sample identification information to output the experimental process time sequence. The experimental process time sequence consists of beat segments arranged in sequence, and each beat segment includes the transition node identifier, the start anchor type, the end anchor type, the duration interval, and the primitive category distribution information.

[0025] In this embodiment, the step of generating the process rhythm vector and calculating the rhythm stability to generate experimental operation behavior authenticity determination information includes: An improved TimeMixer model is constructed, which consists of a rhythm input shaping module, an inter-segment mixing calibration module, and a stability output constraint module, wherein: The rhythm input shaping module receives the experimental procedure time sequence and divides it into multiple rhythm segments according to the operation nodes. It extracts the node interval time, node duration and node sequence encoding of each rhythm segment to form a rhythm input sequence. The inter-segment hybrid calibration module performs inter-segment hybrid calibration processing on the rhythm input sequence. The inter-segment hybrid calibration processing includes aligning adjacent rhythm segments with the start anchor point to the end anchor point as the boundary, and performing proportional calibration on the interval time and duration of adjacent rhythm segments to eliminate time scale differences from different recording sources. The stability output constraint module performs stability output constraint processing on the rhythm input sequence after inter-segment hybrid calibration, generating a process rhythm vector sequence. Specifically, the stability output constraint processing on the rhythm input sequence after inter-segment hybrid calibration involves: The rhythm input sequence after inter-segment hybrid calibration is subjected to node-level constraint screening. The interval time, duration and sequence code of each node are checked one by one according to the order of operation nodes to see if they fall within the preset rhythm fluctuation range. The node components that exceed the preset rhythm fluctuation range are truncated and restricted, while the node components that are within the preset rhythm fluctuation range are retained to obtain the node-constrained rhythm sequence. The rhythm sequence after node constraint is smoothed by adjacent nodes. According to the succession relationship between adjacent operation nodes, the interval time component, duration component and sequential coding component corresponding to the previous node and the next node are continuously adjusted to keep the rhythm change between adjacent nodes within the preset continuous interval, so as to obtain the smoothed rhythm sequence. The smoothed rhythm sequence is output-defined. The interval time component, duration component and sequence code component corresponding to each operation node are combined in a unified order and output sequentially according to the node arrangement order of the sample in the detection process to generate a process rhythm vector sequence. Calculate the rhythm change between adjacent operation nodes based on the process rhythm vector sequence: The process rhythm vectors corresponding to two adjacent nodes are aligned with their components. The interval time component, duration component, and sequence code component are compared one by one to obtain the component difference results. The component difference results are then weighted and summarized in the order of interval time priority, duration second, and sequence code last to obtain the rhythm change of adjacent operation nodes. Rhythm stability is calculated based on the rhythm variation of all adjacent operation nodes. The rhythm changes of each adjacent node are truncated to limit the impact of a single abnormal node on the overall judgment. The truncated rhythm changes are then aggregated to obtain the process rhythm stability. This aggregation includes calculating the center value and dispersion of the rhythm changes separately, and combining the center value and dispersion to form a process rhythm stability index. Center value: Sort all adjacent node rhythm changes after truncation according to their numerical values, and select the rhythm change in the middle position after sorting as the center value; when the number of rhythm changes after truncation is even, take the average of the two middle rhythm changes as the center value. Dispersion: Using the center value as a benchmark, the deviation between the rhythm change of each adjacent node and the center value is calculated, and the average of all deviations is taken to obtain the dispersion. The process rhythm stability index is compared with a preset stability judgment rule. A rhythm stability flag is generated when the process rhythm stability index meets the preset stability conditions; a rhythm anomaly flag is generated when the process rhythm stability index does not meet the preset stability conditions. The preset stability judgment rule is as follows: Using the distribution of process rhythm stability index corresponding to the historical normal testing process of the agricultural testing laboratory as a benchmark, the allowable fluctuation range of the process rhythm stability index is predetermined. When the current process rhythm stability index falls within the allowable fluctuation range, it is determined that the preset stability condition is met. When the current process rhythm stability index exceeds the allowable fluctuation range, it is determined that the preset stability condition is not met. The predetermined allowable fluctuation range of the process rhythm stability index is 0.20 to 0.80. When the process rhythm stability index falls within the range of 0.20 to 0.80, it is determined that the preset stability condition is met. When the process rhythm stability index is less than 0.20 or greater than 0.80, it is determined that the preset stability condition is not met. The authenticity assessment information for experimental procedures is generated based on rhythm stability or rhythm abnormality indicators. Specifically, the generation of this authenticity assessment information involves: Obtain rhythm stability or rhythm abnormality indicators, and use these indicators as the initial judgment results for the experimental operation behavior of the current test sample; Based on the completeness of the execution of nodes in the experimental procedure time sequence of the corresponding sample, the consistency of the initial judgment result is verified. When the execution order of the node corresponding to the rhythm stability indicator is complete and without any missing nodes, a true experimental operation behavior indicator is generated. When the execution order of the node corresponding to the rhythm stability indicator is missing or abnormal, an experimental operation behavior deviation indicator is generated. When the rhythm abnormality indicator appears, an experimental operation behavior abnormality indicator is directly generated. The true experimental operation behavior indicator, experimental operation behavior deviation indicator, or experimental operation behavior abnormality indicator are output as the experimental operation behavior authenticity judgment information.

[0026] In this embodiment, the generation of reagent consumption consistency determination information includes: Continuous monitoring of key reagent containers is conducted to obtain reagent change data recorded in chronological order. The reagent change data includes mass change data or liquid level change data of the reagent containers, and the corresponding collection time identifier is recorded for each reagent change data. A reagent change record set is constructed based on reagent change data. The difference between reagent changes in adjacent time records is calculated. The difference between the current record value and the previous record value of the reagent container is determined as the reagent consumption in the time interval. The reagent consumption record sequence is obtained by arranging the records in the order of collection time. A reagent consumption time series generation module is constructed, and node binding processing is performed on the reagent consumption record sequence. The node binding processing includes: determining the detection time interval corresponding to each test sample according to the experimental process time sequence, assigning the reagent consumption record within the detection time interval to the corresponding test sample, and generating a reagent consumption time series set that corresponds one-to-one with the test sample. The theoretical reagent consumption range is determined based on the number of samples tested and the preset reagent consumption per sample. The theoretical reagent consumption range is the reagent consumption interval obtained by multiplying the number of samples tested by the standard reagent consumption per sample. The actual total reagent consumption in the reagent consumption time series set is compared with the theoretical reagent consumption range. When the actual total reagent consumption is within the theoretical reagent consumption range, reagent consumption consistency judgment information is generated. When the actual total reagent consumption exceeds the theoretical reagent consumption range, reagent consumption anomaly judgment information is generated.

[0027] In this embodiment, the generation of a full-process digital traceability record for agricultural testing laboratories includes: Obtain information on sample consistency, authenticity of experimental procedures, and consistency of reagent consumption, and aggregate and associate this information according to the sample identification information. For each test sample, the sample consistency judgment information, the experimental operation behavior authenticity judgment information, and the reagent consumption consistency judgment information are combined to form the traceability judgment combination information. Perform a credibility assessment on the combined information for tracing and determining the source: Based on the results of sample consistency assessment, experimental operation authenticity assessment, and reagent consumption consistency assessment, corresponding reliability evaluation levels are assigned. These reliability evaluation levels are then comprehensively weighted according to a preset weighting rule to obtain the traceability reliability result for the tested sample. The preset weighting rule is as follows: Based on the different degrees of influence of each judgment result in the testing process on the authenticity of traceability, weights are assigned to the sample consistency judgment result, the experimental operation behavior authenticity judgment result, and the reagent consumption consistency judgment result. The weight of the sample consistency judgment result is higher than the weight of the experimental operation behavior authenticity judgment result, and the weight of the experimental operation behavior authenticity judgment result is higher than the weight of the reagent consumption consistency judgment result. The sum of the weights of the three is guaranteed to be a fixed value. The credibility rating is a graded evaluation value set according to the sample consistency judgment result, the experimental operation behavior authenticity judgment result, and the reagent consumption consistency judgment result respectively. The credibility rating is 1 for the consistency state, 0.5 for the deviation state, and 0 for the abnormal state. The traceability credibility result is compared with the preset credibility threshold. When the traceability credibility result reaches the preset credibility threshold, the test sample test process is deemed credible. When the traceability credibility result does not reach the preset credibility threshold, the test sample test process is deemed abnormal. The preset credibility threshold is 0.75. Based on the reliability determination results of the aforementioned testing process, a full-process digital traceability record for the agricultural testing laboratory is generated and stored in association with sample identification information.

[0028] Example 1: To verify the feasibility of this invention in practice, it was applied to the pesticide residue testing laboratory of the Provincial Agricultural Product Quality and Safety Testing Center. This laboratory mainly undertakes pesticide residue testing of vegetable, fruit, and grain samples. The laboratory processes approximately 80-120 samples daily. The testing process includes sample reception and registration, sample sorting, sample pretreatment, instrument testing, and result recording. The laboratory's original management relied primarily on a laboratory information management system to record sample registration information, equipment operation records, and test result data. However, in actual operation, it was found that relying solely on system records for traceability has significant shortcomings. For example, sample confusion or substitution may occur during sample transfer, yet the system records may remain intact; if the testing procedure is not strictly followed during experimental operations but an operation log is still generated, it is difficult to detect through traditional systems; furthermore, reagent management typically only records inventory changes without correlating reagent consumption with the number of samples tested, thus failing to effectively identify abnormal reagent usage.

[0029] After deploying the method of this invention, the laboratory first sets up a micro-perturbation acquisition platform and a vibration acquisition device in the sample receiving area. When a sample enters the laboratory, the system automatically generates sample identification information and records information such as sample number, sample source, sampling time, sampling location, sample type, and testing items. Subsequently, a slight mechanical excitation is applied to the sample container through the micro-perturbation acquisition platform, and the response signal is acquired. After preprocessing and denoising the acquired response signal, the vibration response frequency, amplitude change, and vibration attenuation rate are calculated. These features are combined and encoded to form initial perturbation feature data, which is stored as the reference feature for sample storage. Since different samples differ in container state, internal density, and structure, the resulting perturbation features can reflect the physical state of the sample relatively stably.

[0030] As samples move to the sample sorting node, pre-processing node, and on-machine testing node, the system again acquires the response signal of the current sample through the micro-perturbation acquisition platform and extracts the perturbation features of the current node. This feature is then compared with the sample entry baseline features to calculate node sensitivity differences, thereby determining whether the sample has been replaced or confused during the transfer process. Simultaneously, the system obtains experimental operation logs, equipment operation records, and testing process step definitions from the laboratory information management system, and uses a process timing reconstruction device to uniformly process data from different sources. The process timing reconstruction device converts the operation logs into standardized operation primitives, eliminates duplicate records, abnormal records, and time jitter records through primitive denoising, and then uses an anchor point calibration method to unify the time signatures of different data sources, thereby generating an experimental process timing sequence that accurately reflects the execution of the experimental process.

[0031] After the experimental procedure time sequence is generated, the system inputs it into an improved TimeMixer model for process rhythm analysis. It calculates the process rhythm vector by encoding the interval time of operation nodes, the duration of nodes, and the sequence of nodes, and then calculates the process rhythm stability. When the rhythm stability of the detection process is significantly lower than the normal range, the system can identify experimental operation anomalies, such as missing steps, incorrect step sequence, or abnormal operation time. During the detection process, the system monitors the quality of key reagent containers, constructing a reagent consumption time series by continuously collecting data on changes in container quality. Based on the number of samples tested and the preset reagent consumption per sample, the theoretical reagent consumption range is calculated to determine whether the reagent consumption matches the detection task. During the experiment, the laboratory completed the testing of 642 agricultural product samples.

[0032] Table 1. Statistical Table of Application of Full-Process Traceability Management Method in Agricultural Testing Laboratories

[0033] As shown in Table 1, after applying the full-process digital traceability management method for agricultural testing laboratories proposed in this invention, the laboratory exhibited a high level of consistency in sample transfer verification. The number of samples in the seven testing tasks ranged from 87 to 96, which is consistent with the daily sample processing scale of agricultural testing laboratories. The consistency rate of sample perturbation characteristics remained between 98.7% and 99.2%, indicating that the physical characteristics of samples acquired through the micro-perturbation acquisition platform have good stability across different transfer nodes. Most samples maintained consistency with the entry-level characteristics during the sampling, pretreatment, and testing processes, effectively identifying potential sample confusion or substitution issues.

[0034] Regarding the experimental procedure analysis, the stability of the procedure rhythm for each group of detection tasks ranged from 0.92 to 0.95, with relatively small overall fluctuations. This indicates that after the experimental operation logs, equipment operation records, and detection procedure step information are uniformly processed by the procedure timing reconstructor, and then combined with the rhythm analysis method, the execution status of the experimental procedure can be reflected relatively accurately. When the experimental operation is performed according to the standard steps, the procedure rhythm stability remains at a high level. However, when individual operations have time anomalies or sequence deviations, the system can promptly identify and record the anomalies.

[0035] Regarding reagent usage monitoring, the reagent consumption matching rate remained between 97.5% and 98.4%, indicating that the actual reagent consumption was basically consistent with the theoretical consumption. By continuously monitoring key reagent containers and correlating this with the number of samples tested, it is possible to effectively determine whether reagent usage meets the requirements of the testing task. The table shows that four abnormal events were identified across the seven testing tasks, including abnormal sample disturbance characteristics and abnormal process rhythms. This demonstrates that the method of this invention can effectively monitor the authenticity of the testing process without affecting normal laboratory testing operations, thereby improving the reliability of traceability management in agricultural testing laboratories.

[0036] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for full-process digital traceability management of agricultural testing laboratories, characterized in that, include: The samples entering the agricultural testing laboratory are registered to generate sample identification information, mechanical excitation is applied and response signals are collected, and the physical disturbance characteristics of the samples are extracted to generate initial disturbance characteristic data as the reference characteristics for sample storage. When a sample is transferred to the sampling node, pre-processing node, or on-machine testing node, the sample entry reference features are retrieved, the disturbance features of the current sample are obtained and the consistency is compared, sample consistency judgment information is generated, and the sample transfer node record is updated. A process timing reconstructor is constructed, which receives experimental operation logs, equipment operation records, detection process step definition information and sample transfer node records, performs primitive denoising, calibration, and beat reconstruction processing on operation events, and generates experimental process timing sequences. The experimental process time sequence is input into the improved TimeMixer model to perform process rhythm rearrangement, generate process rhythm vector and calculate process rhythm stability, and generate experimental operation behavior authenticity judgment information. During the testing process, the quality or liquid level of key reagent containers is monitored to obtain change data during reagent use and construct a reagent consumption time series. Matching analysis is performed based on the number of test samples and the preset reagent consumption per sample to generate reagent consumption consistency judgment information. By comprehensively analyzing the information on sample consistency, the authenticity of experimental operations, and the consistency of reagent consumption, a digital traceability record of the entire process of agricultural testing laboratories is generated.

2. The method for full-process digital traceability management of agricultural testing laboratories according to claim 1, characterized in that, The sample identification information includes sample number, sample source information, sampling time, sampling location, sample type, test items, and sample receipt time.

3. The method for full-process digital traceability management of agricultural testing laboratories according to claim 1, characterized in that, The process of extracting physical disturbance features from the sample to generate initial disturbance feature data includes: The response signal acquired by the vibration acquisition device is preprocessed to remove noise and obtain a stable response signal; Calculate the vibration response frequency, amplitude change, and vibration attenuation rate based on the stable response signal; The vibration response frequency, amplitude change, and vibration attenuation rate are combined and encoded to generate initial disturbance characteristic data.

4. The method for full-process digital traceability management of agricultural testing laboratories according to claim 1, characterized in that, The process of generating sample consistency determination information and updating sample transfer node records includes: When a sample is transferred to a sampling node, pre-processing node, or on-machine testing node, the sample entry reference feature corresponding to the current sample is retrieved, and the node type information of the current transfer node is read. Apply the same mechanical excitation as when the sample was put into storage to the current test sample, collect the response signal of the current test sample, and perform noise reduction, segmentation and time alignment processing on the response signal to obtain the current node perturbation response data; The current node disturbance features are extracted based on the current node disturbance response data. The current node disturbance features include response main frequency features, response amplitude features, attenuation change features, and stage transition features. The current node disturbance features are arranged in the order of response occurrence to form a current node disturbance feature group. Based on the node type information of the current circulation node, perform node sensitivity difference calculation processing on the sample entry benchmark features and the current node disturbance feature group: According to the order of response occurrence, the sample entry reference characteristics and the current node disturbance characteristics are divided into the oscillation stage, the transition stage and the stabilization stage. The changes in the main frequency, amplitude, attenuation and the change in the stage transition position are compared for each stage to generate the difference results for each stage. Based on the node type of the current circulation node, the differences in each stage are sorted by node association to obtain the feature difference value that characterizes the degree of change in the sample state of the current node. Sample consistency determination information is generated based on the feature difference value and written into the sample transfer node record. When the feature difference value is within the preset consistency range, it is determined that the current test sample is consistent with the sample entry benchmark feature. When the feature difference value exceeds the preset consistency range, it is determined that the current test sample has an abnormal sample status.

5. The method for full-process digital traceability management of agricultural testing laboratories according to claim 1, characterized in that, The generated experimental procedure time sequence includes: A process timing reconstructor is constructed, which consists of an operation primitive generation layer, an anchor point calibration layer, and a semantic beat generation layer. The operation primitive generation layer receives experimental operation logs, equipment operation records, detection process step definition information, and sample transfer node records, and converts the input data into standardized operation primitives. The operation primitive generation layer performs primitive denoising processing on the standardized operation primitives. The primitive denoising processing includes merging duplicate reported primitives of the same type, removing primitives with missing key fields, and suppressing high-frequency jitter primitives of the same execution carrier within the same time window, to obtain a clean primitive set. Anchor primitives are extracted from the set of cleanup primitives by the anchor calibration layer. The anchor primitives include sample scanning primitives, sealing or opening primitives, instrument method loading primitives, instrument operation start primitives, instrument operation end primitives, and result writing primitives. Based on the anchor primitives, a unified calibration process is performed on the time stamps from different sources. The unified calibration process includes using the time stamp of the sample scanning primitive as the reference time and aligning and correcting the time stamps of other primitives for the same sample to obtain a set of calibration primitives. The semantic beat generation layer performs beat merging on the calibration primitive set: The primitives of the same sample are segmented by anchor primitives as boundaries. The primitives in each segment are merged according to the event type identifier to generate a beat segment containing the start anchor, end anchor, segment duration interval and primitive category distribution in the segment. Consistency disambiguation is performed on adjacent beat segments. When the anchor point order of adjacent beat segments conflicts, the order is corrected by prioritizing the instrument start primitive and the instrument end primitive. When a beat segment is missing an end anchor point, it is truncated by a preset maximum duration interval and marked as an unclosed segment. The beat segments are aggregated according to the sample identification information to output the experimental process time sequence. The experimental process time sequence consists of beat segments arranged in sequence, and each beat segment includes the transition node identifier, the start anchor type, the end anchor type, the duration interval, and the primitive category distribution information.

6. The method for full-process digital traceability management of agricultural testing laboratories according to claim 1, characterized in that, The process of generating a rhythm vector and calculating rhythm stability to generate experimental operation behavior authenticity determination information includes: An improved TimeMixer model is constructed, which consists of a rhythm input shaping module, an inter-segment mixing calibration module, and a stability output constraint module, wherein: The rhythm input shaping module receives the experimental procedure time sequence and divides it into multiple rhythm segments according to the operation nodes. It extracts the node interval time, node duration and node sequence encoding of each rhythm segment to form a rhythm input sequence. The inter-segment hybrid calibration module performs inter-segment hybrid calibration processing on the rhythm input sequence. The inter-segment hybrid calibration processing includes aligning adjacent rhythm segments with the start anchor point to the end anchor point as the boundary, and performing proportional calibration on the interval time and duration of adjacent rhythm segments to eliminate time scale differences from different recording sources. The stability output constraint module performs stability output constraint processing on the rhythm input sequence after inter-segment hybrid calibration to generate a process rhythm vector sequence. Calculate the rhythm change between adjacent operation nodes based on the process rhythm vector sequence: The process rhythm vectors corresponding to two adjacent nodes are aligned with their components. The interval time component, duration component, and sequence code component are compared one by one to obtain the component difference results. The component difference results are then weighted and summarized in the order of interval time priority, duration second, and sequence code last to obtain the rhythm change of adjacent operation nodes. The process for calculating rhythm stability based on rhythm changes across all adjacent operation nodes is as follows: The rhythm change of each adjacent node is truncated to limit the impact of a single abnormal node on the overall judgment. The truncated rhythm change is then aggregated to obtain the process rhythm stability. The aggregation includes calculating the center value and dispersion of the rhythm change separately, and combining the center value and dispersion to form the process rhythm stability index. The process rhythm stability index is compared with the preset stability judgment rule. When the process rhythm stability index meets the preset stability condition, a rhythm stability label is generated. When the process rhythm stability index does not meet the preset stability condition, a rhythm abnormality label is generated. Information on the authenticity of experimental operations is generated based on rhythm stability markers or rhythm abnormality markers.

7. The method for full-process digital traceability management of agricultural testing laboratories according to claim 1, characterized in that, The consistency determination information for generated reagent consumption includes: Continuous monitoring of key reagent containers is conducted to obtain reagent change data recorded in chronological order. The reagent change data includes mass change data or liquid level change data of the reagent containers, and the corresponding collection time identifier is recorded for each reagent change data. A reagent change record set is constructed based on reagent change data. The difference between reagent changes in adjacent time records is calculated. The difference between the current record value and the previous record value of the reagent container is determined as the reagent consumption in the time interval. The reagent consumption record sequence is obtained by arranging the records in the order of collection time. A reagent consumption time series generation module is constructed, and node binding processing is performed on the reagent consumption record sequence. The node binding processing includes: determining the detection time interval corresponding to each test sample according to the experimental process time sequence, assigning the reagent consumption record within the detection time interval to the corresponding test sample, and generating a reagent consumption time series set that corresponds one-to-one with the test sample. The theoretical reagent consumption range is determined based on the number of samples tested and the preset reagent consumption per sample. The theoretical reagent consumption range is the reagent consumption interval obtained by multiplying the number of samples tested by the standard reagent consumption per sample. The actual total reagent consumption in the reagent consumption time series set is compared with the theoretical reagent consumption range. When the actual total reagent consumption is within the theoretical reagent consumption range, reagent consumption consistency judgment information is generated. When the actual total reagent consumption exceeds the theoretical reagent consumption range, reagent consumption anomaly judgment information is generated.

8. The method for full-process digital traceability management of agricultural testing laboratories according to claim 1, characterized in that, The generated agricultural testing laboratory's full-process digital traceability record includes: Obtain information on sample consistency, authenticity of experimental procedures, and consistency of reagent consumption, and aggregate and associate this information according to the sample identification information. For each test sample, the sample consistency judgment information, the experimental operation behavior authenticity judgment information, and the reagent consumption consistency judgment information are combined to form the traceability judgment combination information. Perform a credibility assessment on the combined information for tracing and determining the source: Based on the results of sample consistency determination, experimental operation behavior authenticity determination, and reagent consumption consistency determination, corresponding credibility evaluation levels are assigned respectively. The credibility evaluation levels are then comprehensively weighted according to the preset weight rules to obtain the traceability credibility result of the tested sample. The traceability credibility result is compared with the preset credibility threshold. When the traceability credibility result reaches the preset credibility threshold, the test sample test process is deemed credible. When the traceability credibility result does not reach the preset credibility threshold, the test sample test process is deemed abnormal. Based on the reliability determination results of the aforementioned testing process, a full-process digital traceability record for the agricultural testing laboratory is generated and stored in association with sample identification information.

Images