CN-122020206-A - Simulation degree digital evaluation method and system for microbial protein meat product

CN122020206ACN 122020206 ACN122020206 ACN 122020206ACN-122020206-A

Abstract

The invention provides a simulation degree digital evaluation method and system of a microbial protein meat product, and relates to the technical field of digital evaluation, wherein the method comprises the following steps of 1, collecting physical and mechanical data, characteristic volatile organic matter chromatographic data, multichannel bionic sensing array response data and structured sensory evaluation data of the microbial protein meat product to be tested in parallel; preprocessing each mode data respectively, and extracting key quantization indexes under the respective modes to form a quality characterization vector. The invention realizes accurate, objective and standardized digital evaluation of the simulation degree, improves the accuracy and normalization of the evaluation result, and fills the defect that the evaluation method ignores the mutual influence among indexes.

Inventors

WANG HAITANG
CHEN XI
LI YINGYING
WANG SHOUWEI
ZOU HAO
LI XIANG
XIONG SUYUE
ZHAO YAN
QU CHAO

Assignees

中国肉类食品综合研究中心

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (10)

1. The method for digitally evaluating the simulation degree of the microbial protein meat product is characterized by comprising the following steps of: step 1, collecting physical and mechanical data, characteristic volatile organic matter chromatographic data, multichannel bionic sensing array response data and structured sensory evaluation data of a to-be-detected microbial protein meat product in parallel, respectively preprocessing each mode data, extracting key quantization indexes under each mode, and forming a quality characterization vector; step 2, respectively calculating information entropy weights of key quantization indexes of all modes, calculating conflict weights between all indexes and the rest indexes, combining the two to generate initial static weights of all indexes; Step 3, simulating an initial interaction matrix as a stress interference field of a multi-crack system in the material mechanics, wherein each index is regarded as a crack source, and the interaction intensity among indexes is simulated as stress interference intensity; And 4, carrying out weighted fusion on the quality characterization vector by using the corrected dynamic weight to generate a comprehensive feature expression vector, inputting the comprehensive feature expression vector into a pre-trained similarity evaluation model, and mapping and outputting quantized simulation degree scores.
2. The method for digitally evaluating the simulation degree of a microbial protein meat product according to claim 1, wherein the preprocessing is performed on each mode data, and key quantization indexes under each mode are extracted to form a quality characterization vector, and the method comprises the steps of: The method comprises the steps of cleaning and standardizing physical and mechanical data, extracting key quantitative indexes representing the quality and characteristics of the microbial protein meat product from the processed data to form physical index vectors, carrying out baseline correction, peak identification and normalization on characteristic volatile organic matter chromatographic data to form flavor substance abundance vectors, carrying out noise reduction and normalization on multichannel bionic sensing array response data, extracting steady state response values or characteristic response modes from response curves of all channels to form sensor response mode vectors, carrying out consistency test and standardization on structural sensory evaluation data, and summarizing scores of all evaluation dimensions to form sensory score vectors; And combining the physical index vector, the flavor substance abundance vector, the sensor response mode vector and the sensory score vector to form a quality characterization vector.
3. The method for digitized evaluation of the simulation degree of a microbial protein meat product according to claim 2, wherein the steps of calculating the entropy weight of each mode key quantization index, calculating the conflict weight between each index and the rest index, and combining the two to generate the initial static weight of each index, respectively, comprise: For all key quantization indexes in the quality characterization vector, calculating the index value duty ratio of each sample under each index and the information entropy value of the corresponding index, and determining the information entropy weight of the corresponding index according to the information entropy value; For each index, calculating the correlation coefficient between each index and the rest indexes respectively, and taking the sum of absolute values of the phase relation numbers as the conflict weight of the corresponding index; combining the information entropy weight and the conflict weight through multiplication synthesis or weighted summation to obtain the initial static weight of each index.
4. The method for digitally evaluating the simulation degree of a microbial protein meat product according to claim 3, wherein the pairing of the index vectors of different modes, analyzing the coupling relationship between any two index data sequences, and constructing an initial interaction matrix reflecting the strength of interaction between cross-mode indexes comprises: Regarding each index in the quality characterization vector as a node, extracting standardized data sequences of the corresponding index in a plurality of samples for any two indexes, and calculating a correlation metric value between the two sequences; Setting the interaction intensity between all indexes as n x n matrix form according to the index sequence, where n is the total number of indexes, and constructing to obtain initial interaction matrix.
5. The method for digitally evaluating the simulation degree of a microbial protein meat product according to claim 4, wherein the step 3 comprises: each element in the initial interaction matrix is regarded as a stress interference coefficient between two corresponding indexes, and each key quantization index is regarded as a crack source; For each crack source, calculating a net interference effect value of the corresponding crack source under the combined action of all the remaining crack sources according to the stress interference coefficients between the corresponding crack source and all the remaining crack sources in the initial interaction matrix; And correcting the initial static weight of the corresponding index according to the net interference effect value of each index, increasing the initial static weight when the net interference effect value is the enhancement effect, and reducing the initial static weight when the net interference effect value is the weakening effect, so as to generate the corrected dynamic weight of each index.
6. The method of claim 5, wherein for each crack source, calculating a net interference effect value of the corresponding crack source under the combined action of all remaining crack sources according to the stress interference coefficients between the corresponding crack source and all remaining crack sources in the initial interaction matrix, comprising: taking each off-diagonal element in the initial interaction matrix as a stress interference coefficient between two corresponding crack sources; and for the ith crack source, extracting stress interference coefficients between the corresponding crack source and all the residual crack sources from the initial interaction matrix, carrying out accumulated summation on all the stress interference coefficients, and taking the summation result as a net interference effect value of the ith crack source under the combined action of all the residual crack sources.
7. The method of digitized evaluation of the simulation of a meat product of a microbial protein of claim 6 wherein modifying the initial static weights of the corresponding indicators based on the net interference effect value of each indicator, increasing the initial static weights when the net interference effect value is an enhancement effect, decreasing the initial static weights when the net interference effect value is a de-enhancement effect, generating a modified dynamic weight for each indicator, comprises: comparing the net interference effect value of each index with zero, wherein if the net interference effect value is greater than zero, the net interference effect value is an enhancement effect, and if the net interference effect value is less than zero, the net interference effect value is a weakening effect; For the index with the net interference effect value being more than zero, the initial static weight of the corresponding index is multiplied by an enhancement coefficient being more than 1 to obtain the dynamic weight after the corresponding index is corrected, and for the index with the net interference effect value being less than zero, the initial static weight of the corresponding index is multiplied by an attenuation coefficient being less than 1 to obtain the dynamic weight after the corresponding index is corrected.
8. The method for digitally evaluating the simulation degree of a microbial protein meat product according to claim 7, wherein the step 4 comprises: The corrected dynamic weights are respectively weighted and multiplied with the physical index vector, the flavor substance abundance vector, the sensor response mode vector and the key quantization indexes corresponding to the sensory score vector contained in the quality characterization vector, and the weighted vectors are spliced or summed to generate a comprehensive feature expression vector; Inputting the comprehensive feature expression vector into a pre-trained similarity evaluation model, wherein the similarity evaluation model calculates the distance or similarity between the corresponding comprehensive feature expression vector and the comprehensive feature expression vector of the standard meat product sample in a feature space; and mapping the distance or the similarity to a quantized simulation degree score in the range of 0 to 100 percent, and taking the quantized simulation degree score as a final simulation degree evaluation result of the microbial protein meat product to be tested.
9. A system for digitally evaluating the simulation of a meat product of a microbial protein, the system implementing the method according to any one of claims 1 to 8, comprising: The acquisition module is used for collecting physical and mechanical data, characteristic volatile organic matter chromatographic data, multichannel bionic sensing array response data and structured sensory evaluation data of the microbial protein meat product to be detected in parallel, respectively preprocessing each mode data, extracting key quantization indexes under each mode to form a quality characterization vector, respectively calculating information entropy weight of key quantization indexes of each mode, calculating conflict weight between each index and the rest index, combining the two indexes to generate initial static weight of each index, pairing index vectors of different modes, analyzing the coupling relation between any two index data sequences, and constructing an initial interaction matrix reflecting the interaction strength between cross-mode indexes; The processing module is used for analogizing an initial interaction matrix into a stress interference field of a multi-crack system in the material mechanics, wherein each index is regarded as a crack source, the interaction intensity among the indexes is analogized into stress interference intensity, calculating a net interference effect of each crack source under the combined action of all the remaining crack sources, dynamically correcting initial static weights of key quantization indexes of all modes according to the net interference effect to generate corrected dynamic weights, carrying out weighted fusion on quality characterization vectors by using the corrected dynamic weights to generate comprehensive feature expression vectors, inputting the comprehensive feature expression vectors into a pre-trained similarity evaluation model, and mapping and outputting quantized simulation scores.
10. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein a program which, when executed by a processor, implements the method according to any of claims 1 to 8.

Description

Simulation degree digital evaluation method and system for microbial protein meat product Technical Field The invention relates to the technical field of digital evaluation, in particular to a simulation degree digital evaluation method and system for a microbial protein meat product. Background Along with the continuous deep application of the microbial protein in the food industry, the meat-like product developed by taking the microbial protein as a raw material gradually becomes a research hot spot, the microbial protein meat product is close to animal meat in nutrition composition, has the advantages of high production efficiency, small environmental impact and the like, however, the conventional microbial protein meat product still has a certain gap from a real meat product in sense attributes such as texture, flavor and the like, and how to scientifically and quantitatively evaluate the simulation degree of the microbial protein meat product and the meat product becomes one of key problems restricting the optimization and market popularization of the product. Currently, quality evaluation of meat products against microbial proteins is mostly dependent on single-dimensional physical and chemical index detection or sensory evaluation. Taking research and development of jerky products as an example, research and development personnel usually respectively measure the texture characteristics and flavor components of the products and comprehensively evaluate the products by combining with sensory scores, however, the multi-source data have the problem that the weight distribution is not reasonable enough in a fusion process, different indexes have different contribution degrees to the simulation degree, and a certain degree of interaction can exist among the indexes, for example, the change of certain flavor substances can have an indirect effect on the sensory scores. When the weight of each index is determined by the existing evaluation method, a subjective assignment or single objective assignment mode is generally adopted, so that the information content of the index and the relevance between the index and other indexes are difficult to fully consider, and the accuracy of an evaluation result can be influenced to a certain extent. Disclosure of Invention The invention aims to provide a simulation degree digital evaluation method and a simulation degree digital evaluation system for a microbial protein meat product, so as to improve the simulation effect of the microbial protein meat product. In order to solve the technical problems, the technical scheme of the invention is as follows: In a first aspect, a method for digitally evaluating the simulation of a microbial protein meat product, the method comprising: step 1, collecting physical and mechanical data, characteristic volatile organic matter chromatographic data, multichannel bionic sensing array response data and structured sensory evaluation data of a to-be-detected microbial protein meat product in parallel, respectively preprocessing each mode data, extracting key quantization indexes under each mode, and forming a quality characterization vector; step 2, respectively calculating information entropy weights of key quantization indexes of all modes, calculating conflict weights between all indexes and the rest indexes, combining the two to generate initial static weights of all indexes; Step 3, simulating an initial interaction matrix as a stress interference field of a multi-crack system in the material mechanics, wherein each index is regarded as a crack source, and the interaction intensity among indexes is simulated as stress interference intensity; And 4, carrying out weighted fusion on the quality characterization vector by using the corrected dynamic weight to generate a comprehensive feature expression vector, inputting the comprehensive feature expression vector into a pre-trained similarity evaluation model, and mapping and outputting quantized simulation degree scores. In a second aspect, a system for digitally evaluating the simulation degree of a microbial protein meat product, comprising: The acquisition module is used for collecting physical and mechanical data, characteristic volatile organic matter chromatographic data, multichannel bionic sensing array response data and structured sensory evaluation data of the microbial protein meat product to be detected in parallel, respectively preprocessing each mode data, extracting key quantization indexes under each mode to form a quality characterization vector, respectively calculating information entropy weight of key quantization indexes of each mode, calculating conflict weight between each index and the rest index, combining the two indexes to generate initial static weight of each index, pairing index vectors of different modes, analyzing the coupling relation between any two index data sequences, and constructing an initial interaction matrix reflecting the interaction strength