Food quality research and judgment method and system based on big data

Abstract

The invention discloses a food quality research and judgment method and system based on big data, and the method comprises the steps: collecting multi-source heterogeneous data of food, and carrying out the preprocessing of the data; constructing and training a convolutional neural network, and calculating a food basic detection index according to the food multi-source heterogeneous data; performing defect identification on the food appearance image by using a convolutional neural network model to obtain food appearance features, and calculating a food appearance feature coefficient; calculating a food environment interference coefficient according to the food multi-source heterogeneous data; calculating a food quality evaluation value according to the food basic detection index, the appearance characteristic coefficient and the food environment interference coefficient; the food quality evaluation value is compared with a set threshold value, and the food quality grade is judged; the method can effectively solve the problem of traditional food quality research and judgment, comprehensively reflect the internal and external quality of food, efficiently and accurately research and judge the food quality, find potential problems in time, and adapt to different food detection requirements.

Description

Technical Field

[0001] The present invention relates to the technical field of food quality detection, and in particular to a food quality analysis method and system based on big data. Background Art

[0002] In the field of food quality testing, traditional methods primarily rely on single data sources, such as chemical analysis, sensory evaluation, or simple image processing. While these methods can reflect certain food characteristics to a certain extent, they have significant limitations. First, the single data source cannot fully reflect the intrinsic and extrinsic quality of food. For example, chemical analysis can only detect food composition but cannot assess its appearance and sensory properties. Sensory evaluation is easily influenced by subjective factors and lacks objectivity and consistency.

[0003] Secondly, manual testing is inefficient, time-consuming, and susceptible to subjective factors, making it difficult to achieve large-scale, rapid testing. Traditional methods lack real-time performance, making it difficult to dynamically monitor changes in food quality during production, storage, and transportation. Traditional statistical methods struggle to handle the complex nonlinear relationships between multi-source heterogeneous data, making it impossible to accurately assess the overall quality of food. With the rapid development of the food industry, higher requirements are being placed on the accuracy and efficiency of food quality testing. These limitations of traditional methods have become increasingly apparent, making it difficult to meet the high quality control demands of the modern food industry.

[0004] This application comprehensively reflects the intrinsic and extrinsic quality of food by integrating multi-source heterogeneous data on food. Convolutional neural networks are used to identify defects in food appearance images and identify defect types and locations. This method is more efficient, accurate, and not affected by subjective factors. At the same time, through machine learning algorithms, modeling is carried out to integrate basic food detection indicators, food appearance characteristic coefficients, and food environmental interference coefficients to calculate a comprehensive evaluation value of food quality and achieve quantitative evaluation of food quality. This system supports real-time data collection and processing, can dynamically monitor changes in food quality, and detect potential problems in a timely manner. The system is highly intelligent and scalable. With the increase in data volume and optimization of the model, it can adapt to different food testing needs and provide strong support for intelligent management and quality control in the food industry. Summary of the Invention

[0005] The present invention provides a method for evaluating food quality based on big data, which includes:

[0006] Collect multi-source heterogeneous food data and pre-process the data;

[0007] Construct and train convolutional neural networks to calculate basic food testing indicators based on multi-source heterogeneous food data;

[0008] Use the convolutional neural network model to identify defects in food appearance images, obtain food appearance features, and calculate food appearance feature coefficients;

[0009] Calculate the food environment interference coefficient based on multi-source heterogeneous food data;

[0010] Calculate the food quality assessment value based on basic food testing indicators, appearance characteristic coefficients, and food environment interference coefficients;

[0011] The food quality level is determined by comparing the food quality assessment value with the set threshold.

[0012] For example, in the above-mentioned food quality assessment method based on big data, a convolutional neural network is constructed and trained to calculate basic food testing indicators based on multi-source heterogeneous food data, including:

[0013] Construct a convolutional neural network architecture to preprocess multi-source heterogeneous food data;

[0014] Use optimization algorithms to train convolutional neural network models and calculate basic food testing indicators.

[0015] For example, in the above-mentioned food quality assessment method based on big data, a convolutional neural network model is used to identify defects in food appearance images, obtain food appearance features, and calculate food appearance feature coefficients, including:

[0016] The processed image is input into the convolutional neural network model to identify the type and location of defects on the food surface and obtain the appearance characteristics of the food;

[0017] By analyzing the appearance characteristics of food, key indicators are extracted and the food appearance characteristic coefficient is calculated.

[0018] For example, in the above-mentioned food quality assessment method based on big data, the food quality assessment value is calculated based on the basic food testing indicators, appearance characteristic coefficient, and food environment interference coefficient, including:

[0019] Use machine learning algorithms to model the complex nonlinear relationships of multi-source heterogeneous food data and build a comprehensive evaluation model;

[0020] The basic food testing indicators, food appearance characteristic coefficients, food environmental interference coefficients and multi-source heterogeneous food data are input into the model, and the food quality assessment value is calculated through nonlinear operations.

[0021] A food quality assessment system based on big data, including:

[0022] The data center module is used to collect multi-source heterogeneous food data and pre-process the data;

[0023] The basic detection index calculation module is used to build and train convolutional neural networks to calculate basic food detection indicators based on multi-source heterogeneous food data;

[0024] The appearance feature recognition module is used to use a convolutional neural network model to identify defects in food appearance images, obtain food appearance features, and calculate food appearance feature coefficients;

[0025] The environmental interference coefficient calculation module is used to calculate the food environmental interference coefficient based on multi-source heterogeneous food data;

[0026] The quality assessment module is used to calculate the food quality assessment value and determine the food quality grade based on the basic food testing indicators, appearance characteristic coefficient, and food environment interference coefficient.

[0027] For example, in the aforementioned food quality assessment system based on big data, the basic detection index calculation module is used to construct and train a convolutional neural network to calculate basic food detection indicators based on multi-source heterogeneous food data, including:

[0028] The model training submodule is used to build a convolutional neural network architecture and preprocess multi-source heterogeneous food data;

[0029] The indicator calculation submodule is used to train the convolutional neural network model using an optimization algorithm and calculate basic food detection indicators.

[0030] For example, in the aforementioned food quality assessment system based on big data, the appearance feature recognition module is used to use a convolutional neural network model to identify defects in food appearance images, obtain food appearance features, and calculate food appearance feature coefficients, including:

[0031] The image recognition module is used to input the processed image into the convolutional neural network model to identify the type and location of defects on the food surface and obtain the appearance characteristics of the food;

[0032] The appearance feature analysis module is used to extract key indicators and calculate food appearance feature coefficients by analyzing food appearance features.

[0033] For example, in the aforementioned food quality assessment system based on big data, the quality assessment module is used to calculate the food quality assessment value and determine the food quality grade based on the basic food testing indicators, appearance characteristic coefficient, and food environment interference coefficient, including:

[0034] The model building submodule is used to use machine learning algorithms to model the complex nonlinear relationships of multi-source heterogeneous food data and build a comprehensive evaluation model;

[0035] The comprehensive evaluation submodule is used to input food basic detection indicators, food appearance characteristic coefficients, food environmental interference coefficients and food multi-source heterogeneous data into the model, and calculate the food quality evaluation value through nonlinear operations.

[0036] The beneficial effects achieved by the present invention are as follows:

[0037] This proposal proposes a food quality assessment method and system based on big data and deep learning. By integrating multi-source heterogeneous food data, it comprehensively reflects the intrinsic and extrinsic quality of food. Convolutional neural networks are used to identify defects in food appearance images and identify defect types and locations. This method is more efficient, accurate, and not affected by subjective factors. At the same time, machine learning algorithms are used to model the food and calculate the comprehensive evaluation value of food quality, thereby achieving a quantitative assessment of food quality. It supports real-time data collection and processing, can dynamically monitor changes in food quality, and promptly identify potential problems. The system is highly intelligent and scalable. With the increase in data volume and optimization of the model, it can adapt to different food testing needs and provide strong support for intelligent management and quality control in the food industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0039] Figure 1 This is a flow chart of a method for evaluating food quality based on big data provided in Example 1 of the present invention;

[0040] Figure 2 This is a schematic diagram of a food quality assessment system based on big data provided in the second embodiment of the present invention; DETAILED DESCRIPTION

[0041] The following is a clear and complete description of the technical solutions in the embodiments of the present invention, in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0042] Example 1

[0043] like Figure 1 As shown, the first embodiment of the present application provides a food quality assessment method based on big data, including:

[0044] S110: Collect multi-source heterogeneous data of food and pre-process the data.

[0045] The system integrates multiple devices and technologies to collect a wide range of food-related multi-source heterogeneous data from different sources.

[0046] The system cleans the large amount of collected data, removes invalid, duplicate or erroneous data, eliminates abnormal data caused by sensor failure, and corrects data entry errors.

[0047] Format data from different sources and convert them into a unified format; normalize numerical data so that its range is unified to [0,1] for model training and optimization.

[0048] Enhance image data through operations such as rotation, scaling, and cropping to generate more diverse training samples and enhance the model's generalization capabilities. Extract key features from complex data, reduce data dimensionality, and improve processing efficiency.

[0049] The pre-processed data is stored in a data warehouse.

[0050] S120: Build and train a convolutional neural network model to calculate basic food testing indicators based on multi-source heterogeneous food data;

[0051] The convolutional neural network model is a computational model that imitates the structure and function of biological neurons. It consists of a large number of nodes (neurons), which are connected to each other through connection weights to form a complex network structure.

[0052] Convolutional neural network models receive data through an input layer, process it layer by layer through multiple hidden layers, and ultimately generate results at the output layer. Each neuron performs a weighted summation of the input signal, and then introduces nonlinear characteristics through the activation function, enabling the network to learn and simulate complex functional relationships.

[0053] The convolutional neural network model uses a large amount of labeled data to adjust the connection weights through the training process to minimize the error between the predicted value and the true value, thereby achieving data classification, regression or feature extraction.

[0054] Construct and train a convolutional neural network model to calculate basic food testing indicators based on multi-source heterogeneous food data. The specific steps include the following:

[0055] S121: Build a convolutional neural network architecture to preprocess multi-source heterogeneous food data.

[0056] Design multiple layers of convolutional layers, pooling layers, and fully connected layers, and divide the training set and test set.

[0057] When building a convolutional neural network architecture for food quality assessment, the system designed multiple convolutional layers, pooling layers, and fully connected layers to effectively process and extract features from multi-source, heterogeneous food data. Multiple convolution kernels (filters) are defined, which slide over the input data, performing convolution operations to extract local features.

[0058] The size, number, and stride of the convolution kernel are key design parameters and are typically adjusted based on the complexity of the data and task requirements. For example, smaller convolution kernels (such as 3×3 or 5×5) can capture finer local features, while more convolution kernels can extract richer feature information.

[0059] The pooling layer is designed to reduce the spatial dimension of features, reduce the amount of computation, and retain important features.

[0060] Pooling operations include maximum pooling and average pooling. Maximum pooling preserves significant features by taking the maximum value of a local region, while average pooling smoothes features by taking the average value. The size and stride of the pooling layer also affect the degree of feature compression. A 2×2 pooling window and a stride of 2 are typically chosen to halve the size of the feature map.

[0061] A fully connected layer integrates the extracted features and outputs the final regression result. In a fully connected layer, each neuron is connected to all neurons in the previous layer, and a nonlinear transformation is performed using weighted summation and an activation function. The number of neurons in the fully connected layer is designed based on the complexity of the task.

[0062] The preprocessed data is divided into training and testing sets.

[0063] S122: Use optimization algorithms to train convolutional neural network models and calculate basic food testing indicators.

[0064] Adjust hyperparameters, learn data features through training sets, and verify model performance using test sets.

[0065] Use the training set to train the convolutional neural network model, and adjust the network parameters through the backpropagation algorithm so that the model can learn the features and patterns in the data.

[0066] Use the test set to verify the performance of the convolutional neural network model and evaluate its generalization ability on unseen data.

[0067] The model is trained by using an optimization algorithm to adjust the weights and bias parameters of the network to better fit the training data by minimizing the loss function (such as cross entropy loss or mean squared error).

[0068] The network parameters are gradually updated according to the data characteristics in the training set, the key features in the multi-source heterogeneous food data are learned, and the hyperparameters are adjusted to optimize the training process.

[0069] Hyperparameters include learning rate, batch size, convolution kernel size, number of layers, etc. The optimal hyperparameter combination is selected through multiple experiments and verifications.

[0070] Use the test set to verify model performance. Measure model performance by calculating metrics such as accuracy, recall, and F1 score on the test set, and further adjust hyperparameters to optimize model performance based on computational needs.

[0071] Input multi-source heterogeneous food data into the trained model to calculate basic food testing indicators.

[0072] The calculation formula for basic food testing indicators is as follows:

[0073]

[0074] β1 is a weight coefficient, whose value range is usually between [0,1] and is used to adjust the importance of the calculation module on the left in the entire calculation. P represents the porosity of the food, which reflects the internal space of the food. The porosity of the food has a significant impact on its physical properties such as water absorption and air permeability, which in turn affects the storage stability, taste and other quality attributes of the food. log(1+P) performs a logarithmic transformation on the porosity and reasonably scales its numerical range to avoid excessive influence on the calculation results due to excessively large or small porosity values, while making the data distribution more in line with the requirements of the calculation model.

[0075] C represents the content of vitamins in food, and its content directly affects the nutritional value of the food. M is the standard reference content of vitamins in food, which is pre-set based on the type and grade of the food. The exponential function is used to highlight the proportional relationship between the actual vitamin content in food and the standard reference content, thereby amplifying the impact of content differences on food quality assessment.

[0076] F represents the water activity of food, which is an indicator that measures the presence of water in food. It is closely related to the microbial growth and chemical reaction activity of food, and greatly affects the shelf life and safety of food. V is the temperature coefficient related to water activity, which is used to correct the effect of water activity on food quality under different temperature conditions. The ratio of water activity to temperature coefficient is mapped to The range should be kept within to prevent the instability of the calculation results caused by excessive fluctuations in the ratio of the two.

[0077] log(1+M) performs logarithmic transformation on the standard reference content of vitamins to avoid calculation overflow caused by excessively large standard reference content values; E represents the energy consumption during food processing;

[0078] β2 weight coefficient, the value range is generally [0,1], which is used to determine the weight of the middle calculation module in the comprehensive index; Calculate the absolute value of the deviation between the sum of all relevant physical and chemical indicators and the average value to measure the degree of dispersion or balance of these physical and chemical indicators; F i represents the value of the i-th specific physical and chemical index related to food quality, and n represents the total number of such physical and chemical indicators.

[0079] S130: Using a convolutional neural network model to identify defects in the food appearance image, obtain food appearance features, and calculate food appearance feature coefficients;

[0080] Furthermore, a convolutional neural network model is used to identify food appearance defects and calculate food appearance characteristic coefficients based on food appearance characteristics, specifically including the following sub-steps:

[0081] S131: Input the processed image into a convolutional neural network model to identify the type and location of defects on the food surface and obtain the appearance characteristics of the food.

[0082] Deep features of the image are gradually extracted through multiple layers of convolution and pooling operations. The convolution kernel in the convolution layer slides across the image, performing convolution operations on each local area to detect features such as edges, textures, and shapes in the image.

[0083] As convolutional layers are stacked one after another, these features are gradually abstracted from low-level simple patterns to higher-level complex patterns. By reducing the spatial dimensions of the feature map, the computational effort is reduced while retaining important features, enhancing the model's robustness to local variations.

[0084] Among them, the deep features extracted by the model contain rich image information.

[0085] These feature maps are passed to the subsequent fully connected layer to integrate and classify the extracted features. Through the training process, the convolutional neural network model learns the characteristic patterns corresponding to different defect types (such as scratches, spots, deformations, etc.) and identifies the specific locations of these patterns in the image.

[0086] Output the appearance characteristics of food, including the type, location, size and other information of defects.

[0087] S132: By analyzing the appearance characteristics of food, extract key indicators and calculate the food appearance characteristic coefficient.

[0088] The calculation formula for food appearance characteristic coefficient is as follows:

[0089]

[0090] α is the weight coefficient, which is used to measure the importance of food color appearance characteristics.i Represents the quantitative index of color characteristics of food appearance, each H i The color-related value corresponding to a certain area or observation point on the food appearance; μ H It is H i The mean of these color feature quantitative indicators is the value obtained by averaging the color-related values ​​of different areas or observation points on the food surface; It reflects the degree of dispersion of food appearance color relative to the average color and reflects the difference in uniformity of color distribution.

[0091] β represents the weight coefficient, which is used to reflect the importance of food appearance and shape; M represents the dimension of the matrix or data set related to food appearance and shape characteristics. Here, the subscripts range from 1 to M, indicating that data of size M*M are involved in the calculation; P(i, j) is an element in a two-dimensional matrix, and the values ​​of i and j both range from 1 to M, representing the similarity probability between different shape elements of food appearance; |ij| is used to measure the distance relationship between the positions of shape elements in the matrix, which can be associated with the spatial position difference of shape elements on food appearance; The product of each element in the shape similarity matrix P(i,j) and the corresponding |ij| is double-summed, reflecting a comprehensive measure based on the similarity of shape elements and spatial position relationships.

[0092] γ represents the weight coefficient, which measures the effect of the maximum and average changes in food appearance texture characteristics on the food appearance characteristic coefficient; η b Represents the texture characteristic parameters of food appearance, such as texture roughness, texture direction and other quantitative values; each η b Corresponding to the texture-related value of a certain local area on the food appearance; max|η b | Take the maximum value to highlight the maximum change in food appearance texture characteristics; It reflects the overall change degree of texture features and the average change of texture features.

[0093] S140: Calculate the food environment interference coefficient based on multi-source heterogeneous food data;

[0094] The calculation formula of food environment interference coefficient is as follows:

[0095]

[0096] The comprehensive quantitative value that represents the impact of temperature fluctuation on food quality; T i represents the temperature value obtained from the i-th measurement, reflecting the actual temperature fluctuation. Temperature fluctuation will affect the chemical reaction rate and microbial growth; T optIndicates the optimal temperature for food storage or processing. It represents the ratio of the actual temperature to the optimal temperature, reflecting the degree to which the actual temperature deviates from the optimal state. n is the total number of temperature measurements, which comprehensively reflects temperature fluctuations through multiple measurements. R is determined based on the characteristics of temperature's impact on food quality and is used to adjust the weight of the impact of temperature deviation on food quality.

[0097] It is the cumulative impact value obtained by calculating the attenuation of light intensity and its impact on food quality at different times during the time period from time 0 to time t; where t represents the cumulative impact of light on food quality from the beginning to time t; I(ε) represents the light exposure at time ε, describing the change of light intensity over time; ε represents the current time; K represents the attenuation coefficient, reflecting the attenuation rate of the impact of light on food quality over time; -k*ε represents the characteristic that the impact of light on food quality weakens over time; as time ε increases, the value of -k*ε will gradually decrease;

[0098] S j represents the quantitative value of the jth environmental factor, such as relative humidity deviation, oxygen exposure concentration, etc.; f j Indicates that S j The corresponding weight coefficient is used to measure the importance of different environmental factors on food quality; M represents the total number of environmental factors, covering a variety of factors affecting food quality; N cur It indicates the current concentration of a substance related to food quality (such as the remaining amount of antioxidants, etc.); γ indicates the index of the impact of substance changes on food quality; H ideal A baseline value under ideal conditions, used to compare the current state, such as ideal antioxidant content.

[0099] S150: Calculate the food quality assessment value based on the basic food testing indicators, appearance characteristic coefficient, and food environment interference coefficient;

[0100] The nonlinear relationship between multi-source heterogeneous food data is modeled and the food quality assessment value is calculated, which specifically includes the following sub-steps:

[0101] S151: Use machine learning algorithms to model the complex nonlinear relationships of multi-source heterogeneous food data and construct a comprehensive evaluation model.

[0102] Use machine learning algorithms to model complex nonlinear relationships between multi-source heterogeneous data.

[0103] The system first selects a machine learning algorithm suitable for processing nonlinear relationships, such as support vector machines or random forests, based on factors such as data characteristics, task requirements, algorithm characteristics, and computing resources.

[0104] For support vector machines, the system maps the input data into a high-dimensional space by selecting a suitable kernel function, and searches for the optimal segmentation hyperplane that maximizes the interval in the space to handle the nonlinear relationship in the data.

[0105] For random forests, the system reduces the variance of the model and improves its generalization ability by constructing multiple decision trees and introducing randomness in the training process of each decision tree, including randomly selecting feature subsets and sample subsets.

[0106] Through these algorithms, the system learns and fits the input food basic detection indicators, food appearance characteristic coefficients, food environmental interference coefficients, food multi-source heterogeneous data, etc., to build a model that can comprehensively consider various factors and output food quality assessment values.

[0107] S152: Input the basic food testing indicators, food appearance characteristic coefficients, food environmental interference coefficients and food multi-source heterogeneous data into the model, and calculate the food quality evaluation value through nonlinear operations.

[0108] The formula for calculating food quality assessment value is as follows:

[0109]

[0110] β1 represents the proportion of the average quantitative value of the combination of basic food testing indicators and appearance characteristics; ω i represents the weight of the i-th factor related to the basic detection index or appearance characteristics of food; C represents the basic detection index of food, which is the quantitative value of the detection data such as the nutritional components and microbial count of food; F i It represents the food appearance characteristic coefficient, which is the quantitative value of appearance attributes such as color, shape, and integrity; n represents the number of factors related to the basic food testing indicators and appearance characteristics;

[0111] β2 weight coefficient reflects the importance of food taste, processing complexity and environmental interference factors in the calculation; z represents the number of factors related to food taste; N represents the food production batch coefficient, and the larger the coefficient, the greater the impact on food quality; u j represents the coefficient of the jth factor related to food taste, such as the quantitative coefficients corresponding to different taste dimensions such as sour, sweet, bitter, salty, and fresh; v i represents the coefficient of the i-th factor related to the complexity of food processing, and the quantitative coefficients corresponding to different processing steps and difficulty levels; T i represents the i-th parameter related to food processing, such as baking time, sterilization temperature, etc.; different T i Describe the processing process from different process dimensions; represent the index of the i-th parameter related to food processing, used to adjust T i, reflecting the sensitivity of certain process parameters to changes; T represents the average storage temperature; S is a benchmark parameter used to convert T into a relative value; E represents the food environmental interference coefficient, reflecting the degree of influence of environmental factors on food quality; K is a parameter that adjusts the overall scale of the formula, adjusting the role of the environmental interference coefficient E in the calculation; m represents the number of factors related to the complexity of food processing.

[0112] S160: Determine the food quality grade based on the food quality assessment value and the set threshold;

[0113] The system compares the food quality assessment value with the preset threshold value.

[0114] Based on the range of the evaluation value, the specific quality grade of the food is determined as excellent, good, qualified, or unqualified.

[0115] Automatically record assessment values ​​and judgment results, and associate them with food batch, production date, source, and other information. Generate corresponding reports or notifications based on the judgment results and provide timely feedback to production management personnel.

[0116] Example 2

[0117] like Figure 2 As shown, the second embodiment of the present application provides a food quality analysis system based on big data, including:

[0118] Data hub module 21: used to collect multi-source heterogeneous data of food and pre-process the data.

[0119] Basic detection index calculation module 22: used to build and train convolutional neural networks to calculate basic food detection indicators based on multi-source heterogeneous food data, specifically including:

[0120] Model training submodule 221: Construct a convolutional neural network architecture to preprocess multi-source heterogeneous food data.

[0121] Index calculation submodule 222: used to use the optimization algorithm to train the convolutional neural network model and calculate the basic food detection indicators.

[0122] Appearance feature recognition module 23: used to use a convolutional neural network model to identify defects in food appearance images, obtain food appearance features, and calculate food appearance feature coefficients, specifically including:

[0123] Image recognition module 231: used to input the processed image into the convolutional neural network model to identify the type and location of defects on the food surface and obtain the appearance characteristics of the food.

[0124] Appearance feature analysis module 232: used to extract key indicators and calculate food appearance feature coefficients by analyzing food appearance features.

[0125] Environmental interference coefficient calculation module 24: used to calculate the food environmental interference coefficient based on multi-source heterogeneous food data;

[0126] Quality Assessment Module 25: This module is used to model the nonlinear relationship between multi-source heterogeneous food data, calculate the food quality assessment value based on the basic food testing indicators, appearance characteristic coefficients, and food environmental interference coefficients; and compare the food quality assessment value with the set threshold to determine the food quality grade. Specifically, it includes:

[0127] Model building submodule 251: used to use machine learning algorithms to model the complex nonlinear relationship of multi-source heterogeneous food data and build a comprehensive evaluation model.

[0128] Comprehensive evaluation submodule 252: used to input food basic detection indicators, food appearance characteristic coefficients, food environmental interference coefficients and food multi-source heterogeneous data into the model, and calculate the food quality evaluation value through nonlinear operations.

[0129] Corresponding to the above embodiment, an embodiment of the present invention provides a computer storage medium, comprising: at least one memory and at least one processor;

[0130] The memory is used to store one or more program instructions;

[0131] The processor is used to run one or more program instructions to execute a food quality assessment method based on big data.

[0132] Corresponding to the above embodiment, an embodiment of the present invention provides a computer-readable storage medium, which contains one or more program instructions, and the one or more program instructions are used by a processor to execute a food quality analysis method based on big data.

[0133] The embodiments disclosed in the present invention provide a computer-readable storage medium, in which computer program instructions are stored. When the computer program instructions are executed on a computer, the computer executes the above-mentioned food quality analysis method based on big data.

[0134] In the embodiments of the present invention, the processor may be an integrated circuit chip having signal processing capabilities. The processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0135] The methods, steps, and logic diagrams disclosed in the embodiments of the present invention can be implemented or executed. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of the present invention can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules within the decoding processor. The software modules can be located in a storage medium well-established in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. The processor reads the information from the storage medium and, in conjunction with its hardware, completes the steps of the aforementioned methods.

[0136] The storage medium may be a memory and may be, for example, a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memory.

[0137] Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory.

[0138] Volatile memory may be random access memory (RAM), which is used as an external cache memory. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct RAM bus random access memory (DRRAM).

[0139] The storage media described in the embodiments of the present invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0140] Those skilled in the art will appreciate that in one or more of the above examples, the functions described herein can be implemented using a combination of hardware and software. When software is used, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media includes any medium that facilitates the transmission of computer programs from one place to another. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0141] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A food quality assessment method based on big data, characterized in that: include: Collect multi-source heterogeneous food data and pre-process the data; Construct and train convolutional neural networks to calculate basic food testing indicators based on multi-source heterogeneous food data; Use the convolutional neural network model to identify defects in food appearance images, obtain food appearance features, and calculate food appearance feature coefficients; Calculate the food environment interference coefficient based on multi-source heterogeneous food data; Calculate the food quality assessment value based on basic food testing indicators, appearance characteristic coefficients, and food environment interference coefficients; The food quality level is determined by comparing the food quality assessment value with the set threshold.

2. A food quality assessment method based on big data according to claim 1, characterized in that: Construct and train convolutional neural networks to calculate basic food testing indicators based on multi-source heterogeneous food data, including: Construct a convolutional neural network architecture to preprocess multi-source heterogeneous food data; Use optimization algorithms to train convolutional neural network models and calculate basic food testing indicators.

3. The method for food quality assessment based on big data according to claim 1, characterized in that: Use the convolutional neural network model to identify defects in food appearance images, obtain food appearance features, and calculate food appearance feature coefficients, including: The processed image is input into the convolutional neural network model to identify the type and location of defects on the food surface and obtain the appearance characteristics of the food; By analyzing the appearance characteristics of food, key indicators are extracted and the food appearance characteristic coefficient is calculated.

4. The method for evaluating food quality based on big data according to claim 1, wherein: Calculate the food quality assessment value based on the basic food testing indicators, appearance characteristic coefficient, and food environment interference coefficient, including: Use machine learning algorithms to model the complex nonlinear relationships of multi-source heterogeneous food data and build a comprehensive evaluation model; The basic food testing indicators, food appearance characteristic coefficients, food environmental interference coefficients and multi-source heterogeneous food data are input into the model, and the food quality assessment value is calculated through nonlinear operations.

5. A food quality assessment system based on big data, characterized in that: include: The data center module is used to collect multi-source heterogeneous food data and pre-process the data; The basic detection index calculation module is used to build and train convolutional neural networks to calculate basic food detection indicators based on multi-source heterogeneous food data; The appearance feature recognition module is used to use a convolutional neural network model to identify defects in food appearance images, obtain food appearance features, and calculate food appearance feature coefficients; The environmental interference coefficient calculation module is used to calculate the food environmental interference coefficient based on multi-source heterogeneous food data; The quality assessment module is used to calculate the food quality assessment value and determine the food quality grade based on the basic food testing indicators, appearance characteristic coefficient, and food environment interference coefficient.

6. A food quality evaluation system based on big data according to claim 5, characterized in that: The basic detection index calculation module is used to build and train convolutional neural networks to calculate basic food detection indicators based on multi-source heterogeneous food data, including: The model training submodule is used to build a convolutional neural network architecture and preprocess multi-source heterogeneous food data; The indicator calculation submodule is used to train the convolutional neural network model using an optimization algorithm and calculate basic food detection indicators.

7. The food quality evaluation system based on big data according to claim 5 is characterized in that: The appearance feature recognition module is used to identify defects in food appearance images using a convolutional neural network model, obtain food appearance features, and calculate food appearance feature coefficients, including: The image recognition module is used to input the processed image into the convolutional neural network model to identify the type and location of defects on the food surface and obtain the appearance characteristics of the food; The appearance feature analysis module is used to extract key indicators and calculate food appearance feature coefficients by analyzing food appearance features.

8. The food quality evaluation system based on big data according to claim 5 is characterized in that: The quality assessment module is used to calculate the food quality assessment value and determine the food quality grade based on the basic food testing indicators, appearance characteristic coefficient, and food environment interference coefficient, including: The model building submodule is used to use machine learning algorithms to model the complex nonlinear relationships of multi-source heterogeneous food data and build a comprehensive evaluation model; The comprehensive evaluation submodule is used to input food basic detection indicators, food appearance characteristic coefficients, food environmental interference coefficients and food multi-source heterogeneous data into the model, and calculate the food quality evaluation value through nonlinear operations.

Images