Food microbiological detection method and device combined with microbial growth prediction

Abstract

The present application provides a food microorganism detection method and device combined with microorganism growth prediction, which relates to the field of microorganism detection technology, including: predicting microorganism growth based on multiple food image feature information to obtain a microorganism category set and multiple first growth scale information; predicting microorganism growth scale based on plastic wrap feature information to obtain second growth scale information; performing growth scale analysis based on multiple first and second growth scale information to obtain a growth scale range, outputting microorganism prediction results in combination with the microorganism category set, and formulating a detection strategy for detection. This application can solve the technical problem that traditional detection methods are difficult to accurately predict the growth trend of microorganisms in the state of plastic wrap packaging, resulting in the inability to set detection strategies in a targeted manner, resulting in greater blindness and randomness in microorganism detection, and can achieve the goal of efficiently predicting the growth trend of microorganisms, thereby facilitating the formulation of more scientific and accurate detection strategies.

Description

Technical Field

[0001] The present application relates to the technical field of food microorganism detection, and in particular to a food microorganism detection method and device combined with microorganism growth prediction. Background Art

[0002] With the improvement of food safety standards, microbial detection has become an important part of food quality control. However, traditional food microbial detection methods mostly rely on laboratory sampling and cultivation, which not only requires the destruction of plastic wrap packaging, but also takes a long time. Especially in the plastic wrap packaging state, traditional detection methods are difficult to quickly obtain dynamic data of microorganisms, resulting in detection strategies that can only rely on fixed schemes or empirical settings, lacking flexibility and pertinence. This method cannot be adjusted according to the actual growth of microorganisms, and is prone to detection blind spots or false detections.

[0003] In summary, traditional food microbial detection methods are difficult to accurately predict the growth trend of microorganisms in the state of plastic wrap packaging, resulting in the inability to set targeted detection strategies, causing technical problems such as high blindness and randomness in microbial detection. Summary of the invention

[0004] The purpose of this application is to provide a food microorganism detection method and device combined with microbial growth prediction, so as to solve the technical problem that traditional food microorganism detection methods are difficult to accurately predict the growth trend of microorganisms under the state of plastic wrap packaging, resulting in the inability to set targeted detection strategies, causing the microbial detection to be relatively blind and random.

[0005] In view of the above problems, the present application provides a food microorganism detection method and device combined with microbial growth prediction.

[0006] In a first aspect, the present application provides a food microorganism detection method combined with microorganism growth prediction, which is implemented by a food microorganism detection device combined with microorganism growth prediction, including: performing image acquisition at a first angle on the food to be detected, obtaining regional food images within multiple acquisition areas, performing identification, and obtaining multiple food image feature information, wherein the food is food wrapped in plastic wrap; performing microorganism growth category prediction and microorganism growth scale prediction based on the multiple food image feature information, and obtaining a microorganism category set and multiple first microorganism growth scale information; performing image acquisition at a second angle on the food to be detected, obtaining a plastic wrap image, performing identification, obtaining plastic wrap feature information, and performing microorganism growth scale prediction to obtain second microorganism growth scale information; performing microorganism growth scale analysis based on the multiple first microorganism growth scale information and the second microorganism growth scale information to obtain a microorganism growth scale range, outputting microorganism prediction results in combination with the microorganism category set, and formulating a detection strategy to perform food microorganism detection.

[0007] In a second aspect, the present application further provides a food microorganism detection device combined with microorganism growth prediction, which is used to execute a food microorganism detection method combined with microorganism growth prediction as described in the first aspect, including: a food image feature information acquisition module, the food image feature information acquisition module is used to capture images of the food to be detected at a first angle, obtain regional food images within multiple acquisition areas, perform identification, and obtain multiple food image feature information, wherein the food is food packaged in plastic wrap; a first growth scale information acquisition module, the first growth scale information acquisition module is used to perform microorganism growth category prediction and microorganism growth scale prediction based on the multiple food image feature information, and obtain microorganism growth. A biological category set and a plurality of first microbial growth scale information; a second growth scale information acquisition module, the second growth scale information acquisition module is used to collect images of the food to be tested at a second angle, obtain a plastic wrap image, perform identification, obtain plastic wrap feature information, and perform microbial growth scale prediction to obtain second microbial growth scale information; a microbial prediction result output module, the microbial prediction result output module is used to perform microbial growth scale analysis based on the plurality of first microbial growth scale information and second microbial growth scale information, obtain a microbial growth scale range, output microbial prediction results in combination with the microbial category set, and formulate a detection strategy to perform food microbial detection.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] By capturing images of the food to be tested at a first angle, identifying regional food images in multiple acquisition areas, and obtaining multiple food image feature information; then predicting the microbial growth category and microbial growth scale based on the multiple food image feature information, and obtaining a microbial category set and multiple first microbial growth scale information; on the other hand, capturing images of the food to be tested at a second angle to obtain a plastic wrap image; identifying the plastic wrap image to obtain plastic wrap feature information, and predicting the microbial growth scale based on the plastic wrap feature information to obtain second microbial growth scale information; further, based on the multiple first microbial growth scale information and the second microbial growth scale information, performing microbial growth scale analysis to obtain a microbial growth scale range, and using the microbial growth scale range and the microbial category set as microbial prediction results; finally, formulating targeted detection strategies based on the microbial prediction results to perform food microbial detection; by combining technical means such as image recognition and intelligent prediction models, the breadth, efficiency and accuracy of microbial growth prediction under plastic wrap packaging can be improved, and the technical goal of efficiently predicting microbial growth trends can be achieved, thereby facilitating the formulation of more scientific and accurate detection strategies and improving microbial detection efficiency and quality.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0012] Figure 1 A schematic diagram of a process for food microbial detection combined with microbial growth prediction in this application;

[0013] Figure 2 This is a schematic diagram of a process for obtaining feature information of multiple food images in a food microbiology detection method combined with microbial growth prediction in this application;

[0014] Figure 3This is a schematic diagram of the structure of a food microorganism detection device combined with microorganism growth prediction in the present application.

[0015] Description of reference numerals:

[0016] A food image feature information obtaining module 11, a first growth scale information obtaining module 12, a second growth scale information obtaining module 13, and a microorganism prediction result output module 14. DETAILED DESCRIPTION

[0017] This application provides a food microbial detection method and device combined with microbial growth prediction, which solves the technical problem that the traditional food microbial detection method is difficult to accurately predict the growth trend of microorganisms in the state of plastic wrap packaging, resulting in the inability to set targeted detection strategies, resulting in greater blindness and randomness in microbial detection. By combining technical means such as image recognition and intelligent prediction models, the breadth, efficiency and accuracy of microbial growth prediction in the state of plastic wrap packaging can be improved, and the technical goal of efficiently predicting the growth trend of microorganisms can be achieved, thereby facilitating the formulation of more scientific and accurate detection strategies and improving the efficiency and quality of microbial detection.

[0018] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0019] For example, please refer to the attached Figure 1 The present application provides a food microorganism detection method combined with microorganism growth prediction, which is applied to a food microorganism detection device combined with microorganism growth prediction, and specifically includes the following steps:

[0020] S100: Capture images of food to be inspected at a first angle to obtain regional food images within a plurality of acquisition areas, perform identification, and obtain feature information of a plurality of food images, wherein the food is food wrapped in plastic wrap.

[0021] Specifically, obtain the food to be tested, wherein the food is food wrapped in plastic wrap, which can be set according to the application scenario, such as fresh meat wrapped in plastic wrap. Configure the first image acquisition angle, wherein the first image acquisition angle is an angle perpendicular to the horizontal line. Divide the food into regions according to a uniform standard and evenly divide it into multiple food regions. Then, at the first image acquisition angle, use a CCD image sensor to capture images of multiple food regions to obtain regional food images within multiple acquisition regions.

[0022] Then, image color recognition is performed on the food images of multiple regions respectively, that is, color feature extraction is performed on the food images of each collected region. The growth of microorganisms (such as bacteria) in food is often accompanied by color changes. For example, fresh meat may gradually change from bright red to gray or brown when bacteria grow. Therefore, the focus of color recognition is to detect changes in the color of the food surface; for the images of each region, the color feature information is extracted, and parameters such as the type of color are identified. These color features can reflect whether abnormal color changes have occurred on the food surface, such as gray, dark yellow, etc., which are usually early signals of microbial growth, and multiple food image feature information is obtained, wherein the food image feature information is image color information, including parameters such as the type of image color.

[0023] By extracting color features, we can quickly capture the color changes of food during the microbial growth process, provide an important basis for subsequent microbial growth prediction, and help to quickly and accurately judge the growth of microorganisms.

[0024] S200: According to the plurality of food image feature information, microorganism growth category prediction and microorganism growth scale prediction are performed to obtain a microorganism category set and a plurality of first microorganism growth scale information.

[0025] Specifically, microbial growth category prediction is performed based on the feature information of the multiple food images, that is, based on the color features in the image (such as gray, dark, green, yellow, etc.), combined with the association between microorganisms and color changes, the category of microorganisms that may exist in the food is predicted. For example, gray or dark colors are usually related to insufficient oxygen and bacterial reproduction, which may indicate the presence of anaerobic bacteria such as Pseudomonas, which usually multiply in large numbers when food is spoiled, causing the surface to become gray and dark; green or yellow may be related to bacteria such as Pseudomonas or Clostridium perfringens, which can decompose proteins and produce pigment changes under certain conditions; a microbial category set is obtained based on the prediction results, wherein the microbial category set is the union of multiple microbial category prediction results, that is, it includes all microbial categories.

[0026] On the other hand, the microbial growth scale is predicted based on the characteristic information of the multiple food images, and the distribution of the microorganisms is inferred based on the depth of color and area. For example, the darker the color, the higher the bacterial reproduction concentration; the wider the color distribution, the larger the scale of bacterial reproduction, and multiple first microbial growth scale information is obtained. Through the prediction method based on color features, the type and growth scale of microorganisms in food can be quickly and non-destructively inferred, significantly improving the efficiency and accuracy of microbial growth trend prediction.

[0027] S300: Capturing an image of the food to be inspected at a second angle to obtain a plastic wrap image, performing recognition to obtain characteristic information of the plastic wrap, and predicting the growth scale of microorganisms to obtain second microorganism growth scale information.

[0028] Specifically, a second image acquisition angle is configured, and the second image acquisition angle is an angle of 30° relative to the horizontal line, and is used to capture images of the food packaging plastic wrap, wherein the second image acquisition angle can better capture detailed information on the surface of the plastic wrap, such as bulging features. According to the second image acquisition angle, the CCD image sensor is used to capture images of the food to be inspected to obtain a plastic wrap image.

[0029] Then, the bulging height of the plastic wrap image is identified. When bacteria and microorganisms grow on food, a certain amount of gas will be discharged, resulting in less bulging of the plastic wrap. By identifying the bulging area on the surface of the plastic wrap, especially the raised part of the surface, the bulging height can be detected. The change in height is usually related to the amount of gas produced by the microorganism. A larger bulge often means that the microorganisms reproduce more actively. The bulging height of the plastic wrap is obtained, such as a bulging height of 0.5 mm, and the bulging height of the plastic wrap is used as the characteristic information of the plastic wrap. By performing feature recognition on the plastic wrap image and obtaining the bulging height information, the judgment of the growth of microorganisms inside the food can be further supplemented, providing a more comprehensive basis for the overall microbial prediction.

[0030] Then, the scale of microbial growth is predicted based on the characteristic information of the plastic wrap, where the bulging height of the plastic wrap is a direct result of gas production caused by microbial metabolism, so the bulging height and area can reflect the activity and growth scale of the microorganisms inside the food; according to the numerical value of the bulging height (such as 0.5 mm), through the preset microbial growth model, the amount of gas produced by the microorganism can be estimated and converted into microbial growth scale information, for example, the larger the bulging height, the more gas is produced, the more active the microbial reproduction is, and the larger the scale is; the second microbial growth scale information is obtained. By predicting the scale of microbial growth based on the bulging height of the plastic wrap, the degree of microbial reproduction can be accurately estimated, providing a basis for subsequent calculation of the range of microbial growth scales.

[0031] S400: Perform a microbial growth scale analysis based on the plurality of first microbial growth scale information and the second microbial growth scale information to obtain a microbial growth scale range, output a microbial prediction result in combination with the microbial category set, formulate a detection strategy, and perform food microbial detection.

[0032] Specifically, a microbial growth scale analysis is performed based on the multiple first microbial growth scale information and the second microbial growth scale information, that is, the prediction errors of the two types of information (such as image recognition accuracy, swelling height measurement error, etc.) are comprehensively considered, and a reasonable microbial growth scale range is set according to the error situation, which covers the minimum growth scale and maximum growth scale of microbial growth in food, and the microbial growth scale range is obtained.

[0033] Then, the microbial growth scale range and microbial category set are used as microbial prediction results, and targeted detection strategies are formulated based on the microbial prediction results, that is, combining the microbial category set, selecting appropriate detection methods for each possible microbial category, where different microbial categories require different detection methods, and the prediction results provide a reference for each type of microorganism, so as to reasonably allocate detection resources. For example, for Pseudomonas, selective detection of culture medium can be used; for the protein decomposition ability of Clostridium perfringens, specific detection methods can be selected. According to the scale range of microbial growth, the frequency and depth of detection are determined. If the prediction results show that the scale of microbial growth is large, the number of tests needs to be increased to ensure that the active area of ​​the microorganism is fully detected; if the scale of the microorganism is small, the detection frequency can be reduced, and the detection range can be appropriately reduced to improve the detection efficiency; and an adaptive detection strategy is obtained. Finally, food microbial detection is carried out according to the adaptive detection strategy.

[0034] By formulating an adaptive detection strategy based on the microbial prediction results and combining different microbial categories and growth scale information, the accuracy and specificity of food microbial detection can be significantly improved. At the same time, the detection means, number of times and detection limits can be flexibly adjusted to ensure efficient and scientific food microbial detection.

[0035] The food microorganism detection method combined with microorganism growth prediction is applied to a food microorganism detection device combined with microorganism growth prediction, which can solve the technical problem that traditional food microorganism detection methods are difficult to accurately predict the growth trend of microorganisms in a plastic wrap packaging state, resulting in the inability to set targeted detection strategies, causing the microorganism detection to be relatively blind and random. By capturing images of the food to be tested at a first angle, identifying regional food images in multiple acquisition areas, and obtaining multiple food image feature information; then predicting the microbial growth category and microbial growth scale based on the multiple food image feature information, and obtaining a microbial category set and multiple first microbial growth scale information; on the other hand, capturing images of the food to be tested at a second angle to obtain a plastic wrap image; identifying the plastic wrap image to obtain plastic wrap feature information, and predicting the microbial growth scale based on the plastic wrap feature information to obtain second microbial growth scale information; further, based on the multiple first microbial growth scale information and the second microbial growth scale information, performing microbial growth scale analysis to obtain a microbial growth scale range, and using the microbial growth scale range and the microbial category set as microbial prediction results; finally, formulating targeted detection strategies based on the microbial prediction results to perform food microbial detection; by combining technical means such as image recognition and intelligent prediction models, the breadth, efficiency and accuracy of microbial growth prediction under plastic wrap packaging can be improved, and the technical goal of efficiently predicting microbial growth trends can be achieved, thereby facilitating the formulation of more scientific and accurate detection strategies and improving microbial detection efficiency and quality.

[0036] Further, as attached Figure 2 As shown, the food to be inspected is imaged at a first angle, regional food images in multiple acquisition areas are obtained, and identification is performed to obtain multiple food image feature information. The application includes:

[0037] The food to be inspected is divided into regions at a first angle to obtain a plurality of acquisition regions; images are acquired from the plurality of acquisition regions to obtain a plurality of regional food images; a food image recognizer is pre-trained; and the plurality of regional food images are recognized using the food image recognizer to obtain a plurality of food image feature information, wherein each food image feature information includes color information.

[0038] Specifically, at the first image acquisition angle (angle perpendicular to the horizontal line), the food to be tested is divided into multiple acquisition areas on average. Then, according to the first image acquisition angle (angle perpendicular to the horizontal line), the CCD image sensor is used to collect images of multiple acquisition areas to obtain multiple regional food images. Then, a food image recognizer is constructed based on a convolutional neural network, and the food image recognizer is trained to a convergence state using a sample data set, wherein the input data of the food image recognizer is the food image, and the output data is the color feature (color type, RGB value, etc.). Then, the multiple regional food images are respectively input into the food image recognizer for recognition, and multiple food image feature information is output, wherein each food image feature information includes color information. By dividing the food into regions and collecting images, and using a pre-trained food image recognizer to recognize the food image, multiple food image feature information is obtained, which provides basic data support for subsequent microbial growth prediction.

[0039] Further, the pre-trained food image recognizer, the present application includes:

[0040] According to the historical detection data of food, a set of sample food images is collected, and the colors in the sample food images are marked to obtain a set of sample food image feature information; the sample food image set and the sample food image feature information set are used as supervised training data, and a food image recognizer is trained until convergence based on a convolutional neural network.

[0041] Specifically, using the historical food inspection data, a large number of sample food image sets are collected. These images cover different types of food and color changes at different microbial growth stages, providing rich training samples for training image recognizers. Then, the color features in the sample food images are annotated one by one, that is, the color information of the food surface is annotated, such as dark red: RGB = (150, 20, 20), yellow: RGB = (200, 180, 60), gray brown: RGB = (130, 100, 90), etc., to obtain a sample food image feature information set, in which the sample food images and the sample food image feature information correspond one to one.

[0042] Then, a food image recognizer is constructed based on a convolutional neural network, wherein the convolutional neural network is a deep learning model specially used for processing image data, which is widely used in the field of computer vision, especially image processing. It can automatically extract spatial features in images through convolution operations and has strong recognition capabilities; the food image recognizer includes an input layer, a convolution layer, a pooling layer, a fully connected layer and an output layer, wherein the input data of the input layer is the food image, and the output data is the food image feature information (color type, etc.); then, the sample food image set and the sample food image feature information set are used as supervised training data to train the food image recognizer. First, the sample food image set is input into the convolutional neural network, and the corresponding The food image feature information is used as the target output label. The food image is propagated layer by layer through the convolution layer, activation function and pooling layer, and finally a prediction result is generated in the output layer. Then the error between the prediction result and the true label is calculated, and the commonly used loss function is the cross entropy loss. Further, according to the loss value, the convolution kernel weight and the fully connected layer weight are adjusted through the back propagation algorithm to gradually optimize the model and reduce the loss function value. Then, the sample food images are continuously input into the model, and the parameters are updated through forward propagation and back propagation. When the loss function value of the model reaches a low level and stabilizes and no longer decreases significantly, the model training converges, indicating that the food image recognizer has learned effective feature extraction capabilities and can accurately predict the microbial growth of new food images.

[0043] Further, according to the plurality of food image feature information, microbial growth category prediction and microbial growth scale prediction are performed to obtain a microbial category set and a plurality of first microbial growth scale information. The present application includes:

[0044] According to the historical detection data of food, a set of sample food image feature information is collected, and the microbial growth category and microbial growth scale under different sample food image feature information are collected and marked as a sample microbial category set and a sample first microbial growth scale information set; the sample food image feature information set is used as the prediction classification input, and the sample microbial category set and the sample first microbial growth scale information set are respectively used as the prediction classification output to construct a microbial category prediction classifier and a first microbial scale prediction classifier; using the microbial category prediction classifier and the first microbial scale prediction classifier, the multiple food image feature information is input into the prediction classification to obtain a microbial category set and multiple first microbial growth scale information.

[0045] Specifically, a set of sample food image feature information is collected based on historical food inspection data, and then the growth category and growth scale information of microorganisms in the food are collected based on the image features of each sample food and combined with the actual inspection data. For example, Pseudomonas is usually accompanied by gray or dark color changes, and Clostridium perfringens is usually associated with green or yellow pigment changes; the microbial growth scale corresponding to each sample food image is recorded, which is usually numerical information obtained by detecting the concentration of microorganisms or colony counts, such as the coverage area of ​​microorganisms, or obtained by experimental testing of bacterial concentrations, such as 1,000CFU / g-100,000CFU / g; for each sample food image, the corresponding microbial growth category and microbial growth scale are labeled according to its image features and experimental test results, to obtain a sample microbial category set and a sample first microbial growth scale information set.

[0046] Based on the BP neural network, a microorganism category prediction classifier and a first microorganism scale prediction classifier are constructed. The microorganism category prediction classifier and the first microorganism scale prediction classifier are deep learning models for predicting the microorganism category and growth scale in food images. The microorganism category prediction classifier includes an input layer, multiple hidden layers and an output layer, the input data of the input layer is food image feature information (RGB values, etc.), and the output data of the output layer is the microorganism category; the first microorganism scale prediction classifier also includes an input layer, multiple hidden layers and an output layer, the data of the input layer is food image feature information (RGB values, etc.), and the output data of the output layer is microorganism growth scale information (such as bacterial concentration).

[0047] Then, the sample food image feature information set is used as the prediction classification input, and the sample microorganism category set is used as the prediction classification output, and the microorganism category prediction classifier is supervised and trained. First, the sample food image feature information is input into the BP neural network, and the prediction result of the microorganism category is obtained through layer-by-layer calculation of the input layer and the hidden layer; then, the error between the prediction result and the actual label (sample microorganism category set) is calculated, usually using the cross entropy loss function for calculation; then, according to the loss function value, the weights and biases in the neural network are adjusted through the back propagation algorithm, so that the prediction result gradually approaches the true label; iterative training is performed, and the weights are continuously updated until the loss function value reaches a predetermined threshold, indicating that the model has converged, and a trained microorganism category prediction classifier is obtained.

[0048] The sample food image feature information set is used as the prediction classification input, and the sample first microbial growth scale information set is used as the prediction classification output, and the first microbial scale prediction classifier is supervised and trained. First, the sample food image feature information is input into the network, and after step-by-step calculation of the input layer and the hidden layer, the predicted value of the microbial growth scale is finally generated in the output layer; then the mean square error loss function is used to calculate the error between the model prediction value and the actual microbial growth scale. The mean square error is suitable for regression problems and can quantify the difference between the predicted value and the true value; then the gradient is calculated through the back propagation algorithm, the weights and biases in the network are adjusted, and the loss function value is gradually reduced; and the forward propagation, loss calculation and back propagation cycle process is repeated. By updating the network parameters each time, the loss value is continuously reduced, so that the model predicts the microbial growth scale more accurately, and the trained first microbial scale prediction classifier is obtained.

[0049] Finally, the multiple food image feature information is used as input data and input into the microbial category prediction classifier and the first microbial scale prediction classifier for classification prediction, multiple microbial categories and multiple first microbial growth scale information are output, and a union operation is performed on the multiple microbial categories to obtain a microbial category set.

[0050] Further, the food to be tested is imaged at a second angle to obtain a cling film image, which is identified to obtain cling film feature information, and microbial growth scale prediction is performed to obtain second microbial growth scale information. The present application includes:

[0051] The food to be inspected is imaged at a second angle to obtain a plastic wrap image; a plastic wrap image recognizer is pre-trained, wherein a set of sample plastic wrap images is collected, and the heights of plastic wrap swelling in different sample plastic wrap images are collected and marked as a sample plastic wrap feature information set, and the plastic wrap image recognizer is trained; the plastic wrap image is input and recognized by the plastic wrap image recognizer to obtain plastic wrap feature information; microbial growth scale prediction and classification are performed based on the plastic wrap feature information to obtain second microbial growth scale information, wherein the sample plastic wrap feature information set and the sample second microbial growth scale information set are collected, and a second microbial scale prediction classifier is constructed to perform prediction and classification.

[0052] Specifically, a second image acquisition angle is configured, and the second image acquisition angle is a 30° angle relative to the horizontal line, or a horizontal angle, wherein the second image acquisition angle can better capture detailed information on the surface of the plastic wrap, such as bulging characteristics. According to the second image acquisition angle, the CCD image sensor is used to capture an image of the food to be inspected to obtain an image of the plastic wrap.

[0053] Then, a cling film image recognizer is constructed based on a convolutional neural network, wherein the cling film image recognizer comprises an input layer, a convolutional layer, a pooling layer, a fully connected layer and an output layer, wherein the input data of the input layer of the cling film image recognizer is the cling film image, and the output data of the output layer is the swelling height of the cling film; then, a set of sample cling film images is collected, and the swelling heights of the cling film in different sample cling film images are collected and annotated as a set of sample cling film feature information, for example, the swelling height of the cling film is 0.5 mm, 0.2 mm, etc., wherein the sample cling film images and the sample cling film feature information correspond one to one; further, the sample cling film image set and the sample cling film feature information set are used to perform supervised training on the cling film image recognizer, firstly, the sample cling film image is input into the convolutional neural network, and ... In the neural network model, the features (such as swelling height) in the cling film image are extracted through the convolution layer and the pooling layer, and then the features are input into the fully connected layer for processing, and finally the predicted value of the swelling height of the cling film is generated in the output layer; then the mean square error loss function is used to calculate the error between the model prediction value and the actual swelling height of the cling film; further, according to the value of the loss function, the weights and biases in the convolutional neural network model are updated through the back propagation algorithm, and the commonly used optimization algorithms such as Adam or stochastic gradient descent can effectively adjust the network parameters and reduce the error; forward propagation, loss calculation and back propagation are continuously performed, and the prediction ability of the model is gradually improved by updating the network parameters each time until the loss function converges to a lower value; and a trained cling film image recognizer is obtained. Then the cling film image is input into the cling film image recognizer for recognition, and the cling film feature information (cling film swelling height) is output.

[0054] A second microbial scale prediction classifier is constructed based on the BP neural network. The second microbial scale prediction classifier is a BP neural network model that can be iteratively optimized in machine learning, including an input layer, multiple hidden layers and an output layer. The input data of the input layer of the second microbial scale prediction classifier is the characteristic information of the cling film (the expansion height of the cling film), and the output data of the output layer is the microbial growth scale information (microbial concentration). Then, the sample cling film feature information set and the sample second microbial growth scale information set are collected as training data, and the second microbial scale prediction classifier is supervised for training. First, after the input layer receives the cling film feature information, the data is processed by multiple hidden layers of nonlinear processing, and finally the predicted value of the microbial growth scale is generated in the output layer; then the mean square error loss function is used to measure the difference between the predicted microbial growth scale and the actual growth scale, and according to the loss function value, the gradient of each layer of the model is calculated by the back propagation algorithm, and the model is optimized by adjusting the weight and bias value of each connection; finally, the iterative process of forward propagation and back propagation is repeated, and each time the weight is adjusted, the prediction error of the model gradually decreases, and the model gradually learns the mapping relationship between the cling film feature information and the microbial growth scale until the convergence condition is met, and the second microbial scale prediction classifier is obtained. Then the cling film feature information is input into the second microbial scale prediction classifier for microbial growth scale prediction classification, and the second microbial growth scale information is output.

[0055] Further, according to the plurality of first microbial growth scale information and the second microbial growth scale information, a microbial growth scale analysis is performed to obtain a microbial growth scale, and the microbial category set is combined to output a microbial prediction result, and a detection strategy is formulated. The present application includes:

[0056] According to the plurality of first microbial growth scale information and the second microbial growth scale information, average microbial growth scale information is calculated; the average microbial growth scale information is compensated to obtain a microbial growth scale range; the microbial growth scale range and the microbial category set are combined to output a microbial prediction result; and a detection strategy is formulated according to the microbial prediction result.

[0057] Specifically, the mean of the plurality of first microbial growth scale information is calculated to obtain the first microbial growth scale mean information; then the first microbial growth scale mean information and the second microbial growth scale information are calculated again to obtain the average microbial growth scale information (such as microbial coverage area). Then the prediction errors of the two types of information (such as the accuracy of image recognition, the measurement error of the bulging height, etc.) are calculated, and the average microbial growth scale information is compensated according to the prediction error, and the microbial growth scale range is obtained according to the compensation result. The microbial growth scale range and the microbial category set are further used as microbial prediction results. Finally, a targeted detection strategy is formulated according to the microbial prediction results, that is, a suitable detection method is selected for each possible microbial category in combination with the microbial category set, wherein different microbial categories require different detection means, and the prediction results provide a reference for each type of microorganism, thereby rationally allocating detection resources, for example, for Pseudomonas, selective detection of culture medium can be used; for the protein decomposition ability of Clostridium perfringens, a specific detection means can be selected. Determine the frequency and depth of detection based on the scale of microbial growth. If the prediction results show that the scale of microbial growth is large, the number of tests needs to be increased to ensure that the active areas of the microorganisms are fully detected. If the scale of the microorganisms is small, the detection frequency can be reduced, and the detection range can be appropriately reduced to improve detection efficiency. Obtain an adaptive detection strategy.

[0058] By formulating an adaptive detection strategy based on the microbial prediction results and combining different microbial categories and growth scale information, the accuracy and specificity of food microbial detection can be significantly improved. At the same time, the detection means, number of times and detection limits can be flexibly adjusted to ensure efficient and scientific food microbial detection.

[0059] Further, the average microbial growth scale information is compensated to obtain the microbial growth scale range. The present application includes:

[0060] A second compensation weight is assigned to the second microbial growth scale information, and a plurality of first compensation weights are assigned to the plurality of first microbial growth scale information, wherein the weights of the plurality of first microbial growth scale information are assigned according to the distances between the plurality of collection area centers and the food center, and the distance between the collection area and the food center is positively correlated with the size of the first compensation weight; based on the data of predicting and detecting the microbial growth scale in historical time, a plurality of historical first microbial growth scale information sets and historical second microbial growth scale information sets are collected, and a historical actual microbial scale information set is obtained by collecting and obtaining a plurality of historical first scale accuracy parameter sets and historical second scale accuracy parameter sets, and the mean values ​​are calculated respectively. , obtain multiple average first scale accuracy parameters and second average scale accuracy parameters; randomly select multiple historical first scale accuracy parameters and historical second scale accuracy parameters from the multiple historical first scale accuracy parameter sets and the historical second scale accuracy parameter sets, calculate the error amplitudes with the multiple average first scale accuracy parameters and the second average scale accuracy parameters, and obtain multiple first error ranges and second error ranges; use the multiple first compensation weights and the second compensation weights to perform weighted calculation on the multiple first error ranges and the second error ranges to obtain the error range; use the error range to perform error compensation calculation on the average microbial growth scale information to obtain the microbial growth scale range.

[0061] Specifically, a second compensation weight is assigned to the second microbial growth scale information, wherein the second compensation weight is 0.5; a plurality of first compensation weights are assigned to the plurality of first microbial growth scale information, wherein the sum of the plurality of first compensation weights is 0.5. First, the distance between the center of each collection area and the center of the food is obtained, wherein microorganisms generally start to grow from the edge of the food, and areas farther from the center of the food are more likely to show early signs of microbial growth. Therefore, the farther the collection area is from the center of the food, the weight of its microbial growth scale should be appropriately increased; then, the weights of the plurality of first microbial growth scale information are assigned according to the distances between the plurality of collection area centers and the center of the food, that is, the distance between the collection area and the center of the food is positively correlated with the size of the first compensation weight, and the larger the distance, the larger the corresponding weight. For example, the sum of the distances between the plurality of collection area centers and the center of the food is calculated, and the ratio of each distance to the sum of the plurality of distances is used as a weight to obtain a plurality of first compensation weights, for example, a certain first compensation weight is 0.1.

[0062] Then, based on the data of predicting and detecting the scale of microbial growth in the historical time, multiple historical first microbial growth scale information sets (historical data sets for predicting the scale of microbial growth from the first image acquisition angle) and historical second microbial growth scale information sets (historical data sets for predicting the scale of microbial growth from the second image acquisition angle) are collected, and the historical actual microbial scale information sets are collected. Then, deviation calculation is performed based on multiple historical first microbial growth scale information sets, historical second microbial growth scale information sets, and historical actual microbial scale information sets, and then scale accuracy calculation is performed, for example, the accuracy is set to the inverse of the scale information deviation, wherein the smaller the deviation, the larger the scale accuracy parameter; multiple historical first scale accuracy parameter sets and historical second scale accuracy parameter sets are obtained; then, the mean of multiple historical first scale accuracy parameter sets and historical second scale accuracy parameter sets is calculated respectively, and multiple average first scale accuracy parameters and second average scale accuracy parameters are obtained.

[0063] For example, the historical first microbial growth scale information is 2000 CFU / g, and the historical actual microbial scale information is 2500 CFU / g. The deviation is the ratio of 500 CFU / g to 2500 CFU / g, which is 0.2. The calculated historical first scale accuracy parameter is the reciprocal of 0.2, which is 5.

[0064] Then, from the multiple historical first scale accuracy parameter sets and the historical second scale accuracy parameter sets, multiple historical first scale accuracy parameters and historical second scale accuracy parameters are randomly selected respectively, and then the error amplitudes of the multiple historical first scale accuracy parameters and the multiple average first scale accuracy parameters are calculated to obtain multiple first error amplitudes, wherein the error amplitude can be randomly selected multiple times and obtained by averaging to improve the accuracy and reliability of the error amplitude calculation, and the multiple first error amplitudes are used as first error ranges to obtain multiple first error ranges, wherein the error range represents the possible fluctuation range of the prediction result, reflecting the instability and error size of the prediction of a certain acquisition area or angle, for example, the average first scale accuracy parameter is 10, and the randomly selected historical first scale accuracy parameter is 9, then the first error amplitude is the ratio of the difference 1 and 10, and the error range is ±10%; on the other hand, the error amplitude of the historical second scale accuracy parameter and the second average scale accuracy parameter is calculated, for example, the ratio of the difference between the historical second scale accuracy parameter and the second average scale accuracy parameter to the second average scale accuracy parameter to obtain the second error amplitude, and the second error amplitude is used as the second error range.

[0065] Among them, the larger the error amplitude, the larger the error in predicting the microbial growth scale in the corresponding collection area or the second angle, and the larger the error range.

[0066] Then, the multiple first compensation weights and the second compensation weights are used to weightedly calculate the multiple first error ranges and the second error ranges, and the weighted calculation result is used as the final error range; finally, the error range is used to perform error compensation calculation on the average microbial growth scale information, that is, the average microbial growth scale information is added to the error range to obtain the microbial growth scale range. For example, assuming that the average microbial growth scale information is a microbial coverage area of ​​10 square centimeters and the error range is ±3%, the compensated microbial growth scale range is a microbial coverage area of ​​9.97 square centimeters to 10.03 square centimeters. By weightedly calculating the multiple first error ranges and the second error ranges, a comprehensive weighted error range is obtained, and then the error range is used to compensate the average microbial growth scale information, and finally a more accurate microbial growth scale range is obtained, which can not only improve the stability of the prediction, but also dynamically adjust the prediction results to cope with different error sizes, and provide a reference for subsequent microbial detection scheme decisions.

[0067] In summary, the food microbial detection method combined with microbial growth prediction provided by the present application has the following technical effects:

[0068] By capturing images of the food to be tested at a first angle, identifying regional food images in multiple acquisition areas, and obtaining multiple food image feature information; then predicting the microbial growth category and microbial growth scale based on the multiple food image feature information, and obtaining a microbial category set and multiple first microbial growth scale information; on the other hand, capturing images of the food to be tested at a second angle to obtain a plastic wrap image; identifying the plastic wrap image to obtain plastic wrap feature information, and predicting the microbial growth scale based on the plastic wrap feature information to obtain second microbial growth scale information; further, based on the multiple first microbial growth scale information and the second microbial growth scale information, performing microbial growth scale analysis to obtain a microbial growth scale range, and using the microbial growth scale range and the microbial category set as microbial prediction results; finally, formulating targeted detection strategies based on the microbial prediction results to perform food microbial detection; by combining technical means such as image recognition and intelligent prediction models, the breadth, efficiency and accuracy of microbial growth prediction under plastic wrap packaging can be improved, and the technical goal of efficiently predicting microbial growth trends can be achieved, thereby facilitating the formulation of more scientific and accurate detection strategies and improving microbial detection efficiency and quality.

[0069] Embodiment 2: Based on the food microorganism detection method combined with microorganism growth prediction in the previous embodiment, the present application also provides a food microorganism detection device combined with microorganism growth prediction, please refer to the attached Figure 3 ,include:

[0070] A food image feature information acquisition module 11, the food image feature information acquisition module 11 is used to capture images of the food to be detected at a first angle, obtain regional food images within multiple acquisition areas, perform identification, and obtain multiple food image feature information, wherein the food is food packaged in plastic wrap; a first growth scale information acquisition module 12, the first growth scale information acquisition module 12 is used to perform microbial growth category prediction and microbial growth scale prediction based on the multiple food image feature information, obtain a microbial category set and multiple first microbial growth scale information; a second growth scale information acquisition module 13, the second growth scale information acquisition module 13 is used to capture images of the food to be detected at a second angle, obtain a plastic wrap image, perform identification, obtain plastic wrap feature information, and perform microbial growth scale prediction to obtain second microbial growth scale information; a microbial prediction result output module 14, the microbial prediction result output module 14 is used to perform microbial growth scale analysis based on the multiple first microbial growth scale information and the second microbial growth scale information, obtain a microbial growth scale range, output a microbial prediction result in combination with the microbial category set, and formulate a detection strategy to perform food microbial detection.

[0071] Furthermore, the food microorganism detection device combined with microorganism growth prediction is also used for:

[0072] The food to be inspected is divided into regions at a first angle to obtain a plurality of acquisition regions; images are acquired from the plurality of acquisition regions to obtain a plurality of regional food images; a food image recognizer is pre-trained; and the plurality of regional food images are recognized using the food image recognizer to obtain a plurality of food image feature information, wherein each food image feature information includes color information.

[0073] Furthermore, the food microorganism detection device combined with microorganism growth prediction is also used for:

[0074] According to the historical detection data of food, a set of sample food images is collected, and the colors in the sample food images are marked to obtain a set of sample food image feature information; the sample food image set and the sample food image feature information set are used as supervised training data, and a food image recognizer is trained until convergence based on a convolutional neural network.

[0075] Furthermore, the food microorganism detection device combined with microorganism growth prediction is also used for:

[0076] According to the historical detection data of food, a set of sample food image feature information is collected, and the microbial growth category and microbial growth scale under different sample food image feature information are collected and marked as a sample microbial category set and a sample first microbial growth scale information set; the sample food image feature information set is used as the prediction classification input, and the sample microbial category set and the sample first microbial growth scale information set are respectively used as the prediction classification output to construct a microbial category prediction classifier and a first microbial scale prediction classifier; using the microbial category prediction classifier and the first microbial scale prediction classifier, the multiple food image feature information is input into the prediction classification to obtain a microbial category set and multiple first microbial growth scale information.

[0077] Furthermore, the food microorganism detection device combined with microorganism growth prediction is also used for:

[0078] The food to be inspected is imaged at a second angle to obtain a plastic wrap image; a plastic wrap image recognizer is pre-trained, wherein a set of sample plastic wrap images is collected, and the heights of plastic wrap swelling in different sample plastic wrap images are collected and marked as a sample plastic wrap feature information set, and the plastic wrap image recognizer is trained; the plastic wrap image is input and recognized by the plastic wrap image recognizer to obtain plastic wrap feature information; microbial growth scale prediction and classification are performed based on the plastic wrap feature information to obtain second microbial growth scale information, wherein the sample plastic wrap feature information set and the sample second microbial growth scale information set are collected, and a second microbial scale prediction classifier is constructed to perform prediction and classification.

[0079] Furthermore, the food microorganism detection device combined with microorganism growth prediction is also used for:

[0080] According to the plurality of first microbial growth scale information and the second microbial growth scale information, average microbial growth scale information is calculated; the average microbial growth scale information is compensated to obtain a microbial growth scale range; the microbial growth scale range and the microbial category set are combined to output a microbial prediction result; and a detection strategy is formulated according to the microbial prediction result.

[0081] Furthermore, the food microorganism detection device combined with microorganism growth prediction is also used for:

[0082] A second compensation weight is assigned to the second microbial growth scale information, and a plurality of first compensation weights are assigned to the plurality of first microbial growth scale information, wherein the weights of the plurality of first microbial growth scale information are assigned according to the distances between the plurality of collection area centers and the food center, and the distance between the collection area and the food center is positively correlated with the size of the first compensation weight; based on the data of predicting and detecting the microbial growth scale in historical time, a plurality of historical first microbial growth scale information sets and historical second microbial growth scale information sets are collected, and a historical actual microbial scale information set is obtained by collecting and obtaining a plurality of historical first scale accuracy parameter sets and historical second scale accuracy parameter sets, and the mean values ​​are calculated respectively. , obtain multiple average first scale accuracy parameters and second average scale accuracy parameters; randomly select multiple historical first scale accuracy parameters and historical second scale accuracy parameters from the multiple historical first scale accuracy parameter sets and the historical second scale accuracy parameter sets, calculate the error amplitudes with the multiple average first scale accuracy parameters and the second average scale accuracy parameters, and obtain multiple first error ranges and second error ranges; use the multiple first compensation weights and the second compensation weights to perform weighted calculation on the multiple first error ranges and the second error ranges to obtain the error range; use the error range to perform error compensation calculation on the average microbial growth scale information to obtain the microbial growth scale range.

[0083] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The food microorganism detection method combined with microorganism growth prediction and the specific examples in the aforementioned embodiment 1 are also applicable to the food microorganism detection device combined with microorganism growth prediction in this embodiment. Through the aforementioned detailed description of the food microorganism detection method combined with microorganism growth prediction, those skilled in the art can clearly know the food microorganism detection device combined with microorganism growth prediction in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0084] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0085] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A method for detecting food microorganisms in combination with prediction of microbial growth, characterized in that: Methods include: Capturing images of food to be inspected at a first angle, obtaining regional food images within a plurality of acquisition areas, and performing identification to obtain feature information of a plurality of food images, wherein the food is food packaged in a plastic wrap; According to the plurality of food image feature information, performing microbial growth category prediction and microbial growth scale prediction to obtain a microbial category set and a plurality of first microbial growth scale information; Capturing an image of the food to be tested at a second angle to obtain a plastic wrap image, performing recognition to obtain characteristic information of the plastic wrap, and predicting the growth scale of microorganisms to obtain second microorganism growth scale information; According to the plurality of first microbial growth scale information and the second microbial growth scale information, a microbial growth scale analysis is performed to obtain a microbial growth scale range, and the microbial category set is combined to output a microbial prediction result, and a detection strategy is formulated to perform food microbial detection.

2. The food microorganism detection method combined with microbial growth prediction according to claim 1, characterized in that: The food to be inspected is imaged at a first angle to obtain regional food images within a plurality of acquisition areas, and recognition is performed to obtain a plurality of food image feature information, including: Divide the food to be tested into regions at a first angle to obtain multiple collection regions; Performing image acquisition on the multiple acquisition areas to obtain multiple regional food images; Pre-trained food image recognizer; The food image identifier is used to identify the multiple regional food images to obtain multiple food image feature information, wherein each food image feature information includes color information.

3. The food microorganism detection method combined with microbial growth prediction according to claim 2, characterized in that: Pre-trained food image recognizer, including: According to the historical detection data of food, a set of sample food images is collected, and the colors in the sample food images are marked to obtain a set of sample food image feature information; The sample food image set and the sample food image feature information set are used as supervised training data, and a food image recognizer is trained until convergence based on a convolutional neural network.

4. The method for detecting food microorganisms in combination with prediction of microbial growth according to claim 1, characterized in that: According to the plurality of food image feature information, microorganism growth category prediction and microorganism growth scale prediction are performed to obtain a microorganism category set and a plurality of first microorganism growth scale information, including: According to the historical detection data of food, a sample food image feature information set is collected, and the microbial growth category and microbial growth scale under different sample food image feature information are collected, and marked as a sample microbial category set and a sample first microbial growth scale information set; Using the sample food image feature information set as the prediction classification input, and using the sample microorganism category set and the sample first microorganism growth scale information set as the prediction classification output, respectively, to construct a microorganism category prediction classifier and a first microorganism scale prediction classifier; The microorganism category prediction classifier and the first microorganism scale prediction classifier are used to input the plurality of food image feature information into the prediction classification to obtain a microorganism category set and a plurality of first microorganism growth scale information.

5. The food microorganism detection method combined with microorganism growth prediction according to claim 1, characterized in that: The food to be tested is imaged at a second angle to obtain a plastic wrap image, and is identified to obtain plastic wrap feature information, and microbial growth scale prediction is performed to obtain second microbial growth scale information, including: Capturing an image of the food to be inspected at a second angle to obtain a plastic wrap image; Pre-training a cling film image recognizer, wherein a set of sample cling film images is collected, and the heights of bulging cling films in different sample cling film images are collected and marked as a set of sample cling film feature information, and the cling film image recognizer is trained; Using the cling film image recognizer, inputting and recognizing the cling film image to obtain cling film feature information; According to the cling film characteristic information, microbial growth scale prediction and classification are performed to obtain second microbial growth scale information, wherein a sample cling film characteristic information set and a sample second microbial growth scale information set are collected, and a second microbial scale prediction classifier is constructed for prediction and classification.

6. The method for detecting food microorganisms in combination with prediction of microbial growth according to claim 1, characterized in that: According to the plurality of first microbial growth scale information and the second microbial growth scale information, a microbial growth scale analysis is performed to obtain a microbial growth scale, and in combination with the microbial category set, a microbial prediction result is outputted, and a detection strategy is formulated, including: Calculating and obtaining average microbial growth scale information according to the plurality of first microbial growth scale information and second microbial growth scale information; Compensating the average microbial growth scale information to obtain a microbial growth scale range; Outputting microbial prediction results by combining the microbial growth scale range and the microbial category set; A detection strategy is formulated based on the microbial prediction results.

7. The method for detecting food microorganisms in combination with prediction of microbial growth according to claim 6, characterized in that: The average microbial growth scale information is compensated to obtain a microbial growth scale range, including: Allocating a second compensation weight to the second microbial growth scale information, and allocating a plurality of first compensation weights to the plurality of first microbial growth scale information, wherein the weights of the plurality of first microbial growth scale information are allocated according to the distances between the centers of the plurality of collection areas and the centers of the food, and the distances between the collection areas and the centers of the food are positively correlated with the magnitudes of the first compensation weights; According to the data of predicting and detecting the growth scale of microorganisms in historical time, a plurality of historical first microorganism growth scale information sets and historical second microorganism growth scale information sets are collected, and a historical actual microorganism scale information set is collected, and a plurality of historical first scale accuracy parameter sets and historical second scale accuracy parameter sets are calculated, and the averages are calculated respectively to obtain a plurality of average first scale accuracy parameters and second average scale accuracy parameters; Randomly selecting a plurality of historical first scale accuracy parameters and a plurality of historical second scale accuracy parameters from the plurality of historical first scale accuracy parameter sets and the plurality of historical second scale accuracy parameter sets, respectively, and calculating the error margins with the plurality of average first scale accuracy parameters and the second average scale accuracy parameters to obtain a plurality of first error ranges and a second error range; Using the plurality of first compensation weights and the second compensation weights, weighted calculation is performed on the plurality of first error ranges and the second error ranges to obtain an error range; The error range is used to perform error compensation calculation on the average microbial growth scale information to obtain the microbial growth scale range.

8. A food microorganism detection device combined with microorganism growth prediction, characterized in that: The steps for implementing the food microbial detection method combined with microbial growth prediction as described in any one of claims 1 to 7 include: A food image feature information acquisition module, the food image feature information acquisition module is used to capture images of the food to be detected at a first angle, obtain regional food images within a plurality of acquisition areas, perform identification, and obtain a plurality of food image feature information, wherein the food is food packaged in a plastic wrap; A first growth scale information acquisition module, the first growth scale information acquisition module is used to perform microbial growth category prediction and microbial growth scale prediction according to the plurality of food image feature information, and obtain a microbial category set and a plurality of first microbial growth scale information; A second growth scale information acquisition module, which is used to collect images of the food to be inspected at a second angle, obtain a cling film image, perform recognition, obtain cling film feature information, and perform microbial growth scale prediction to obtain second microbial growth scale information; A microbial prediction result output module is used to perform microbial growth scale analysis based on the multiple first microbial growth scale information and the second microbial growth scale information, obtain the microbial growth scale range, output the microbial prediction result in combination with the microbial category set, and formulate a detection strategy to perform food microbial detection.

Images