Food processing safety supervision method, device and equipment and storage medium

Abstract

The invention discloses a food processing safety supervision method and device, equipment and a storage medium. The method comprises the following steps: acquiring respective corresponding category data acquired by a plurality of category sensors, wherein the plurality of category sensors are arranged in a food processing area; determining a food processing safety result of the food processing area based on the plurality of category data; and based on the food processing safety result, executing a food processing safety supervision system corresponding to the food processing safety result. According to the scheme, the food processing environment is monitored from multiple dimensions by collecting multi-sensor data, and the processing safety supervision strength is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of food safety, and specifically relates to a food processing safety supervision method, device, equipment and storage medium. Background Art

[0002] Food safety is directly related to the lives and health of the people and social stability. Food safety is the foundation for protecting public health and promoting harmonious social development. Ensuring that food is non-toxic and harmless and meets the nutritional requirements can effectively reduce the risk of foodborne diseases and improve people's quality of life.

[0003] Food factories are an important part of the food industry chain. Their food safety issues are related to the food safety of the entire industry chain. However, the current operating procedures and food safety in food factories cannot be guaranteed. There are inevitably various possible contamination behaviors in the food production line. For example, contaminated food raw materials are mistakenly added back into processed food, causing food products to be contaminated, which endangers the health of consumers. Therefore, effective safety supervision during the food processing process is crucial. Summary of the invention

[0004] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a food processing safety supervision method, device, equipment and storage medium to solve the problem of low food processing safety supervision in the prior art.

[0005] According to one aspect of the present application, a food processing safety supervision method is disclosed, the method comprising:

[0006] Acquire the respective corresponding category data acquired by a plurality of category sensors, wherein the plurality of category sensors are arranged in a food processing area;

[0007] Determining a food processing safety result of the food processing area based on a plurality of the category data;

[0008] Based on the food processing safety results, a food processing safety supervision system corresponding to the food processing safety results is implemented.

[0009] In some embodiments, the category sensor includes at least a camera sensor, the category data includes at least video information corresponding to the camera sensor, and the determining of the food processing safety information of the food processing area based on a plurality of the category data includes:

[0010] Inputting the plurality of category data into the corresponding prediction models respectively for prediction, and obtaining a plurality of prediction results, wherein each category data corresponds to one prediction result;

[0011] Based on the multiple prediction results, food processing safety results corresponding to each of the multiple prediction results are determined, wherein the food processing safety results at least include operation actions determined based on the video information, and the operation actions are based on operations performed by a target operator with a unique target identifier.

[0012] In some embodiments, determining the food processing safety result of the food processing area based on the plurality of category data comprises:

[0013] Performing data processing on the multiple categories of data to obtain multiple target category data, wherein the data processing includes format processing;

[0014] Performing multimodal fusion on a plurality of target category data to obtain a multimodal data set;

[0015] Model training is performed based on the multimodal data set to obtain a behavior recognition model, and the behavior recognition model is used to comprehensively recognize multi-category data to obtain food processing safety results for the food processing area.

[0016] In some embodiments, when the category sensor includes a camera sensor, before inputting the plurality of category data into the corresponding category prediction models for prediction, the method further includes:

[0017] Establishing a prediction model corresponding to each of the category data, wherein establishing a frame prediction model corresponding to the camera sensor includes:

[0018] Acquire diversified security anomaly data, wherein the security anomaly data includes security anomaly images and security anomaly videos;

[0019] Performing a first marking process on the safety abnormality image to obtain a first marked image, where the first marked image is an abnormal image associated with the food processing abnormality in the safety abnormality image;

[0020] Extracting and processing video frames of the security anomaly video to obtain target video frames;

[0021] Performing a second marking process on the target video frame to obtain a second marked image, where the second marked image is an associated frame image of the food processing abnormality in the target video frame;

[0022] Model training is performed based on the first labeled image and the second labeled image to obtain a frame prediction model corresponding to the camera sensor.

[0023] In some embodiments, extracting and processing the video frame of the security abnormality video to obtain the target video frame includes:

[0024] Extracting video frames from the abnormal security video to obtain abnormal video frames;

[0025] The abnormal video frame is cleaned and enhanced to obtain the target video frame.

[0026] In some embodiments, performing a first marking process on the security abnormality image to obtain a first marked image includes:

[0027] Based on target marking software or web page tools or automatic marking algorithms, target frames are marked on images containing food processing anomalies in the safety anomaly images.

[0028] In some embodiments, determining the food processing safety results corresponding to each of the plurality of prediction results based on the plurality of prediction results comprises:

[0029] Identifying target attribute information in the video information;

[0030] Comparing the target attribute information with the stored attribute information in the target database to determine an initial operator set;

[0031] Identify target identification information of each of the initial operators in the initial operator set;

[0032] A target operator associated with the food processing safety result is determined based on the target identification information.

[0033] According to another aspect of the present application, a food processing safety monitoring device is also disclosed, characterized in that the device comprises:

[0034] A category data acquisition module, used to acquire the category data corresponding to each other acquired by a plurality of category sensors, wherein the plurality of category sensors are arranged in a food processing area;

[0035] A food processing safety result determination module, used for determining a food processing safety result of the food processing area based on a plurality of the category data;

[0036] A safety supervision system execution module is used to execute the food processing safety supervision system corresponding to the food processing safety result based on the food processing safety result.

[0037] According to another aspect of the present application, an electronic device is also disclosed, which includes a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the electronic device executes each step of the food processing plant safety supervision method as described in any of the above items.

[0038] According to another aspect of the present application, a computer-readable storage medium is also disclosed, wherein instructions are stored on the computer-readable storage medium, wherein when the instructions are executed by a processor, each step of the food processing plant safety supervision method as described in any one of the above is implemented.

[0039] The present invention includes but is not limited to the following beneficial effects: (1) This solution monitors the food processing environment from multiple dimensions by integrating multi-sensor data, thereby improving the intensity of processing safety supervision; (2) This solution inputs multiple categories of data into their respective prediction models, and can perform specialized analysis based on different data features to improve prediction accuracy, and determine the target operator based on the extracted video frame information and unique target identification information, thereby improving the accuracy of determining personnel associated with food processing safety results; (3) Through multimodal fusion, data from different sensors are integrated into a multimodal data set, thereby fully utilizing the advantages of various types of sensors, improving the comprehensiveness and accuracy of data, and enabling the behavior recognition model trained based on the multimodal data set to comprehensively identify data of different categories and improve the ability to accurately identify operating behaviors; (4) By acquiring and processing diversified safety anomaly data, a more accurate and effective prediction model can be established. In particular, by extracting and processing key frames in safety anomaly videos, dynamic information can be effectively utilized to enhance the model's understanding and recognition capabilities of time series data, thereby improving the ability to identify and monitor safety risks in food processing areas, and providing stronger protection for food safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below.

[0041] Figure 1 is a flow chart of the food processing safety supervision method of the embodiment of the present application;

[0042] Figure 2 It is another flow chart of the food processing safety supervision method of the embodiment of the present application;

[0043] Figure 3 This is an example diagram of the principle of distinguishing operators of food processing according to an embodiment of the present application;

[0044] Figure 4 It is another flow chart of the food processing safety supervision method of the embodiment of the present application;

[0045] Figure 5 It is another flow chart of the food processing safety supervision method of the embodiment of the present application;

[0046] Figure 6It is another flow chart of the food processing safety supervision method of the embodiment of the present application;

[0047] Figure 7 is a structural block diagram of a food processing safety monitoring device according to an embodiment of the present application;

[0048] Figure 8 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The embodiment of the present invention discloses a method for monitoring food processing safety, the method comprising: obtaining corresponding category data obtained by multiple category sensors, the multiple category sensors being arranged in a food processing area; determining a food processing safety result of the food processing area based on the multiple category data; and executing a food processing safety monitoring system corresponding to the food processing safety result based on the food processing safety result. This solution monitors the food processing environment from multiple dimensions by integrating multi-sensor data, thereby improving the intensity of processing safety monitoring.

[0050] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0051] For ease of understanding, the specific process of the embodiment of the present invention is described below. Specifically, Figure 1 It is a flow chart of the food processing safety supervision method, specifically, Figure 1 As shown, the following steps are included:

[0052] S100: Obtaining category data corresponding to each of the plurality of category sensors.

[0053] Specifically, multiple categories of sensors may include but are not limited to temperature sensors, humidity sensors, gas sensors, pH sensors, motion sensors, etc., wherein temperature sensors are used to monitor temperature changes during food processing to provide feedback on whether the ingredients are stored and processed at an appropriate temperature. Humidity sensors are used to monitor the humidity of air and materials to provide feedback on the moisture content during food processing and provide effective data for the prediction of deterioration or mold. Gas sensors are used to detect the concentration of carbon dioxide, ammonia and other volatile gases to monitor gas changes during fermentation and provide data for the evaluation of the freshness of food. pH sensors are used to measure the acidity and alkalinity of food to provide feedback on whether the pH value during fermentation, pickling and other processes is within a safe range. Motion sensors are used to detect the movement of equipment and personnel, monitor the operating efficiency of the production line, and provide data for food processing safety events. Further, multiple categories of sensors are dispersedly arranged in the food processing area. For example, sensors can be arranged in key locations such as processing areas, storage areas, and packaging areas. According to the type of sensor, the appropriate height and angle are selected to ensure the best detection effect. Further, the data of the category sensors are transmitted to the central data processing system wirelessly or wired. Furthermore, the monitoring system can display the data of various sensors in real time to facilitate timely response.

[0054] By deploying various types of sensors in the food processing area, the food production process can be effectively monitored and managed.

[0055] S102. Determine food processing safety results of food processing areas based on multiple categories of data.

[0056] In one example, after obtaining multiple categories of data, the obtained category data can be cleaned first, such as removing outliers, that is, checking outliers in each category of data (such as erroneous data caused by sensor failure), and removing or correcting them. Or missing value processing, that is, processing missing data, can be performed, and interpolation, mean filling and other methods can be selected. Further, data standardization is performed, such as converting data from different sensors into a unified unit (such as temperature unified as degrees Celsius, humidity unified as a percentage, etc.), and ensuring that the timestamps of all data are consistent for time series analysis. Furthermore, through data visualization tools (such as line charts, bar charts, etc.), the changing trends of key indicators such as temperature and humidity are identified, and the current data is compared with the historical data to evaluate the safety of the current food processing environment as a food processing safety result.

[0057] It is understandable that in some examples, the safety results may include food processing as a safety normal event or food processing as a safety abnormal event, wherein the safety normal event or safety abnormal event may be based on standards or events such as relevant laws and regulations or standard documents or on-site visits to food processing plants or questionnaire surveys as reference standards for safety abnormal events or safety normal events. For example, some standard documents stipulate that the event of mixing old food with fresh food is considered a safety abnormal event.

[0058] S104. Based on the food processing safety results, implement the food processing safety supervision system corresponding to the food processing safety results.

[0059] In one example, in combination with the above examples, a food processing safety supervision system corresponding to the food processing safety results can be implemented based on the food processing safety results, such as adjusting temperature settings, increasing ventilation, etc.

[0060] Furthermore, Figure 2 is another flow chart of the food processing safety supervision method of the embodiment of the present application. Specifically, the method of the flow chart is an exemplary description of determining food processing safety information of a food processing area based on multiple category data when the category sensor includes at least a camera sensor and the category data includes at least video information corresponding to the camera sensor. Specifically, refer to Figure 2 , including the following steps:

[0061] S200, inputting multiple category data into their corresponding prediction models for prediction, and obtaining multiple prediction results.

[0062] It is understandable that each category of data corresponds to a prediction result. Each category of data is input into its corresponding prediction model for prediction to obtain the prediction result corresponding to the category of data. For example, the acquired video information is directly input into the frame prediction model corresponding to the pre-trained camera sensor for prediction to obtain the frame anomaly prediction result.

[0063] S202. Based on the multiple prediction results, determine the food processing safety results corresponding to each of the multiple prediction results.

[0064] Among them, the food processing safety results at least include frame anomaly prediction results, that is, frame images containing operating actions determined based on video information, and the operating actions are based on the operations of target operators with unique target identifiers.

[0065] It is understandable that in the present application, operators can be distinguished in advance and marked with unique identification. Figure 3As shown, the distinction can be made based on the nature of the work: for example, the dress distinction can be made according to the nature of the work of the operators in the food processing plant, including but not limited to the dress distinction based on different working methods, such as production line operators, quality inspectors, etc., based on different operating spaces, such as workshops, warehouses, laboratories, etc., based on different operating properties, such as the dress distinction between personnel who directly contact food and those who do not directly contact food, based on different technologies, such as the dress distinction between personnel involved in the use of different equipment or technologies, etc. For example, production line operators wear blue, inspectors wear green, etc. Further, the unique identification mark can be a unique number assigned to each employee, and the numbering method can be "capital first letter of the nature + number". For example, the production line operator can be marked as "O1" and the inspector can be marked as "Q1". The number can be printed on a conspicuous position of the work clothes, such as the chest, back, etc., or it can be deployed in multiple parts such as the chest, back, hat, etc. to ensure easy identification in the work environment. Furthermore, in some other examples, in order to facilitate the recognition of the camera sensor, a mark with identification characteristics can be designed for each operator. These marks can include but are not limited to specific shapes or patterns, company or department logos, stripes or patterns of different colors, QR codes printed on work clothes, etc. Among them, the QR code printed on the work clothes can be linked to the employee's detailed information, training records, etc., for quick query. It can be understood that through the above-mentioned marks and color distinctions, the system can quickly identify different individuals and their work nature in reading the acquired video information, thereby improving management efficiency.

[0066] It can be understood that the above-mentioned embodiment inputs multiple categories of data into respective prediction models, and can perform specialized analysis based on different data features to improve prediction accuracy, and determine the target operator based on the extracted video frame information and unique target identification information, thereby improving the accuracy of determining personnel associated with food processing safety results.

[0067] Furthermore, the prediction model is pre-trained, in one example, Figure 4 As shown, the steps for building a frame prediction model are introduced:

[0068] S400. Obtain diversified security exception data.

[0069] In one example, food processing safety warning videos or materials can be collected through online collection, field scene interpretation, algorithm generation and other methods, and illegal operations or food safety hazards that occur during processing provided by the factory can be collected. Video data sets can be obtained by performing on-site food processing unsafe scene interpretation and collection in the factory; table data sets can be obtained through operation data or sensor data provided by the factory when illegal operations or food safety hazards occur. Additional video data sets can be obtained through algorithms such as GAN (generative adversarial network), and additional table data sets can be obtained through algorithms such as SMOTE. The above collected safety anomaly data include but are not limited to video files in formats such as JPG, MP4, AVI, and table files in formats such as CSV, TXT, XLSX, etc.

[0070] S402: Perform a first marking process on the security abnormality image to obtain a first marked image.

[0071] Specifically, the first marking process is performed on the safety abnormality image to obtain the first marked image, which can be specifically marking the abnormal images associated with the food processing abnormality from all the safety abnormality images. The marking process can be marking by box selection or the like. The marking process can be implemented by marking the image using image marking software or scripts. The images with food processing abnormalities in the safety abnormality images can also be marked with target boxes based on target marking software or webpage tools or automatic marking algorithms.

[0072] S404: extract and process video frames of the security anomaly video to obtain target video frames.

[0073] Specifically, the safety accident video is segmented frame by frame to convert the video into separate images. Further, data processing is performed on the converted multiple frames of images, that is, target video frames are obtained after removing low-quality images.

[0074] S406: Perform a second marking process on the target video frame to obtain a second marked image.

[0075] Specifically, the target video frame is subjected to the second marking process to obtain the second marked image, which may be specifically a marking process of abnormal images associated with food processing abnormalities from all target video frames, and the marking process may be a marking such as box selection. The marking process may be implemented by marking the image using image marking software or scripts. The images with food processing abnormalities in the safety abnormal images may also be marked with target boxes based on target marking software or webpage tools or automatic marking algorithms.

[0076] S408 . Perform model training based on the first labeled image and the second labeled image to obtain a frame prediction model corresponding to the camera sensor.

[0077] Specifically, after obtaining the first labeled image and the second labeled image, data of the first labeled image and the second labeled image are divided to obtain a training set and a validation set, and then the training set is input into a self-built model or an existing model is retrained. The existing models here include but are not limited to Resnet, YOLO, DeepLabV3+, and Transformer models.

[0078] It can be understood that steps S400-S408 are only an illustration of the frame prediction model. Each prediction model needs to be pre-trained, that is, based on multi-category data, it can be divided into different categories according to different regulations, and can be trained separately according to different illegal operations, and different weight files can be generated. Through this step, the model can combine different weight files when performing action prediction to form diverse and personalized supervision plans for illegal operations and dangerous operations, thereby meeting the needs of food processing plants with different functions.

[0079] Furthermore, if the security anomaly data includes table files in formats such as CSV, TXT, XLSX, etc., it is necessary to preprocess the obtained CSV, TXT, XLSX, etc. format files. Specifically, the preprocessing steps include: cleaning the data and removing low-quality data. The data when the violation occurs can be marked simply by 0 and 1 classification, marking the data when the violation occurs as 1, and marking the data when no violation occurs as 0. After the data is cleaned and marked, the processed data is placed in a self-built model or the existing model is retrained. The model here includes but is not limited to a transformer model, a BERT-based model, a deep neural network, a convolutional neural network, a logistic regression, a support vector machine, a random forest, a gradient boosting tree, a K-nearest neighbor algorithm, a linear discriminant analysis, a decision tree, a naive Bayes or a perceptron. Thus, a prediction model is established for the sample data based on the table format.

[0080] Further, such as Figure 5 As shown, Figure 5 is another flow chart of the food processing safety supervision method of the embodiment of the present application, which is another exemplary description of step S102, determining the food processing safety result of the food processing area based on multiple categories of data. For details, refer to Figure 5 , including the following steps:

[0081] S500: Process multiple categories of data to obtain multiple target category data.

[0082] The multiple categories of data include, but are not limited to, sound data in a food processing plant environment collected by a sound collector installed in the food factory. These sound data may include, but are not limited to, equipment operation sounds, personnel conversation sounds, alarm sounds, etc. The category data may also include video data, temperature data, etc. Among them, data processing of multiple categories of data may be labeling the collected sound data to distinguish normal sounds from abnormal sounds. For example, labeling normal equipment operation sounds and abnormal equipment failure sounds. Extracting the sound data, for example, using Fourier transform and other methods to convert the sound signal into a spectrum, extracting sound features such as frequency components, extracting security event-associated image frames from the video data, for example, using deep learning models such as convolutional neural networks (CNN) to extract features of video frames and obtain image features.

[0083] S502: Perform multimodal fusion on multiple target category data to obtain a multimodal data set.

[0084] Specifically, the sound features are combined with data collected by other sensors (such as video surveillance, temperature sensors, etc.) to form a multimodal data set. The multimodal data set may include image or video data at a specific time point, the sound features corresponding to the video frame, and information such as temperature and humidity corresponding to the video frame.

[0085] S504. Perform model training based on the multimodal data set to obtain a behavior recognition model, where the behavior recognition model is used to perform comprehensive recognition on multi-category data to obtain food processing safety results in the food processing area.

[0086] Specifically, a suitable multimodal learning model is selected, for example, a deep learning model (such as CNN, RNN, etc.) is used to process image and sound data. During the training process, the sound features are input into the deep learning model together with the video data, and the model is trained to recognize different behavior features to obtain a behavior recognition model, so that the behavior recognition model can learn the correlation between sound and video, thereby improving the ability to recognize abnormal behavior.

[0087] Furthermore, Figure 6 is another flow chart of the food processing safety supervision method of the embodiment of the present application, and this step is an exemplary description of step S104. Figure 6 As shown, the following steps are included:

[0088] S600: Identify target attribute information in video information.

[0089] The target attribute information refers to the characteristics and status of a specific object identified in the video, including but not limited to the type, color, shape, size, position, motion status, etc. of the target. In the food factory scene, the target may be an operator, equipment, food product, etc.

[0090] S602: Compare the target attribute information with the stored attribute information in the target database to determine an initial operator set.

[0091] Specifically, by comparing the target attribute information with the stored attribute information in the target database, the attribute information in the target database that matches the target attribute information can be determined, and then the initial set of operators corresponding to the attribute information can be determined. The attribute information matches the target attribute information, which can mean that the target type, color, location, etc. are similar. By matching the target attribute information first, some obviously irrelevant operators can be screened out, narrowing the scope of operators and reducing the difficulty of identification.

[0092] S604: Identify target identification information of each initial operator in the initial operator set.

[0093] Exemplarily, based on the above introduction, each operator has unique identification information. In this step, the target identification information corresponding to each initial operator in the initial operator set is identified. It is not necessary to identify from all video frame images, which narrows the scope of personnel identification and improves identification efficiency.

[0094] S606. Determine the target operator associated with the food processing safety result based on the target identification information.

[0095] Specifically, since each operator has unique identification information, after identifying the target identification information, the action behavior information of each initial operator is determined, so as to determine whether the initial operator threatens food processing safety, and then identify the operator who threatens food processing safety as the target operator associated with the food processing safety result.

[0096] Furthermore, Figure 7 The structural diagram of the food processing safety monitoring device is as follows: Figure 7 As shown, the device comprises:

[0097] A category data acquisition module, used to acquire the category data corresponding to each other acquired by a plurality of category sensors, wherein the plurality of category sensors are arranged in a food processing area;

[0098] A food processing safety result determination module, used for determining a food processing safety result of a food processing area based on multiple category data;

[0099] The safety supervision system execution module is used to execute the food processing safety supervision system corresponding to the food processing safety results based on the food processing safety results.

[0100] The application introduction of the relevant modules of the device in this example can refer to the relevant introduction of the principle of the above method, which will not be repeated here.

[0101] According to another aspect of the present application, the present application also discloses an electronic device, which includes a memory and at least one processor, wherein instructions are stored in the memory; at least one processor calls the instructions in the memory to enable the electronic device to execute each step of the above-mentioned food processing safety supervision method.

[0102] The above figure describes in detail the food processing safety monitoring device in the embodiment of the present invention from the perspective of modular functional entities, and the following figure describes in detail the electronic device in the embodiment of the present invention from the perspective of hardware processing.

[0103] Figure 8 8 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device 800 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 810 (for example, one or more processors) and a memory 820, and one or more storage media 830 (for example, one or more mass storage devices) storing application programs 833 or data 832. Among them, the memory 820 and the storage medium 830 can be short-term storage or permanent storage. The program stored in the storage medium 830 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the electronic device 800. Furthermore, the processor 810 can be configured to communicate with the storage medium 830 to execute a series of instruction operations in the storage medium 830 on the electronic device 800.

[0104] The electronic device 800 may also include one or more power supplies 840, one or more wired or wireless network interfaces 850, one or more input and output interfaces 860, and / or one or more operating systems 831, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Figure 8 The structure of the electronic device shown does not constitute a limitation on the electronic device, and may include more or less components than shown in the figure, or combine some components, or arrange the components differently.

[0105] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the food processing safety supervision method.

[0106] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device, or unit can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0107] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0108] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A food processing safety supervision method, characterized in that: The method comprises: Acquire the respective corresponding category data acquired by a plurality of category sensors, wherein the plurality of category sensors are arranged in a food processing area; Determining a food processing safety result of the food processing area based on a plurality of the category data; Based on the food processing safety results, a food processing safety supervision system corresponding to the food processing safety results is implemented.

2. The food processing safety supervision method according to claim 1, characterized in that: The category sensor includes at least a camera sensor, the category data includes at least video information corresponding to the camera sensor, and the food processing safety information of the food processing area is determined based on a plurality of the category data, including: Inputting the plurality of category data into the corresponding prediction models respectively for prediction, and obtaining a plurality of prediction results, wherein each category data corresponds to one prediction result; Based on the multiple prediction results, food processing safety results corresponding to each of the multiple prediction results are determined, wherein the food processing safety results at least include operation actions determined based on the video information, and the operation actions are based on operations performed by a target operator with a unique target identifier.

3. The food processing safety supervision method according to claim 1, characterized in that: Determining the food processing safety result of the food processing area based on the plurality of category data comprises: Performing data processing on the multiple category data to obtain multiple target category data, wherein the data processing includes format processing; Performing multimodal fusion on a plurality of target category data to obtain a multimodal data set; Model training is performed based on the multimodal data set to obtain a behavior recognition model, and the behavior recognition model is used to comprehensively recognize multi-category data to obtain food processing safety results for the food processing area.

4. The food processing safety supervision method according to claim 2, characterized in that: When the category sensor includes a camera sensor, before inputting the plurality of category data into the corresponding category prediction models for prediction, the method further includes: Establishing a prediction model corresponding to each of the category data, wherein establishing a frame prediction model corresponding to the camera sensor includes: Acquire diversified security anomaly data, wherein the security anomaly data includes security anomaly images and security anomaly videos; Performing a first marking process on the safety abnormality image to obtain a first marked image, where the first marked image is an abnormal image associated with the food processing abnormality in the safety abnormality image; Extracting and processing video frames of the security anomaly video to obtain target video frames; Performing a second marking process on the target video frame to obtain a second marked image, where the second marked image is an associated frame image of the food processing abnormality in the target video frame; Model training is performed based on the first labeled image and the second labeled image to obtain a frame prediction model corresponding to the camera sensor.

5. The food processing safety supervision method according to claim 4, characterized in that: The extracting and processing the video frame of the security abnormality video to obtain the target video frame includes: Extracting video frames from the abnormal security video to obtain abnormal video frames; The abnormal video frame is cleaned and enhanced to obtain the target video frame.

6. The food processing safety supervision method according to claim 4, characterized in that: The performing a first marking process on the security abnormality image to obtain a first marked image comprises: Based on target marking software or web page tools or automatic marking algorithms, target frames are marked on images containing food processing anomalies in the safety anomaly images.

7. The food processing safety supervision method according to claim 2, characterized in that: Determining the food processing safety results corresponding to each of the plurality of prediction results based on the plurality of prediction results includes: Identifying target attribute information in the video information; Comparing the target attribute information with the stored attribute information in the target database to determine an initial operator set; Identify target identification information of each of the initial operators in the initial operator set; A target operator associated with the food processing safety result is determined based on the target identification information.

8. A food processing safety monitoring device, characterized in that: The device comprises: A category data acquisition module, used to acquire the category data corresponding to each other acquired by a plurality of category sensors, wherein the plurality of category sensors are arranged in a food processing area; A food processing safety result determination module, used for determining a food processing safety result of the food processing area based on a plurality of the category data; A safety supervision system execution module is used to execute the food processing safety supervision system corresponding to the food processing safety result based on the food processing safety result.

9. An electronic device, characterized in that: The electronic device includes a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the electronic device executes each step of the food processing plant safety supervision method as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the various steps of the food processing plant safety supervision method as described in any one of claims 1-7 are implemented.

Images