Food storage early warning method and system based on artificial intelligence

Abstract

The invention provides a food storage early warning method and system based on artificial intelligence. The method comprises the following steps: acquiring an internal shot image of a target refrigerator, a multi-source sensing signal and the putting time of each food; performing image denoising and enhancement processing on the internal shot image to obtain a pre-processed image; performing image recognition according to the preprocessed image to obtain food information of each type of food; the food information comprises food types, food numbers and food colors; performing signal analysis on the multi-source sensing signal to obtain temperature information, humidity information and gas information; inputting the putting time, the temperature information, the humidity information, the gas information and the food information into a preset food freshness prediction model to obtain a prediction result; and performing early warning according to the prediction result. The food freshness can be intelligently monitored, early warning can be given out in time, and therefore food waste is effectively reduced, and the food management efficiency and safety are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of food storage, and in particular to an artificial intelligence-based food storage early warning method and system. Background Art

[0002] With the acceleration of modern life and people's attention to food safety and health, the intelligent degree of household refrigerators as important equipment for food storage has gradually become the focus of consumers. However, traditional refrigerators have many shortcomings in food management, which are mainly reflected in the following aspects:

[0003] 1. Traditional refrigerators are usually unable to monitor the status of food inside in real time. Users can only rely on memory or regular inspections to judge the freshness and shelf life of food. This method not only easily leads to food expiration or deterioration, but also may cause unnecessary food waste.

[0004] 2. When using the refrigerator, users often cannot obtain accurate and timely information about the storage status of food. For example, information such as the storage time and shelf life of food is difficult to be effectively recorded and reminded, resulting in users lacking necessary reference when using the refrigerator.

[0005] 3. Environmental factors such as temperature and humidity inside the refrigerator are crucial to the preservation of food. Traditional refrigerators usually lack real-time monitoring of these environmental parameters and cannot adjust the operating status in time to adapt to the storage needs of different foods.

[0006] 4. Due to the lack of effective management and reminder mechanisms, food often expires or spoils in households, resulting in a large amount of food waste. This not only puts a burden on the family economy, but also has a negative impact on the environment.

[0007] Therefore, there is an urgent need for an artificial intelligence-based food storage early warning method to improve the intelligence level of refrigerators, achieve efficient food management, reduce food waste, and ensure household food safety. Summary of the invention

[0008] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a food storage warning method and system based on artificial intelligence, which can intelligently monitor the freshness of food and issue warnings in time by acquiring and analyzing the images and environmental data of food inside the refrigerator in real time, thereby effectively reducing food waste and improving the efficiency and safety of food management.

[0009] To achieve the above object, the present invention provides the following solutions:

[0010] A food storage early warning method based on artificial intelligence, comprising:

[0011] Obtaining internal images of the target refrigerator, multi-source sensor signals, and the time when each food was placed in the refrigerator;

[0012] Performing image denoising and enhancement processing on the internally captured image to obtain a preprocessed image;

[0013] Performing image recognition based on the preprocessed image to obtain food information of various types of food; the food information includes: food type, food quantity and food color;

[0014] Performing signal analysis on the multi-source sensor signals to obtain temperature information, humidity information and gas information;

[0015] Inputting the placement time, the temperature information, the humidity information, the gas information and the food information into a preset food freshness prediction model to obtain a prediction result;

[0016] An early warning is issued based on the prediction results.

[0017] Preferably, performing image denoising and enhancement processing on the internally captured image to obtain a pre-processed image includes:

[0018] Performing filtering processing on the internal captured image to obtain a filtered image;

[0019] Performing color difference balance on the filtered image to obtain a color difference balanced image;

[0020] Construct an image enhancement window, and use the image enhancement window to traverse the image after chromatic aberration balance to perform denoising, and obtain a denoised image;

[0021] The denoised image is subjected to histogram equalization processing to obtain the pre-processed image.

[0022] Preferably, filtering the internal captured image to obtain a filtered image comprises:

[0023] Performing wavelet decomposition on the internal captured image at multiple scales to obtain corresponding wavelet coefficients;

[0024] constructing a filtering threshold according to the size and decomposition scale of the internal captured image;

[0025] constructing an image filtering model using the filtering threshold;

[0026] The internal captured image is filtered using the image filtering model to obtain the filtered image.

[0027] Preferably, performing color balance on the filtered image to obtain a color-balanced image includes:

[0028] Calculate the pixel mean of the filtered image in the three RGB channels; wherein the pixel mean calculation formula is: Where D represents the pixel mean, represents the pixel mean of the filtered image in the red channel, represents the pixel mean of the filtered image in the green channel, Represents the pixel mean of the filtered image in the blue channel;

[0029] Calculating a balance coefficient for each color channel according to the pixel mean;

[0030] The balancing coefficient is used to perform color balance on the filtered image to obtain a color-balanced image.

[0031] Preferably, performing color difference balance on the filtered image using the balance coefficient to obtain a color difference balanced image includes:

[0032] Using formula Perform color difference balance on the filtered image to obtain a color difference balanced image; wherein K R Indicates the balance coefficient of the red channel, K G Indicates the balance coefficient of the green channel, K B Represents the balance coefficient of the blue channel, R′ represents the pixel value of the red channel after balance, G′ represents the pixel value of the green channel after balance, B′ represents the pixel value of the blue channel after balance, R represents the pixel value of the filtered image in the red channel, G represents the pixel value of the filtered image in the green channel, and B represents the pixel value of the filtered image in the blue channel.

[0033] Preferably, image recognition is performed based on the preprocessed image to obtain food information of various types of food, including:

[0034] Using a convolutional neural network to extract features from the preprocessed image to obtain visual features of different foods;

[0035] Inputting the visual features into a preset food classification model to classify the food in the image to obtain the category of each food;

[0036] Identifying the boundaries of each food in the visual features by an image segmentation method, and calculating the number of food instances of the same category to determine the number of each food;

[0037] Calculating the main color tone of each food using a color histogram extraction method, and determining the color of the food according to the main color tone;

[0038] The identified food types, food quantities and food colors are collected to construct a food information list.

[0039] Preferably, the method for constructing the food freshness prediction model includes:

[0040] Constructing an initial convolutional neural network; the placement time, the temperature information, the humidity information, the gas information and the food information;

[0041] The initial convolutional neural network is trained according to the sample placement time data set, the sample temperature information data set, the sample humidity information data set, the sample gas information data set and the sample food information data set to obtain a trained first classifier, a second classifier, a third classifier, a fourth classifier and a fifth classifier;

[0042] Cascading the trained first classifier, second classifier, third classifier, fourth classifier and fifth classifier to obtain a classification network;

[0043] Connecting the trained LSTM neural network after the classification network to obtain a CNN-LSTM combined neural network model;

[0044] The placement time, the temperature information, the humidity information, the gas information and the food information are predicted according to the CNN-LSTM combined neural network model to obtain the prediction result.

[0045] An artificial intelligence-based food storage early warning system, comprising:

[0046] A data acquisition module is used to acquire internal images of the target refrigerator, multi-source sensor signals, and the time when each food is put in;

[0047] An image preprocessing module, used for performing image denoising and enhancement processing on the internally captured image to obtain a preprocessed image;

[0048] An image recognition module, used to perform image recognition based on the pre-processed image to obtain food information of various types of food; the food information includes: food type, food quantity and food color;

[0049] A signal preprocessing module, used for performing signal analysis on the multi-source sensor signals to obtain temperature information, humidity information and gas information;

[0050] A prediction module, used for inputting the placement time, the temperature information, the humidity information, the gas information and the food information into a preset food freshness prediction model to obtain a prediction result;

[0051] The early warning module is used to issue an early warning according to the prediction result.

[0052] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0053] The present invention provides a food storage early warning method and system based on artificial intelligence, the method comprising: obtaining an internal photographic image of a target refrigerator, a multi-source sensor signal, and the time when each food is put in; performing image denoising and enhancement processing on the internal photographic image to obtain a pre-processed image; performing image recognition according to the pre-processed image to obtain food information of each type of food; the food information includes: food type, number of foods, and food color; performing signal analysis on the multi-source sensor signal to obtain temperature information, humidity information, and gas information; inputting the time when the food is put in, the temperature information, the humidity information, the gas information, and the food information into a preset food freshness prediction model to obtain a prediction result; and issuing an early warning according to the prediction result. The present invention can intelligently monitor the freshness of food and issue an early warning in a timely manner by acquiring and analyzing the image and environmental data of food inside the refrigerator in real time, thereby effectively reducing food waste and improving the efficiency and safety of food management. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] 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 will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0055] Figure 1 A flow chart of a method provided by an embodiment of the present invention;

[0056] Figure 2 A schematic diagram of the system structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] The purpose of the present invention is to provide a food storage early warning method and system based on artificial intelligence. By acquiring and analyzing the images and environmental data of food inside the refrigerator in real time, it can intelligently monitor the freshness of food and issue early warnings in time, thereby effectively reducing food waste and improving the efficiency and safety of food management.

[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] Figure 1 A flow chart of a method provided by an embodiment of the present invention, such as Figure 1 As shown, the present invention provides a food storage early warning method based on artificial intelligence, comprising:

[0061] Step 100: Acquire the internal image of the target refrigerator, multi-source sensor signals, and the time when each food is put in;

[0062] Step 200: performing image denoising and enhancement processing on the internally captured image to obtain a pre-processed image;

[0063] Step 300: Perform image recognition based on the preprocessed image to obtain food information of various types of food; the food information includes: food type, food quantity and food color;

[0064] Step 400: performing signal analysis on multi-source sensor signals to obtain temperature information, humidity information and gas information;

[0065] Step 500: input the placement time, temperature information, humidity information, gas information and food information into a preset food freshness prediction model to obtain a prediction result;

[0066] Step 600: Issue an early warning based on the prediction results.

[0067] Specifically, step 100 of this embodiment includes:

[0068] 1. Obtain images of the interior of the target refrigerator. Install a high-definition camera (such as a USB camera or a webcam) inside the refrigerator to ensure that it covers the main storage area of ​​the refrigerator. The camera has night vision so that it can capture clear images even in low-light environments. Set up a timed shooting mechanism to automatically capture images of the interior of the refrigerator regularly (such as every 10 minutes). Use motion detection technology to only capture images when the refrigerator door is open or an object moves to save storage space and improve efficiency. Upload the captured images to cloud storage or local servers in real time to ensure data security and accessibility. Use file naming conventions (such as "Refrigerator ID_Timestamp.jpg") to facilitate subsequent retrieval and management.

[0069] 2. Obtain multi-source sensor signals and install multiple sensors inside the refrigerator, including:

[0070] Temperature sensor: used to monitor the temperature inside the refrigerator to ensure it is within a safe range.

[0071] Humidity sensor: used to monitor the humidity in the refrigerator to prevent food from spoiling due to excessive humidity.

[0072] Gas sensor: used to detect the gas composition in the refrigerator (such as ethylene, ammonia, etc.) to determine the freshness and deterioration of food.

[0073] Use a microcontroller (such as Arduino or Raspberry Pi) to connect various sensors and read sensor data regularly. Set the data collection frequency (such as once a minute) and transmit the data to the central processing unit in real time. Send the sensor data to the cloud or local server via Wi-Fi or Bluetooth module for storage and processing.

[0074] 3. Obtain the time when each food was put into the refrigerator. When the user puts the food into the refrigerator, provide a user interface (such as a mobile phone APP or refrigerator display) to allow the user to enter the name of the food and the time it was put into the refrigerator. You can also paste a QR code label on the food packaging, and the user will automatically record the time of putting it into the refrigerator after scanning. The recorded time of putting the food into the refrigerator is stored in the database together with the food information (such as name, type, etc.) to ensure the integrity and traceability of the information.

[0075] Through the above scheme, this embodiment can realize real-time image acquisition of the interior of the target refrigerator, acquisition of multi-source sensor signals, and recording of the time when each food is put in. This process combines image processing, sensor technology, and user interaction design, providing a reliable data basis for subsequent food freshness prediction and early warning.

[0076] Preferably, performing image denoising and enhancement processing on the internally captured image to obtain a pre-processed image includes:

[0077] Performing filtering processing on the internal captured image to obtain a filtered image;

[0078] Performing color difference balance on the filtered image to obtain a color difference balanced image;

[0079] Construct an image enhancement window, and use the image enhancement window to traverse the image after chromatic aberration balance to perform denoising, and obtain a denoised image;

[0080] The denoised image is subjected to histogram equalization processing to obtain the pre-processed image.

[0081] Preferably, filtering the internal captured image to obtain a filtered image comprises:

[0082] Performing wavelet decomposition on the internal captured image at multiple scales to obtain corresponding wavelet coefficients;

[0083] constructing a filtering threshold according to the size and decomposition scale of the internal captured image;

[0084] constructing an image filtering model using the filtering threshold;

[0085] The internal captured image is filtered using the image filtering model to obtain the filtered image.

[0086] Specifically, this embodiment uses an image filtering model to perform a self-adjusting threshold transformation on the wavelet coefficients at each scale, which can highlight the detailed features of the image at different decomposition levels, suppress the image noise, and enhance the layering of the image.

[0087] Preferably, performing color balance on the filtered image to obtain a color-balanced image includes:

[0088] Calculate the pixel mean of the filtered image in the three RGB channels; wherein the pixel mean calculation formula is: Where D represents the pixel mean, represents the pixel mean of the filtered image in the red channel, represents the pixel mean of the filtered image in the green channel, Represents the pixel mean of the filtered image in the blue channel;

[0089] Calculating a balance coefficient for each color channel according to the pixel mean;

[0090] The balancing coefficient is used to perform color balance on the filtered image to obtain a color-balanced image.

[0091] Preferably, performing color difference balance on the filtered image using the balance coefficient to obtain a color difference balanced image includes:

[0092] Using formula Perform color difference balance on the filtered image to obtain a color difference balanced image; wherein K R Indicates the balance coefficient of the red channel, K G Indicates the balance coefficient of the green channel, K B Represents the balance coefficient of the blue channel, R′ represents the pixel value of the red channel after balance, G′ represents the pixel value of the green channel after balance, B′ represents the pixel value of the blue channel after balance, R represents the pixel value of the filtered image in the red channel, G represents the pixel value of the filtered image in the green channel, and B represents the pixel value of the filtered image in the blue channel.

[0093] Specifically, this embodiment considers that when the external lighting conditions change, the surface color of the object remains unchanged, but the color of the object captured by the imaging device will deviate with the change of lighting, gradually deviating from the true color of the object. The present invention uses a color difference balance algorithm to compensate for this defect of the imaging device and correct the color difference of the image.

[0094] Preferably, image recognition is performed based on the preprocessed image to obtain food information of various types of food, including:

[0095] Using a convolutional neural network to extract features from the preprocessed image to obtain visual features of different foods;

[0096] Inputting the visual features into a preset food classification model to classify the food in the image to obtain the category of each food;

[0097] Identifying the boundaries of each food in the visual features by an image segmentation method, and calculating the number of food instances of the same category to determine the number of each food;

[0098] Calculating the main color tone of each food using a color histogram extraction method, and determining the color of the food according to the main color tone;

[0099] The identified food types, food quantities and food colors are collected to construct a food information list.

[0100] Specifically, the process of feature extraction using convolutional neural network (CNN) in this embodiment is as follows:

[0101] 1. Select an appropriate deep learning architecture, such as ResNet, VGG or MobileNet. This embodiment uses ResNet, which has been pre-trained on a large food image dataset and can effectively extract visual features.

[0102] 2. The pre-processed image after denoising and enhancement is input into the selected CNN model. In this embodiment, the image size is adjusted to meet the input requirements of the model (such as 224x224 pixels).

[0103] 3. In the forward propagation process of CNN, multiple convolutional layers and pooling layers are used to extract multi-layer feature representations of the image. Finally, the features output by the fully connected layer can represent the visual features of different foods in the image.

[0104] Furthermore, the steps of classifying food using the extracted visual features in this embodiment are as follows:

[0105] 1. Using transfer learning, the extracted features are input into a preset food classification model. This model can have multiple output classes, representing different types of food.

[0106] 2. Fine-tune the classification model based on existing food images and corresponding label data to improve recognition accuracy.

[0107] 3. For each preprocessed image, obtain the food category information through the output of the model and record the type of food in each category.

[0108] Furthermore, this embodiment uses an image segmentation method to identify the boundaries of each food in the visual features:

[0109] 1. Use advanced image segmentation algorithms, such as Mask R-CNN, U-Net, or YOLO, to identify the area of ​​each food. The mask obtained through segmentation can clearly distinguish the boundaries of different foods. For each category, calculate the area where it appears in the image. Count the food instances of the same category to determine the number of each food. This is usually achieved by analyzing the number of identified boundaries.

[0110] 2. Use the color histogram extraction method to calculate the main color of each food. Convert the image from RGB color space to HSV color space to extract the main color. Calculate the color histogram for each food area and count the brightness and saturation of each color channel. Find the color value with the highest frequency from the histogram to determine the main color of the food. This can be obtained by calculating the average color value of each category area.

[0111] Furthermore, the specific steps of constructing the food information list in this embodiment include:

[0112] 1. Organize the identified food types, food quantities, and food colors into a structured data format, such as a dictionary, JSON, or database record.

[0113] 2. Form a food information list containing multiple food items, each of which contains its category, quantity and color, to facilitate storage and subsequent management and analysis.

[0114] The above process uses the knowledge of deep learning, image processing and data structure. This embodiment realizes the extraction and recording of detailed information of food in the refrigerator from the pre-processed image, providing important data support for the subsequent intelligent early warning mechanism.

[0115] Preferably, the method for constructing the food freshness prediction model includes:

[0116] Constructing an initial convolutional neural network; the placement time, the temperature information, the humidity information, the gas information and the food information;

[0117] The initial convolutional neural network is trained according to the sample placement time data set, the sample temperature information data set, the sample humidity information data set, the sample gas information data set and the sample food information data set to obtain a trained first classifier, a second classifier, a third classifier, a fourth classifier and a fifth classifier;

[0118] Cascading the trained first classifier, second classifier, third classifier, fourth classifier and fifth classifier to obtain a classification network;

[0119] Connecting the trained LSTM neural network after the classification network to obtain a CNN-LSTM combined neural network model;

[0120] The placement time, the temperature information, the humidity information, the gas information and the food information are predicted according to the CNN-LSTM combined neural network model to obtain the prediction result.

[0121] Specifically, this embodiment first constructs an initial convolutional neural network (CNN), designs and constructs an initial convolutional neural network model, so as to process the feature extraction of food information and environmental data. The network architecture includes the following parts:

[0122] Input layer: Set up multiple input nodes to receive the placement time, temperature, humidity, gas information and food information respectively.

[0123] Convolutional layer: Feature extraction is performed through multiple convolutional layers to extract the spatial features of the image and gradually reduce the dimension of the feature map in subsequent layers.

[0124] Activation layer: Add ReLU activation function after the convolution layer to introduce nonlinear features.

[0125] Pooling layer: The size of the feature map is reduced through the Max Pooling layer, thereby reducing the computational complexity.

[0126] Fully connected layer: A fully connected layer is added at the end of the network to map the features extracted by the convolutional layer to the label space for classification.

[0127] Secondly, this embodiment trains each classifier and establishes multiple corresponding data sets to train CNN. The specific steps are as follows:

[0128] Construct independent data sets based on each input feature (placement time, temperature, humidity, gas, and food information), and ensure that each data set contains enough sample data to improve the effectiveness of training. Use the above sample data sets to train the initial CNN separately, as follows:

[0129] First classifier: Use sample placement time data to train CNN, focusing on learning the relationship between placement time and food freshness.

[0130] The second classifier: Use sample temperature information data to train CNN and extract the characteristics of the impact of temperature changes on food freshness.

[0131] The third classifier: Use sample humidity information data to train CNN and establish the impact of humidity on food storage conditions.

[0132] The fourth classifier is trained using sample gas information data to identify the indication of food status by changes in gas composition.

[0133] Fifth classifier: Use sample food information data to train CNN to understand the preservation characteristics of different foods.

[0134] Specifically, this embodiment cascades multiple trained classifiers to form a comprehensive classification network, as follows:

[0135] The output features of each classifier are concatenated to form a more comprehensive feature representation. This involves connecting the classifiers to combine the effects of different types of inputs. Using fusion strategies such as weighted averaging and concatenation, the features of multiple classifiers are combined to form a powerful feature representation. This can be achieved through a simple connection layer or feature fusion layer.

[0136] Furthermore, in this embodiment, a long short-term memory network (LSTM) is connected behind the cascaded classification network. LSTM introduces time series characteristics: since the time, temperature, humidity and gas information are often time series data, LSTM can capture the time series characteristics therein and improve the accuracy of the model's freshness prediction. The feature sequence obtained by the classification network is input into the LSTM, and the LSTM will learn the correlation at different time points through the internal gating mechanism to predict the future state of food. The merged data set is used to train the entire CNN-LSTM combined model, and all weights in the joint network are updated to ensure that the model can effectively integrate the learning results of the convolutional layer and the LSTM layer. According to the trained CNN-LSTM combined network, prediction is performed on the new input data to obtain the freshness prediction result of the food.

[0137] Through the above steps, the CNN-LSTM combined neural network model of this embodiment can effectively combine visual features and environmental information to accurately predict the freshness of food, providing a scientific basis for intelligent food storage management.

[0138] Optionally, the prediction results of this embodiment include multiple key information, covering the freshness status of the food, the estimated remaining shelf life, and environmental factors that affect the quality of the food. Specifically, the prediction results will provide the current freshness score of each food (such as fresh, about to go bad, spoiled, etc.), combined with the time of placement, temperature, humidity and gas information, to evaluate the shelf life changes of the food in the next few days. In addition, this embodiment will also analyze the impact of environmental conditions on the freshness of food, provide corresponding suggestions and warnings, and help users take timely measures to ensure food safety and quality. With this information, users can better manage the food in the refrigerator, reduce waste and ensure healthy diet.

[0139] Furthermore, this embodiment sets the freshness threshold of each food (such as remaining shelf life, temperature and humidity safety range, etc.) according to historical data and food type. When the prediction result of the CNN-LSTM combined model shows that the state of a certain food is close to or lower than the set threshold, the system will automatically trigger the early warning mechanism, and promptly remind the user that the food is about to go bad or is no longer suitable for consumption through mobile phone APP push notifications, refrigerator display warnings or sound prompts, etc., so as to help the user take corresponding measures, such as eating, processing or discarding the food as soon as possible, so as to reduce food waste and protect family health.

[0140] Specifically, the present embodiment also establishes a database containing food characteristics (such as optimal storage temperature, humidity, shelf life, etc.) and user dietary preferences. By analyzing the user's historical consumption data and preferences, the present embodiment can identify the types of food that the user often purchases and their storage habits. Combined with real-time monitored environmental data and food freshness prediction results, the present embodiment will generate personalized storage recommendations, such as providing more frequent inspection reminders for perishable foods, or suggesting that users adjust refrigerator settings under specific conditions (such as when the temperature is too high or the humidity is too high). In addition, the present embodiment can also push notifications to remind users to eat certain foods as soon as possible before the best time to eat, or provide recipe suggestions to encourage users to use ingredients that are about to expire, thereby improving the user's overall experience and satisfaction.

[0141] like Figure 2 As shown, this embodiment provides a food storage early warning system based on artificial intelligence, including:

[0142] A data acquisition module is used to acquire internal images of the target refrigerator, multi-source sensor signals, and the time when each food is placed in the refrigerator;

[0143] An image preprocessing module, used for performing image denoising and enhancement processing on the internally captured image to obtain a preprocessed image;

[0144] An image recognition module, used to perform image recognition based on the pre-processed image to obtain food information of various types of food; the food information includes: food type, food quantity and food color;

[0145] A signal preprocessing module, used for performing signal analysis on the multi-source sensor signals to obtain temperature information, humidity information and gas information;

[0146] A prediction module, used for inputting the placement time, the temperature information, the humidity information, the gas information and the food information into a preset food freshness prediction model to obtain a prediction result;

[0147] The early warning module is used to issue an early warning according to the prediction result.

[0148] The beneficial effects of the present invention are as follows:

[0149] (1) The present invention monitors and predicts the freshness of food in real time, and promptly reminds users to handle food that is about to go bad, thereby effectively reducing food waste in the home.

[0150] (2) The present invention can provide early warnings based on environmental changes and food status, helping users avoid eating expired or spoiled food and protecting the health and safety of family members.

[0151] (3) The present invention combines artificial intelligence technology to automatically manage food information in the refrigerator, eliminating the need for users to record information manually, thereby improving the convenience and efficiency of food management.

[0152] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system 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.

[0153] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A food storage early warning method based on artificial intelligence, characterized in that: include: Obtaining internal images of the target refrigerator, multi-source sensor signals, and the time when each food was placed; Performing image denoising and enhancement processing on the internally captured image to obtain a preprocessed image; Performing image recognition based on the preprocessed image to obtain food information of various types of food; The food information includes: food type, food quantity and food color; Performing signal analysis on the multi-source sensor signals to obtain temperature information, humidity information and gas information; Inputting the placement time, the temperature information, the humidity information, the gas information and the food information into a preset food freshness prediction model to obtain a prediction result; An early warning is issued based on the prediction results.

2. The artificial intelligence-based food storage early warning method according to claim 1, characterized in that: Performing image denoising and enhancement processing on the internally captured image to obtain a pre-processed image, including: Performing filtering processing on the internal captured image to obtain a filtered image; Performing color difference balance on the filtered image to obtain a color difference balanced image; Construct an image enhancement window, and use the image enhancement window to traverse the image after chromatic aberration balance to perform denoising, and obtain a denoised image; The denoised image is subjected to histogram equalization processing to obtain the pre-processed image.

3. The artificial intelligence-based food storage early warning method according to claim 2, characterized in that: Performing filtering processing on the internal captured image to obtain a filtered image, including: Performing wavelet decomposition on the internal captured image at multiple scales to obtain corresponding wavelet coefficients; constructing a filtering threshold according to the size and decomposition scale of the internal captured image; constructing an image filtering model using the filtering threshold; The internal captured image is filtered using the image filtering model to obtain the filtered image.

4. The artificial intelligence-based food storage early warning method according to claim 3, characterized in that: Performing color difference balancing on the filtered image to obtain a color difference balanced image includes: Calculate the pixel mean of the filtered image in the three RGB channels; wherein the pixel mean calculation formula is: Where D represents the pixel mean, represents the pixel mean of the filtered image in the red channel, G represents the pixel mean of the filtered image in the green channel, and B represents the pixel mean of the filtered image in the blue channel; Calculating a balance coefficient for each color channel according to the pixel mean; The balancing coefficient is used to perform color balance on the filtered image to obtain a color-balanced image.

5. The artificial intelligence-based food storage early warning method according to claim 4, characterized in that: Performing color difference balance on the filtered image using the balance coefficient to obtain a color difference balanced image, including: Using formula Perform color difference balance on the filtered image to obtain a color difference balanced image; wherein K R Indicates the balance coefficient of the red channel, K G Indicates the balance coefficient of the green channel, K B Represents the balance coefficient of the blue channel, R′ represents the pixel value of the red channel after balance, G′ represents the pixel value of the green channel after balance, B′ represents the pixel value of the blue channel after balance, R represents the pixel value of the filtered image in the red channel, G represents the pixel value of the filtered image in the green channel, and B represents the pixel value of the filtered image in the blue channel.

6. The artificial intelligence-based food storage early warning method according to claim 1, characterized in that: Image recognition is performed based on the pre-processed image to obtain food information of various types of food, including: Using a convolutional neural network to extract features from the preprocessed image to obtain visual features of different foods; Inputting the visual features into a preset food classification model to classify the food in the image to obtain the category of each food; Identifying the boundaries of each food in the visual features by an image segmentation method, and calculating the number of food instances of the same category to determine the number of each food; Calculating the main color tone of each food using a color histogram extraction method, and determining the color of the food according to the main color tone; The identified food types, food quantities and food colors are collected to construct a food information list.

7. The food storage early warning method based on artificial intelligence according to claim 1 is characterized in that: The method for constructing the food freshness prediction model comprises: Constructing an initial convolutional neural network; the placement time, the temperature information, the humidity information, the gas information and the food information; The initial convolutional neural network is trained according to the sample placement time data set, the sample temperature information data set, the sample humidity information data set, the sample gas information data set and the sample food information data set to obtain a trained first classifier, a second classifier, a third classifier, a fourth classifier and a fifth classifier; Cascading the trained first classifier, second classifier, third classifier, fourth classifier and fifth classifier to obtain a classification network; Connecting the trained LSTM neural network after the classification network to obtain a CNN-LSTM combined neural network model; The placement time, the temperature information, the humidity information, the gas information and the food information are predicted according to the CNN-LSTM combined neural network model to obtain the prediction result.

8. A food storage early warning system based on artificial intelligence, characterized in that: include: A data acquisition module is used to acquire internal images of the target refrigerator, multi-source sensor signals, and the time when each food is put in; An image preprocessing module, used for performing image denoising and enhancement processing on the internally captured image to obtain a preprocessed image; An image recognition module, used to perform image recognition based on the pre-processed image to obtain food information of various types of food; The food information includes: food type, food quantity and food color; A signal preprocessing module, used for performing signal analysis on the multi-source sensor signals to obtain temperature information, humidity information and gas information; A prediction module, used for inputting the placement time, the temperature information, the humidity information, the gas information and the food information into a preset food freshness prediction model to obtain a prediction result; The early warning module is used to issue an early warning according to the prediction result.

Images