Article classification and storage management system based on gravity sensing

Abstract

The invention discloses an article classification and storage management system based on gravity sensing, and relates to the technical field of storage management, the article classification and storage management system comprises an article storage device, a sensor array module, an article category analysis module, an article moving module and a user interface module are integrated on the article storage device, and the modules are in electric signal connection; the sensor array module collects multi-dimensional data of an article through various types of deployed sensors. According to the article classification and storage management system based on gravity sensing, the touch screen and the user interface module with the voice prompt and gesture recognition functions are combined, visual and convenient operation experience is provided for a user, the user can easily check the storage state, adjust the storage rule and customize the interface style according to personal preference through the touch screen, and meanwhile, the user can conveniently use the system. The voice prompt function enables the operation process to be more intuitive and understandable, thereby not only enhancing the sense of participation and satisfaction of the user, but also improving the use experience of the whole storage management system.

Description

Technical Field

[0001] The present invention relates to the technical field of storage management, and in particular to an article classification and storage management system based on gravity sensing. Background Art

[0002] With the continuous development of science and technology, automation and intelligent equipment have been widely used in daily life and work, greatly improving people's work efficiency and quality of life. In the field of storage, the traditional manual classification and storage method is not only time-consuming and labor-intensive, but also prone to errors, and cannot meet the fast-paced and efficient needs of modern society. Therefore, it is particularly important to develop a system that can automatically identify, classify and store items.

[0003] Most of the existing storage systems rely on technologies such as barcode scanning and RFID recognition, and items need to be marked, which not only increases the cost but is also not suitable for all items, limiting its scope of application. In order to solve this problem, an item classification and storage management system based on gravity sensing is proposed. Summary of the invention

[0004] To achieve the above objectives, the present invention is implemented through the following technical solutions: an article classification and storage management system based on gravity sensing, comprising an article storage device, on which a sensor array module, an article category analysis module, an article movement module and a user interface module are integrated, wherein electrical signals between the modules are connected; The sensor array module collects multi-dimensional data of the object through various types of deployed sensors, wherein the sensors include weighing sensors, infrared sensors, ultrasonic sensors and stereo vision sensors; The object category analysis module receives and processes the sensor data from the sensor array module, uses a machine learning algorithm to determine the category of the object, efficiently processes the data and outputs the classification results, controls subsequent storage actions, and realizes intelligent management; The article moving module receives the classification result output by the article category analysis module, controls the article moving mechanism to move the article to the designated storage location, and has high-precision positioning and motion control capabilities to ensure that the article is safely and accurately transported and placed, thereby improving the storage efficiency. The article moving mechanism is a robotic arm or a conveying system; The user interface module includes a touch screen, voice prompts and gesture recognition functions, allowing users to view storage status, adjust storage rules and view statistical information, providing an intuitive and friendly operation and information display interface, allowing users to easily manage the system, optimize storage effects and enhance user experience.

[0005] Preferably, the sensor array module includes a gravity sensing unit and an auxiliary sensor unit; The gravity sensing unit uses a high-precision weighing sensor to measure the weight of the items placed on it, accurately measure items of different weight levels, and provide the system with core item weight data; The auxiliary sensor unit uses infrared sensors, ultrasonic sensors and stereo vision sensors to detect the position, size and shape of objects, and provides the system with a comprehensive description of the objects by monitoring the three-dimensional information and dynamic changes of the objects in real time, thereby enhancing the accuracy of classification.

[0006] Preferably, in the sensor array module, the process of collecting multi-dimensional data of an object includes: Weighing sensors, infrared sensors, ultrasonic sensors and stereo vision sensors are deployed in the item storage to fully cover the storage area, and all sensors are initialized and calibrated to ensure their accuracy and stability, including adjusting the sensitivity of the sensors and calibrating the zero point to ensure the accuracy of the measurement results; Click the Place button on the touch screen of the user interface module, the drawer of the item storage pops out and the item is placed in it, the drawer retracts into the item storage, and the high-precision weighing sensor in the gravity sensing unit measures the pressure change of the item on the surface of the weighing sensor to calculate the weight of the item; The infrared sensor emits infrared light and receives its reflected light, measures the time change of the reflected light to detect the position of the object, and uses the infrared sensor to monitor the movement and position change of the object in real time to obtain the object's location information; Use multiple ultrasonic sensors to emit ultrasonic pulses and receive their echoes, calculate the distance between the object and the sensor by measuring the echo time, and combine the measurement results of multiple ultrasonic sensors to calculate the three-dimensional size and shape of the object using the triangulation principle; Use multiple cameras to form a stereo vision sensor to capture images of objects from different angles, and reconstruct the 3D model of the object through image processing and 3D reconstruction algorithms; Integrate the multi-dimensional sensor data of weight, position, size and shape collected by the sensor array module, and transmit the sensor data to the item category analysis module.

[0007] Preferably, the item category analysis module includes a data acquisition unit, a data preprocessing unit, a feature extraction unit, a category classification unit and an output unit, wherein the units are connected by electrical signals; The data acquisition unit is used to collect data from various sensors, including weight, position, size and shape information; The data preprocessing unit is used to preprocess the collected sensor data, including filtering, denoising and normalization operations, so as to improve data quality, reduce errors and interference, and ensure the accuracy of subsequent classification algorithms; The feature extraction unit is used to extract the identification features of weight, length, width, height and contour from the preprocessed sensor data to accurately describe the properties of the object; The category classification unit applies a machine learning algorithm to train an object category recognition model based on the extracted features, thereby determining the category of the object; The output unit is used to output the classification result of the article category and control the article moving mechanism of the article moving module to move the article to a designated storage position.

[0008] Preferably, in the item category analysis module, the process of determining the item category includes: The sensor data of weight, position, size and shape information are received from the sensor array module by the data acquisition unit in a wireless manner, and the sensor data is preprocessed by the data preprocessing unit, including filtering, denoising and normalization operations, wherein the filtering operation removes high-frequency noise in the data to improve the smoothness of the data, the denoising operation eliminates random errors and interference in the data to improve the accuracy of the data, and the normalization operation is used to convert data of different dimensions into the same dimension range for subsequent processing and analysis; Using a feature extraction unit, extracting identification features of weight, length, width, height and contour from the preprocessed sensor data, and arranging the extracted features to form an identification feature sequence; Collect data on items of known categories, extract identification features from them, and train an item category recognition model in combination with a convolutional neural network model, so as to use the item category recognition model to perform identification and analysis on item categories; The acquired sensor data is input into the object category recognition model, the acquired weight, length, width, height and contour sensor data are integrated with the object recognition feature data of known categories, the object category matching index is calculated, the matching degree of the recognition features is analyzed, and then the sensor data is classified to determine the category of the object; If the data of items of known categories is matched successfully, the output unit is used to output the classification result of the identified item category to the item movement module. The item movement module performs corresponding operations according to the identified item category. If the match is unsuccessful, the item type and weight information are manually input in the user interface module, and then the item movement module is used to perform corresponding item movement operations.

[0009] Preferably, the training process of the item category recognition model is: Collect a large amount of data on known categories of objects, including sensor data of weight, length, width, height and corresponding contour data. The contour data is extracted from the image through image recognition technology (edge ​​detection, shape matching), and the known category object data is preprocessed by filtering, denoising and normalization to ensure the quality and consistency of the data. The object category data set is obtained, and then the object category data set is divided into a training set and a test data set; Extract the identification features of weight, length, width, height and outline from the object category dataset, and arrange the obtained identification features into an identification feature sequence as the input of the model; Design a CNN model structure that processes multi-dimensional recognition feature input, including input layer, convolution layer, pooling layer, fully connected layer and output layer. The input layer is used to receive the recognition feature sequence, the convolution layer is used to extract features, the pooling layer is used to reduce dimensionality and reduce the amount of calculation, the fully connected layer is used to integrate features and output classification results, and the output layer is used to output the item category. The extracted recognition feature sequences and corresponding item category labels are input into the CNN model as training data sets. The weight and bias parameters of the model are updated through iterative training using the optimization method of the back-propagation algorithm. During the training process, the cross-validation technique is used to evaluate the performance of the model, and the model is adjusted and optimized based on the evaluation results. Calculate the accuracy, recall, and F1 score indicators, evaluate the performance of the trained item category recognition model in combination with the test data set, and analyze the classification effect of the model, and then integrate the trained item category recognition model into the item classification and storage management system; The calculation formula for updating the model weights is: ; in, and Respectively represent the weight values ​​before and after the update, Represents the learning rate, which is used to control the step size of weight update. Represents the loss function About weight The gradient of is used to guide the update direction of the weights; The calculation formula for updating the bias parameter is: ; in, and Respectively represent the bias values ​​before and after the update, represents the learning rate, Represents the loss function About Bias gradient.

[0010] Preferably, the calculation formula of the item category matching index is: ; ; in, is the item category matching index, is a sensor dataset, including weight, length, width, height and contour data, is the item to be classified. For items of known categories, For items and The ratio of the intersection and union of features reflects the similarity of the features of two items. For items and The size of the intersection of the features, For items and The size of the union of features, when When ,express and The characteristics are exactly the same. tends to infinity, indicating a perfect match.

[0011] Preferably, in the item moving module, the item moving process includes: The object moving module receives the output result from the object classification analysis module through the interface, which includes the classification information of the objects to be moved. According to the characteristics of the objects (size, weight, fragility) and the classification results, the received classification information is parsed, and the corresponding object moving mechanism, i.e., the robot arm or the conveying system, is selected. The robot arm is a multi-joint type or SCARA type, with high-precision positioning and motion control capabilities, and is suitable for objects of different sizes and weights, or objects that need to be placed accurately. The conveying system includes a conveyor belt, a roller or other types of conveying mechanisms, which are suitable for objects of similar size and moderate weight, and can smoothly carry objects and accurately place them at the designated location. After determining the moving mechanism, the object moving module plans the optimal moving path according to the information of the target storage location to ensure that the object can reach the designated location safely and quickly. During the moving process, the state and location of the object are monitored in real time to ensure that the object is not damaged during the moving process and reaches the designated location accurately. When the item arrives at the designated location, the item moving module controls the moving mechanism to place the item, confirms that the storage is completed, and records the storage completion information in the user interface module for subsequent reference.

[0012] Preferably, the confirmation process of the mobile mechanism includes: If the moving mechanism is a robotic arm, the robotic arm is initialized before starting to check the flexibility of each joint and confirm the sensor status. According to the output result of the object category analysis module, the robotic arm adjusts its grasping mechanism to adapt to the shape and size of the object and then grasps the object. On the planned moving path, the robotic arm moves the object at a predetermined speed and acceleration to maintain the stability of the object. When the robotic arm reaches the specified position, it adjusts the grasping mechanism to release the object and places it in the correct position. After completing a handling task, the robotic arm returns to the initial position and waits for the next handling task. If the mobile mechanism is a conveying system, a safety check is performed on the conveying system before startup to ensure that all components are operating normally and there are no obstacles. According to the output results of the item category analysis module, the speed and direction of the conveyor line are adjusted to meet the needs of different items, and the items are placed at the starting position of the conveyor line. During the conveying process, the position and status of the items are monitored in real time to ensure that the items are not damaged during the conveying process. When the items arrive at the designated location, the conveying system cooperates with the sorting mechanism to sort the items to the designated storage location. After completing a conveying task, the conveying system waits for the next conveying task.

[0013] Preferably, the user interface module specifically includes: When the item classification and storage management system is started, the user interface module automatically starts, performs initialization operations, loads interface elements, configures voice prompts and gesture recognition functions, and requires users to authenticate based on the set login permissions to ensure that only authorized users can access the system; The user interface module displays the current storage status through the touch screen, including the quantity, type, and location information of the stored items, and announces the current storage status to the user through the voice prompt function, or switches the display mode of the storage status through specific gestures to obtain detailed information; Adjust storage rules according to user needs. Storage rules include setting storage priorities for different items, specifying specific storage locations, or creating custom storage categories. After the user submits a new storage rule, the user interface module verifies it to ensure the validity and feasibility of the rule. If the rule is conflicting or unreasonable, corresponding prompts and suggestions are given. Once the rule is verified, the user interface module applies it to the item classification and storage management system and updates the storage status in real time to reflect the new rule.

[0014] The present invention provides an article classification and storage management system based on gravity sensing. It has the following beneficial effects: 1. This gravity-sensing-based item classification and storage management system, combined with a user interface module with touch screen, voice prompts and gesture recognition functions, provides users with an intuitive and convenient operation experience. Users can easily view the storage status, adjust the storage rules, and customize the interface style according to personal preferences through the touch screen. At the same time, the voice prompt function makes the operation process more intuitive and easy to understand, which not only enhances the user's sense of participation and satisfaction, but also improves the user experience of the entire storage management system.

[0015] 2. This gravity-sensing-based item classification and storage management system monitors the weight changes of items through gravity-sensing technology, thereby accurately judging whether the items are placed correctly, greatly improving the efficiency and accuracy of storage management. It not only reduces the tediousness of manual classification, but also avoids the loss or confusion of items caused by misplacement, and then monitors the usage of storage space to further optimize space utilization.

[0016] 3. This gravity-sensing-based item classification and storage management system uses sensing technology and intelligent algorithms to achieve automatic identification, classification and storage of items without any special marking of the items. It not only improves operational efficiency, but also significantly reduces labor costs, making it an ideal choice for retail, warehousing, home and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a module diagram of the gravity-sensing-based article classification and storage management system of the present invention; Figure 2 This is a flow chart of determining the category of an item according to the present invention. DETAILED DESCRIPTION

[0018] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention are provided for the purpose of illustration and description, and are not intended to be exhaustive or to limit the present invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention, and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

[0019] The first embodiment, as Figure 1 As shown, the present invention provides a technical solution: an item classification and storage management system based on gravity sensing, comprising an item storage device, on which a sensor array module, an item category analysis module, an item movement module and a user interface module are integrated, wherein electrical signals are connected between the modules; The sensor array module collects multi-dimensional data of objects through the deployment of various types of sensors, wherein the sensors include weighing sensors, infrared sensors, ultrasonic sensors and stereo vision sensors, and the sensor array module includes a gravity sensing unit and an auxiliary sensor unit; wherein the gravity sensing unit uses a high-precision weighing sensor to measure the weight of the objects placed on it, accurately measures objects of different weight levels, and provides the system with core object weight data; the auxiliary sensor unit uses infrared sensors, ultrasonic sensors and stereo vision sensors to detect the position, size and shape of objects, and provides the system with a comprehensive object description by real-time monitoring of the three-dimensional information and dynamic changes of the objects, thereby enhancing the accuracy of classification; Furthermore, in the sensor array module, the process of collecting multi-dimensional data of an object includes: Weighing sensors, infrared sensors, ultrasonic sensors and stereo vision sensors are deployed in the item storage to fully cover the storage area, and all sensors are initialized and calibrated to ensure their accuracy and stability, including adjusting the sensitivity of the sensor and calibrating the zero point to ensure the accuracy of the measurement results. When the placement button is clicked on the touch screen of the user interface module, the item storage drawer pops out and the item is placed in it. The drawer is retracted into the item storage, and the high-precision weighing sensor in the gravity sensing unit measures the pressure change of the item on the surface of the weighing sensor to calculate the weight of the item. The infrared sensor emits infrared light and receives its reflected light, and the time change of the reflected light is measured to detect Measure the position of the object, use infrared sensors to monitor the movement and position changes of the object in real time, obtain the position information of the object, use multiple ultrasonic sensors to emit ultrasonic pulses and receive their echoes, calculate the distance between the object and the sensor by measuring the time of the echo, combine the measurement results of multiple ultrasonic sensors, use the triangulation principle to calculate the three-dimensional size and shape of the object, use multiple cameras to form a stereo vision sensor, capture the image of the object from different angles, reconstruct the three-dimensional model of the object through image processing and three-dimensional reconstruction algorithm, integrate the multi-dimensional sensor data of weight, position, size and shape collected by the sensor array module, and transmit the sensor data to the item category analysis module; The item category analysis module receives and processes the sensor data from the sensor array module, uses the machine learning algorithm to determine the category of the item, efficiently processes the data and outputs the classification result, controls the subsequent storage action, and realizes intelligent management. The item category analysis module includes a data acquisition unit, a data preprocessing unit, a feature extraction unit, a category classification unit, and an output unit, wherein the electrical signals between the units are connected. The data acquisition unit is used to collect data from each sensor, including weight, position, size and shape information. The data preprocessing unit is used to preprocess the collected sensor data, including filtering, denoising and normalization operations to improve data quality, reduce errors and interference, and ensure the accuracy of the subsequent classification algorithm. The feature extraction unit is used to extract the identification features of weight, length, width, height and contour from the preprocessed sensor data to accurately describe the attributes of the item. The category classification unit applies the machine learning algorithm to train the item category recognition model according to the extracted features, and then determines the category of the item. The output unit is used to output the classification result of the item category and control the item moving mechanism of the item moving module to move the item to the designated storage position; The item moving module receives the classification results output by the item category analysis module, controls the item moving mechanism to move the items to the designated storage location, and has high-precision positioning and motion control capabilities to ensure that the items are safely and accurately transported and placed, thereby improving storage efficiency. The item moving mechanism is a robotic arm or a conveying system; The user interface module, including touch screen, voice prompt and gesture recognition functions, allows users to check storage status, adjust storage rules and view statistical information, providing an intuitive and friendly operation and information display interface, allowing users to easily manage the system, optimize storage effects and enhance user experience.

[0020] The second embodiment is based on the first embodiment. Figure 2 As shown, in the item category analysis module, the process of determining the item category includes: The sensor data of weight, position, size and shape information is received from the sensor array module by the data acquisition unit in a wireless manner, and the sensor data is preprocessed by the data preprocessing unit, including filtering, denoising and normalization operations, wherein the filtering operation removes high-frequency noise in the data and improves the smoothness of the data, and the denoising operation eliminates random errors and interference in the data and improves the accuracy of the data, and the normalization operation is used to convert data of different dimensions into the same dimension range for subsequent processing and analysis, and the feature extraction unit is used to extract the identification features of weight, length, width, height and contour from the preprocessed sensor data, and the extracted features are arranged to form an identification feature sequence, collect the data of items of known categories, and extract the identification features therefrom, and combine the convolutional neural network The network model is used to train the item category recognition model, so as to use the item category recognition model to recognize and analyze the item category, input the acquired sensor data into the item category recognition model, comprehensively obtain the weight, length, width, height and profile sensor data and the item recognition feature data of the known category, calculate the item category matching index, analyze the matching degree of the recognition feature, and then classify the sensor data to judge the category of the item. If the item data of the known category is matched successfully, the output unit is used to output the classification result of the identified item category to the item movement module. According to the identified item category, the item movement module performs corresponding operations. If the match is unsuccessful, the item type and weight information are manually input in the user interface module, and then the item movement module is used to perform the corresponding item movement operation; Furthermore, the training process of the object category recognition model is: A large amount of data on known categories of objects is collected, including sensor data of weight, length, width, height and corresponding contour data. The contour data is extracted from the image through image recognition technology (edge ​​detection, shape matching), and the known category object data is filtered, denoised and normalized to ensure the quality and consistency of the data. The object category data set is obtained, and then the object category data set is divided into a training set and a test data set. The identification features of weight, length, width, height and contour are extracted from the object category data set, and the obtained identification features are arranged into an identification feature sequence as the input of the model. The CNN model structure that processes multi-dimensional identification feature input is designed, including input layer, convolution layer, pooling layer, fully connected layer and output layer. The input layer is used Receive the recognition feature sequence, the convolution layer is used to extract features, the pooling layer is used to reduce dimension and reduce the amount of calculation, the fully connected layer is used to integrate features and output classification results, the output layer is used to output the item category, the extracted recognition feature sequence and the corresponding item category label are input into the CNN model as the training data set, the back propagation algorithm optimization method is used, the weight and bias parameters of the model are updated through iterative training, during the training process, the cross-validation technology is used to evaluate the performance of the model, and the model is adjusted and optimized according to the evaluation results, the accuracy, recall rate, and F1 score indicators are calculated, and the performance of the trained item category recognition model is evaluated in combination with the test data set, and the classification effect of the model is analyzed, and then the trained item category recognition model is integrated into the item classification and storage management system; The calculation formula for updating the model weights is: ; in, and Respectively represent the weight values ​​before and after the update, Represents the learning rate, which is used to control the step size of weight update. Represents the loss function About weight The gradient of is used to guide the update direction of the weights; The calculation formula for updating the bias parameter is: ; in, and Respectively represent the bias values ​​before and after the update, represents the learning rate, Represents the loss function About Bias The gradient of Furthermore, the calculation formula of the item category matching index is: ; ; in, is the item category matching index, is a sensor dataset, including weight, length, width, height and contour data, is the item to be classified. For items of known categories, For items and The ratio of the intersection and union of features reflects the similarity of the features of two items. For items and The size of the intersection of the features, For items and The size of the union of features, when When ,express and The characteristics are exactly the same. tends to infinity, indicating a perfect match, when When it is close to 0, it means and The features of have almost no intersection. Close to 0, indicating a low matching degree. increase, will also increase, reflecting the improvement in feature matching; In the item movement module, the item movement process includes: The object moving module receives the output result from the object category analysis module through the interface, which includes the classification information of the objects to be moved. According to the characteristics of the objects (size, weight, fragility) and the classification results, the received classification information is parsed, and the corresponding object moving mechanism, i.e., a robot arm or a conveying system is selected, wherein the robot arm is a multi-joint type or SCARA type, and has high-precision positioning and motion control capabilities, and is suitable for objects of different sizes and weights, or objects that need to be accurately placed. The conveying system includes a conveyor belt, a roller or other types of conveying mechanisms, which are suitable for objects of similar size and moderate weight, and can smoothly carry the objects and accurately place them at the designated location. After determining the moving mechanism, the object moving module plans the optimal moving path according to the information of the target storage location to ensure that the objects can reach the designated location safely and quickly, and monitors the status and position of the objects in real time during the moving process to ensure that the objects are not damaged during the moving process and accurately reach the designated location. When the objects arrive at the designated location, the object moving module controls the moving mechanism to place the objects, confirms that the storage is completed, and records the information of the storage completion into the user interface module for subsequent reference; Furthermore, the confirmation process of the mobile mechanism includes: If the moving mechanism is a robotic arm, the robotic arm is initialized before starting to check the flexibility of each joint and confirm the sensor status. According to the output result of the object category analysis module, the robotic arm adjusts its grasping mechanism to adapt to the shape and size of the object and then grasps the object. On the planned moving path, the robotic arm moves the object at a predetermined speed and acceleration to maintain the stability of the object. When the robotic arm reaches the specified position, it adjusts the grasping mechanism to release the object and places it in the correct position. After completing a handling task, the robotic arm returns to the initial position and waits for the next handling task. If the mobile mechanism is a conveying system, a safety check is performed on the conveying system before starting to ensure that all components are operating normally and there are no obstacles. According to the output results of the item category analysis module, the speed and direction of the conveying line are adjusted to meet the needs of different items, and the items are placed at the starting position of the conveying line. During the conveying process, the position and status of the items are monitored in real time to ensure that the items are not damaged during the conveying process. When the items arrive at the designated location, the conveying system cooperates with the sorting mechanism to sort the items to the designated storage location. After completing a conveying task, the conveying system waits for the next conveying task; The user interface module specifically includes: When the item classification and storage management system is started, the user interface module automatically starts, performs initialization operations, loads interface elements, configures voice prompts and gesture recognition functions, and requires users to authenticate their identities based on the set login permissions to ensure that only authorized users can access the system. The user interface module displays the current storage status through the touch screen, including the number, type, and location information of the stored items, and announces the current storage status to the user through the voice prompt function, or switches the display mode of the storage status through specific gestures to obtain detailed information. According to user needs, the storage rules are adjusted. The storage rules include setting storage priorities for different items, specifying specific storage locations, or creating custom storage categories. After the user submits a new storage rule, the user interface module verifies it to ensure the validity and feasibility of the rule. If the rule conflicts or is unreasonable, corresponding prompts and suggestions are given. After the rule is verified, the user interface module applies it to the item classification and storage management system and updates the storage status in real time to reflect the new rule.

[0021] Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field and related fields without creative work should fall within the scope of protection of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention are implemented according to the conventional means in the field unless otherwise specified and limited.

Claims

1. An article classification and storage management system based on gravity sensing, including an article storage device, characterized in that: The article storage device is integrated with a sensor array module, an article category analysis module, an article movement module and a user interface module, wherein the modules are connected by electrical signals; The sensor array module collects multi-dimensional data of the object through various types of deployed sensors, wherein the sensors include weighing sensors, infrared sensors, ultrasonic sensors and stereo vision sensors; The object category analysis module receives and processes the sensor data from the sensor array module, uses a machine learning algorithm to determine the category of the object, and outputs a classification result; The article moving module receives the classification result output by the article category analysis module and controls the article moving mechanism to move the article to a designated storage location, wherein the article moving mechanism is a robotic arm or a conveying system; The user interface module includes a touch screen, voice prompts and gesture recognition functions, allowing users to view storage status, adjust storage rules and view statistical information.

2. The article classification and storage management system based on gravity sensing according to claim 1 is characterized in that: The sensor array module includes a gravity sensing unit and an auxiliary sensor unit; Wherein, the gravity sensing unit uses a high-precision weighing sensor to measure the weight of the items placed on it and measure items of different weight levels; The auxiliary sensor unit uses an infrared sensor, an ultrasonic sensor and a stereoscopic vision sensor to detect the position, size and shape of the object.

3. The article classification and storage management system based on gravity sensing according to claim 2 is characterized in that: In the sensor array module, the process of collecting multi-dimensional data of an object includes: Deploy weighing sensors, infrared sensors, ultrasonic sensors, and stereo vision sensors in the item storage to fully cover the storage area, and initialize and calibrate all sensors, including adjusting the sensitivity of the sensors and calibrating the zero point; Click the Place button on the touch screen of the user interface module, the drawer of the item storage pops out and the item is placed in it, the drawer retracts into the item storage, and the high-precision weighing sensor in the gravity sensing unit measures the pressure change of the item on the surface of the weighing sensor to calculate the weight of the item; The infrared sensor emits infrared light and receives its reflected light, measures the time change of the reflected light to detect the position of the object, and uses the infrared sensor to monitor the movement and position change of the object in real time to obtain the object's location information; Use multiple ultrasonic sensors to emit ultrasonic pulses and receive their echoes, calculate the distance between the object and the sensor by measuring the echo time, and combine the measurement results of multiple ultrasonic sensors to calculate the three-dimensional size and shape of the object using the triangulation principle; Use multiple cameras to form a stereo vision sensor to capture images of objects from different angles, and reconstruct the 3D model of the object through image processing and 3D reconstruction algorithms; Integrate the multi-dimensional sensor data of weight, position, size and shape collected by the sensor array module, and transmit the sensor data to the item category analysis module.

4. The article classification and storage management system based on gravity sensing according to claim 3 is characterized in that: The item category analysis module includes a data acquisition unit, a data preprocessing unit, a feature extraction unit, a category classification unit, and an output unit, wherein the units are connected by electrical signals; The data acquisition unit is used to collect data from various sensors, including weight, position, size and shape information; The data preprocessing unit is used to preprocess the collected sensor data, including filtering, denoising and normalization operations; The feature extraction unit is used to extract identification features of weight, length, width, height and contour from the preprocessed sensor data to describe the attributes of the object; The category classification unit applies a machine learning algorithm to train an object category recognition model based on the extracted features, thereby determining the category of the object; The output unit is used to output the classification result of the article category and control the article moving mechanism of the article moving module to move the article to a designated storage position.

5. The article classification and storage management system based on gravity sensing according to claim 4 is characterized in that: In the item category analysis module, the process of determining the item category includes: receiving sensor data of weight, position, size and shape information from the sensor array module by wireless means through a data acquisition unit, and preprocessing the sensor data using a data preprocessing unit; Using a feature extraction unit, extracting identification features of weight, length, width, height and contour from the preprocessed sensor data, and arranging the extracted features to form an identification feature sequence; Collect data on items of known categories, extract identification features from them, and train an item category recognition model in combination with a convolutional neural network model, so as to use the item category recognition model to perform identification and analysis on item categories; The acquired sensor data is input into the object category recognition model, the acquired weight, length, width, height and contour sensor data are integrated with the object recognition feature data of known categories, the object category matching index is calculated, the matching degree of the recognition features is analyzed, and then the sensor data is classified to determine the category of the object; If the data of items of known categories is matched successfully, the output unit is used to output the classification result of the identified item category to the item movement module. The item movement module performs corresponding operations according to the identified item category. If the match is unsuccessful, the item type and weight information are manually input in the user interface module, and then the item movement module is used to perform corresponding item movement operations.

6. The article classification and storage management system based on gravity sensing according to claim 5 is characterized in that: The training process of the item category recognition model is as follows: Collect a large amount of data on known categories of items, including sensor data of weight, length, width, height and corresponding contour data. The contour data is extracted from the image through image recognition technology, and the known category item data is preprocessed by filtering, denoising and normalization to obtain the item category data set, which is then divided into a training set and a test data set. Extract the identification features of weight, length, width, height and outline from the object category dataset, and arrange the obtained identification features into an identification feature sequence as the input of the model; Design a CNN model structure that processes multi-dimensional recognition feature input, including input layer, convolution layer, pooling layer, fully connected layer and output layer. The input layer is used to receive the recognition feature sequence, the convolution layer is used to extract features, the pooling layer is used to reduce dimensionality and reduce the amount of calculation, the fully connected layer is used to integrate features and output classification results, and the output layer is used to output the item category. The extracted recognition feature sequences and corresponding item category labels are input into the CNN model as training data sets. The weight and bias parameters of the model are updated through iterative training using the optimization method of the back-propagation algorithm. During the training process, the cross-validation technique is used to evaluate the performance of the model, and the model is adjusted and optimized based on the evaluation results. Calculate the accuracy, recall, and F1 score indicators, evaluate the performance of the trained item category recognition model in combination with the test data set, analyze the classification effect of the model, and then integrate the trained item category recognition model into the item classification and storage management system.

7. The article classification and storage management system based on gravity sensing according to claim 6 is characterized in that: The calculation formula of the item category matching index is: ； ； in, is the item category matching index, is a sensor dataset, including weight, length, width, height and contour data, is the item to be classified. For items of known categories, For items and The ratio of the intersection and union of features reflects the similarity of the features of two items. For items and The size of the intersection of the features, For items and The size of the union of features, when When ,express and The characteristics are exactly the same. tends to infinity, indicating a perfect match.

8. The article classification and storage management system based on gravity sensing according to claim 7 is characterized in that: In the item moving module, the process of item moving includes: The object moving module receives the output result from the object classification analysis module through the interface, including the classification information of the object to be moved, analyzes the received classification information according to the characteristics of the object and the classification result, and selects the corresponding object moving mechanism, i.e., the robot arm or the conveying system; After determining the moving mechanism, the object moving module plans the optimal moving path according to the information of the target storage location, and monitors the status and location of the object in real time during the moving process; When the item arrives at the designated location, the item moving module controls the moving mechanism to place the item, confirms that the storage is completed, and records the storage completion information in the user interface module for subsequent reference.

9. The article classification and storage management system based on gravity sensing according to claim 8, characterized in that: The confirmation process of the mobile mechanism includes: If the moving mechanism is a robotic arm, the robotic arm is initialized before starting to check the flexibility of each joint and confirm the sensor status. According to the output result of the object category analysis module, the robotic arm adjusts its grasping mechanism to adapt to the shape and size of the object and then grasps the object. On the planned moving path, the robotic arm moves the object at a predetermined speed and acceleration. When the robotic arm reaches the specified position, it adjusts the grasping mechanism to release the object and places it in the correct position. After completing a handling task, the robotic arm returns to the initial position and waits for the next handling task. If the mobile mechanism is a conveying system, a safety check is performed on the conveying system before startup to ensure that all components are operating normally and there are no obstacles. According to the output results of the item category analysis module, the speed and direction of the conveyor line are adjusted to meet the needs of different items, and the items are placed at the starting position of the conveyor line. During the conveying process, the position and status of the items are monitored in real time. When the items arrive at the designated location, the conveying system cooperates with the sorting mechanism to sort the items to the designated storage location. After completing a conveying task, the conveying system waits for the next conveying task.

10. The article classification and storage management system based on gravity sensing according to claim 9, characterized in that: The user interface module specifically includes: When the item classification and storage management system is started, the user interface module automatically starts, performs initialization operations, loads interface elements, configures voice prompts and gesture recognition functions, and requires the user to authenticate based on the set login permissions; The user interface module displays the current storage status through the touch screen, including the quantity, type, and location information of the stored items, and announces the current storage status to the user through the voice prompt function; Adjust storage rules according to user needs. Storage rules include setting storage priorities for different items, specifying specific storage locations, or creating custom storage categories. After the user submits a new storage rule, the user interface module verifies it. If the rule conflicts or is unreasonable, corresponding prompts and suggestions are given. Once the rule is verified, the user interface module applies it to the item classification and storage management system and updates the storage status in real time to reflect the new rule.

Images