Pet behavior monitoring and correcting method, system and equipment based on deep learning visual large model technology

Abstract

The invention belongs to the technical field of pet behavior monitoring and correction, and particularly relates to a pet behavior monitoring and correction method, system and device based on a deep learning visual large model technology, and the device comprises a mounting substrate, a main control board, a camera, a sensor group, and an edge calculation terminal. The mounting substrate is provided with hole sites for mounting the camera and the sensor group, and the camera and the sensor group are mounted in the corresponding hole sites of the mounting substrate; the camera and the sensor group are respectively used for detecting a pet motion image, living environment information and motion information. The accuracy and the real-time performance of pet behavior monitoring are improved. The workload of a pet owner is reduced, and the convenience of pet behavior management is improved. Through data analysis, more personalized and scientific nursing suggestions can be provided for pets.

Description

Technical Field

[0001] The present invention relates to the technical field of pet behavior monitoring and correction, and specifically to a method, system and device for pet behavior monitoring and correction based on deep learning visual large model technology. Background Art

[0002] Pets refer to creatures that people keep for spiritual purposes rather than economic purposes. Traditional pets refer to mammals or birds, which are kept for entertainment and companionship. In real life, pets include fish, reptiles, amphibians, insects, and even plants, which are used for viewing, companionship, and to relieve people's mental stress.

[0003] With the development of the times, the scope of pets is very wide, including animals, plants, virtual pets, electronic pets, etc. However, the laws of most countries still limit pets to animals. Raising pets is an opportunity for humans to get close to nature and can meet human psychological needs. It is a very healthy and normal hobby.

[0004] With the increase in the number of pet households, pet behavior management has become a concern for more and more pet owners. Traditional pet behavior management methods mainly rely on human observation and training, which is not only time-consuming and labor-intensive, but also difficult to quantify the effect. In recent years, with the development of computer vision technology and deep learning, it has become possible to use these technologies to automatically monitor and correct pet behavior. Summary of the invention

[0005] The purpose of the present invention is to provide a method, system and device for pet behavior monitoring and correction based on deep learning visual large model technology to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a device for pet behavior monitoring and correction based on deep learning visual large model technology, comprising:

[0007] Install the baseboard, main control board, camera, sensor group, and edge computing terminal;

[0008] The mounting substrate is provided with holes for mounting the camera and the sensor group, and the camera and the sensor group are mounted in the holes corresponding to the mounting substrate;

[0009] The camera and sensor group are used to detect pet motion images, living environment information and motion information respectively. The main control board is used to control the normal operation of the system. The edge computing terminal is responsible for processing the video image data collected by the camera and using the CNN convolution model for inference and prediction; and analyzing the data sent back by the sensor group. The main control board and the edge computing terminal are integrated in a device box and connected to the installation substrate through a data bus.

[0010] Preferably, the sensor group includes a temperature sensor, a humidity sensor, a vibration sensor and a noise sensor.

[0011] Preferably, a warning speaker is also integrated on the mounting substrate.

[0012] Preferably, a buckle is provided on the back side of the mounting substrate, and the buckle is hollow and has an opening at the rear side.

[0013] A pet behavior monitoring and correction system based on deep learning visual big model technology, the pet behavior monitoring and correction system based on deep learning visual big model technology includes: a visual acquisition and reasoning system, a cloud visual server, an application server, and a local end, wherein the local end includes a user end, an institution end, a database, and an inference server group;

[0014] Among them, the camera establishes a connection with the cloud video server, and the video server provides traditional video storage and playback services; the camera is connected to the edge computing terminal, and the video data stream is transmitted to the edge computing terminal. The terminal runs the Linux system or the Android system, calls the independently developed business application to organize the video data stream, and then inputs the deep learning visual model for reasoning and prediction; the signal output end of the sensor group establishes a connection with the application server, and the visual acquisition and reasoning system is integrated on the edge computing terminal. The edge computing terminal establishes a connection with the cloud vision server, the application server, and the local end.

[0015] A method for pet behavior monitoring and correction based on deep learning visual large model technology, the specific steps of the method for pet behavior monitoring and correction based on deep learning visual large model technology are as follows:

[0016] The information collected by the camera and sensor group is used by the visual acquisition and reasoning system using deep learning visual large model technology to achieve accurate recognition of pet behavior, including but not limited to eating, defecation, sleeping, and playing;

[0017] As the data processing center, the edge computing terminal is responsible for receiving the raw data from the smart door edge and using the powerful computing power of the NPU and the self-developed deep learning visual model algorithm to train and infer static image and dynamic video data.

[0018] The application server provides a user interface that allows pet owners to view their pet’s behavior reports and set behavior correction strategies;

[0019] Data storage and data mining analysis are performed locally. Small data storage devices are deployed in users' homes to store and quickly access historical data. Simple data mining analysis is performed to assist in optimizing behavior monitoring and correction algorithms. Server groups for reasoning learning and data storage management are deployed on the institutional side to manage user information, clean the massive amounts of stored images and video data, and build a data set suitable for visual model reasoning. Large model algorithms are optimized and generated on the institutional side, and the completed model is uploaded to the application server. An information transmission channel is established between the application server and the user side to complete the update and download of the visual model.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] Improves the accuracy and real-time performance of pet behavior monitoring.

[0022] It reduces the workload of pet owners and improves the convenience of pet behavior management.

[0023] Through data analysis, more personalized and scientific care recommendations can be provided for pets. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the structure of the present invention.

[0025] In the figure: 1. Mounting substrate; 2. Camera; 3. Sensor group; 4. Buckle. DETAILED DESCRIPTION

[0026] 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.

[0027] In the description of the present invention, it is necessary to understand that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0028] Embodiment 1:

[0029] See also Figure 1The present invention provides a technical solution: a device for monitoring and correcting pet behavior based on deep learning visual large model technology, comprising:

[0030] Install substrate 1, main control board, camera 2, sensor group 3, and edge computing terminal;

[0031] The mounting substrate 1 is provided with holes for mounting the camera 2 and the sensor group 3, and the camera 2 and the sensor group 3 are mounted in the holes corresponding to the mounting substrate 1;

[0032] The camera 2 and sensor group 3 are used to detect pet motion images, living environment information and motion information respectively. The main control board is used to control the normal operation of the system. The edge computing terminal is responsible for processing the video image data collected by the camera 2 and using the CNN convolution model for inference and prediction; and analyzing the data sent back by the sensor group 3. The main control board and the edge computing terminal are integrated in a device box and connected to the installation substrate 1 through a data bus.

[0033] The sensor group 3 includes a temperature sensor, a humidity sensor, a vibration sensor and a noise sensor. A warning speaker is also integrated on the mounting substrate 1. A buckle 4 is provided on the back of the mounting substrate 1, and the buckle 4 is hollow and has an opening at the rear.

[0034] The temperature sensor and humidity sensor are used to detect information about the pet's living environment, such as temperature and humidity. The vibration sensor and noise sensor are used to detect whether the pet cage is vibrating and whether there is noise, which serves as a basis for determining whether the pet is exercising.

[0035] Through buckle 4, the present solution can be installed on the pet fence door to provide a pet behavior monitoring and correction device based on deep learning visual large model technology. The device has a built-in camera and sensor for capturing pet behavior data.

[0036] Embedded software: The smart door edge device is equipped with embedded software that can initially process and analyze the collected data, such as identifying the time points of pet entry and exit, pet health and movement status, etc.

[0037] A pet behavior monitoring and correction system based on deep learning visual big model technology, the pet behavior monitoring and correction system based on deep learning visual big model technology is based on the pet behavior monitoring and correction device based on deep learning visual big model technology, the pet behavior monitoring and correction system based on deep learning visual big model technology includes: a visual acquisition and reasoning system, a cloud vision server, an application server, and a local end, wherein the local end includes a user end, an institution end, a database, and an inference server group;

[0038] Among them, the camera 2 establishes a connection with the cloud video server, and the video server provides traditional video storage and playback services; the camera 2 is connected to the edge computing terminal, and the video data stream is transmitted to the edge computing terminal. The terminal runs the Linux system or the Android system, calls the independently developed business application to sort the video data stream, and then inputs the deep learning visual model for reasoning and prediction; the signal output end of the sensor group 3 establishes a connection with the application server, and the visual acquisition and reasoning system is integrated on the edge computing terminal. The edge computing terminal establishes a connection with the cloud vision server, the application server, and the local end.

[0039] The visual acquisition and reasoning system integrates the image recognition model of deep learning. Image recognition refers to the technology of using computers to process, analyze and understand images to identify targets and objects of various different patterns. It is a practical application of deep learning algorithms. At present, image recognition technology is generally divided into face recognition and product recognition. Face recognition is mainly used in security inspection, identity verification and mobile payment; product recognition is mainly used in the process of commodity circulation, especially in unmanned retail fields such as unmanned shelves and smart retail cabinets.

[0040] Image recognition may be based on the main features of the image. Each image has its own features, such as the letter A has a point, P has a circle, and the center of the Y has an acute angle. Studies on eye movements during image recognition show that the line of sight is always focused on the main features of the image, that is, where the image contour has the largest curvature or where the contour direction suddenly changes. These places have the largest amount of information. Moreover, the scanning route of the eyes always switches from one feature to another in turn. It can be seen that in the process of image recognition, the perceptual mechanism must exclude the redundant input information and extract the key information. At the same time, there must be a mechanism in the brain responsible for integrating information, which can organize the information obtained in stages into a complete perceptual image.

[0041] In the human image recognition system, the recognition of complex images often requires information processing at different levels. For familiar graphics, since we have mastered its main features, we will recognize it as a unit and no longer pay attention to its details. This whole unit composed of isolated unit materials is called a block, and each block is perceived at the same time. In the recognition of text materials, people can not only group the strokes or radicals of a Chinese character into a block, but also group characters or words that often appear together into block units for recognition.

[0042] In computer vision recognition systems, image content is usually described by image features. In fact, image retrieval based on computer vision can also be divided into three steps similar to text search engines: feature extraction, index building, and query.

[0043] A method for pet behavior monitoring and correction based on deep learning visual large model technology, the method for pet behavior monitoring and correction based on deep learning visual large model technology is based on the pet behavior monitoring and correction system based on deep learning visual large model technology, and the specific steps of the method for pet behavior monitoring and correction based on deep learning visual large model technology are as follows:

[0044] The information collected by the camera 2 and the sensor group 3 is used by the visual acquisition and reasoning system using deep learning visual large model technology to achieve accurate recognition of pet behaviors, including but not limited to eating, excretion, sleeping, and playing;

[0045] As the data processing center, the edge computing terminal is responsible for receiving the raw data from the smart door edge and using the powerful computing power of the NPU and the self-developed deep learning visual model algorithm to train and infer static image and dynamic video data.

[0046] The application server provides a user interface that allows pet owners to view their pet’s behavior reports and set behavior correction strategies;

[0047] Data storage and data mining analysis are performed locally. Small data storage devices are deployed in users' homes to store and quickly access historical data. Simple data mining analysis is performed to assist in optimizing behavior monitoring and correction algorithms. Server groups for reasoning learning and data storage management are deployed on the institutional side to manage user information, clean the massive amounts of stored images and video data, and build a data set suitable for visual model reasoning. Large model algorithms are optimized and generated on the institutional side, and the completed model is uploaded to the application server. An information transmission channel is established between the application server and the user side to complete the update and download of the visual model.

[0048] Take the following table as an example to record data on pet monitoring status:

[0049] Statistics of pet entry and exit time

[0050]

[0051]

[0052] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention; therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is limited by the attached claims rather than the above description. Therefore, it is intended to include all changes within the meaning and scope of the equivalent elements of the claims in the present invention, and any figure marks in the claims should not be regarded as limiting the claims involved.

[0053] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A device for pet behavior monitoring and correction based on deep learning visual large model technology, characterized in that: include: Installing a substrate (1), a main control board, a camera (2), a sensor group (3), and an edge computing terminal; The mounting substrate (1) is provided with holes for mounting the camera (2) and the sensor group (3), and the camera (2) and the sensor group (3) are mounted in the holes corresponding to the mounting substrate (1); The camera (2) and the sensor group (3) are used to detect pet motion images, living environment information and motion information respectively; the main control board is used to control the normal operation of the system; the edge computing terminal is responsible for processing the video image data collected by the camera (2) and performing inference prediction using a CNN convolution model; and analyzing the data returned by the sensor group (3); the main control board and the edge computing terminal are integrated in a device box and connected to the mounting base plate (1) via a data bus.

2. A device for pet behavior monitoring and correction based on deep learning visual large model technology according to claim 1, characterized in that: The sensor group (3) comprises a temperature sensor, a humidity sensor, a vibration sensor and a noise sensor.

3. A device for pet behavior monitoring and correction based on deep learning visual large model technology according to claim 1, characterized in that: A warning speaker is also integrated on the installation substrate (1).

4. The device for pet behavior monitoring and correction based on deep learning visual large model technology according to claim 1, characterized in that: A buckle (4) is provided on the back side of the mounting substrate (1); the buckle (4) is hollow and has an opening at the rear side.

5. A system for pet behavior monitoring and correction based on deep learning visual large model technology, characterized by: The pet behavior monitoring and correction system based on deep learning visual big model technology is based on the pet behavior monitoring and correction device based on deep learning visual big model technology as described in any one of claims 1-4, and the pet behavior monitoring and correction system based on deep learning visual big model technology includes: a visual acquisition and reasoning system, a cloud vision server, an application server, and a local end, wherein the local end includes a user end, an institution end, a database, and an inference server group; The camera (2) is connected to a cloud video server, and the video server provides traditional video storage and playback services; the camera (2) is connected to an edge computing terminal, and the video data stream is transmitted to the edge computing terminal. The terminal runs a Linux system or an Android system, and calls a self-developed business application to sort the video data stream, and then inputs a deep learning visual model for inference and prediction; the signal output end of the sensor group (3) is connected to an application server, and the visual acquisition and inference system is integrated on the edge computing terminal. The edge computing terminal is connected to the cloud visual server, the application server, and the local end.

6. A method for pet behavior monitoring and correction based on deep learning visual large model technology, characterized in that: The method for pet behavior monitoring and correction based on deep learning visual large model technology is based on the system for pet behavior monitoring and correction based on deep learning visual large model technology according to claim 5. The specific steps of the method for pet behavior monitoring and correction based on deep learning visual large model technology are as follows: The information collected by the camera (2) and the sensor group (3) is used by the visual acquisition and reasoning system using deep learning visual large model technology to achieve accurate recognition of the pet's behavior, including but not limited to eating, defecation, sleeping, and playing; As the data processing center, the edge computing terminal is responsible for receiving the raw data from the smart door edge and using the powerful computing power of the NPU and the self-developed deep learning visual model algorithm to train and infer static image and dynamic video data. The application server provides a user interface that allows pet owners to view their pet’s behavior reports and set behavior correction strategies; Data storage and data mining analysis are performed locally. Small data storage devices are deployed in users' homes to store and quickly access historical data. Simple data mining analysis is performed to assist in optimizing behavior monitoring and correction algorithms. Server groups for reasoning learning and data storage management are deployed on the institutional side to manage user information, clean the massive amounts of stored images and video data, and build a data set suitable for visual model reasoning. Large model algorithms are optimized and generated on the institutional side, and the completed model is uploaded to the application server. An information transmission channel is established between the application server and the user side to complete the update and download of the visual model.

Images