Patrol robot based on computer vision and security and protection system

Abstract

The invention discloses a patrol robot based on computer vision. The visual language navigation module is combined with an environment image captured by a camera and a natural language processing technology, analyzes a human instruction and carries out navigation and positioning; the autonomous patrol and monitoring module is responsible for sensing the surrounding environment in real time, avoiding collision, optimizing a patrol route and monitoring animal behaviors and environment states at the same time; the intelligent identification and alarm module is responsible for intelligently identifying animal behaviors and judging animal states, and once abnormal conditions are detected, the patrol robot immediately sends alarm information; the environmental parameter monitoring and regulating module is responsible for monitoring environmental parameters in the farm in real time, and the patrol robot automatically or manually adjusts environmental equipment according to a monitoring result; according to the invention, through a visual language navigation technology, a voice instruction sent by a human can be accurately understood, an optimal patrol path is planned according to the voice instruction, and the problems of fixed path and insufficient flexibility in traditional navigation are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of patrol robots, and in particular to a patrol robot and a security system based on computer vision. Background Art

[0002] Computer vision is an important branch of artificial intelligence. It enables computer systems to simulate the functions of the human visual system and realize the perception, recognition and understanding of image or video data. With the continuous development of technologies such as deep learning and neural networks, computer vision has made significant progress in target detection, image recognition, image segmentation, etc., providing strong technical support for patrol robots and security systems. In the livestock industry, computer vision technology can be applied to individual animal identification, behavior monitoring, health status assessment, etc., providing strong support for the intelligent management of farms.

[0003] In the prior art, the path in traditional navigation is fixed and lacks flexibility. The fixed path of the traditional navigation system means that patrol robots or security equipment can only move and monitor along a preset route, which limits the improvement of management efficiency. In actual applications, the layout of the farm, the activity habits of animals, and emergencies may change, and the key areas that need to be patrolled and monitored will also be adjusted accordingly. However, the traditional navigation system cannot flexibly respond to these changes, resulting in limited management efficiency. Therefore, a patrol robot and security system based on computer vision are proposed. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art, that is, the fixed path and insufficient flexibility in traditional navigation, which leads to limited management efficiency, and to propose a patrol robot and security system based on computer vision.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A patrol robot based on computer vision, comprising:

[0007] Visual language navigation module: It combines the environmental images captured by the camera and natural language processing technology to interpret human instructions and perform navigation and positioning;

[0008] Autonomous patrol and monitoring module: responsible for real-time perception of the surrounding environment, avoiding collisions, and optimizing patrol routes. At the same time, cameras and infrared thermal imaging equipment are used to monitor animal behavior and environmental conditions, such as animal activity, body temperature changes, and environmental temperature and humidity;

[0009] Intelligent identification and alarm module: responsible for intelligent identification of animal behavior, judging whether the animal is sick or frightened, etc. Once an abnormal situation is detected, the patrol robot will immediately send an alarm message;

[0010] Environmental parameter monitoring and control module: responsible for real-time monitoring of environmental parameters in the farm, such as temperature, humidity, ammonia concentration, etc. According to the monitoring results, the patrol robot automatically or manually adjusts environmental equipment, such as turning on ventilation fans and adjusting lighting, to optimize the living environment of animals;

[0011] The visual language navigation module sends the parsed instruction information to the autonomous patrol and monitoring module to guide the autonomous patrol and monitoring module to patrol according to a preset route. The animal behavior and environmental status data collected by the autonomous patrol and monitoring module during the patrol process are fed back to the visual language navigation module for optimizing the navigation algorithm and path planning. The intelligent recognition and alarm module receives the environmental images and animal behavior and environmental status data provided by the visual language navigation module and the autonomous patrol and monitoring module for intelligent recognition. Once an abnormal situation is identified, the intelligent recognition and alarm module will trigger an alarm and quickly locate the abnormal position through the visual language navigation module and the path information recorded by the autonomous patrol and monitoring module. The environmental parameter monitoring and control module receives the real-time environmental status data provided by the autonomous patrol and monitoring module, and automatically or manually adjusts the environmental equipment (such as ventilation fans, lighting equipment, etc.) according to these data.

[0012] The above technical solution further includes:

[0013] Furthermore, the visual language navigation module includes a voice input unit, a natural language processing unit, a camera image capture unit, an image processing unit, a path planning unit and a positioning and navigation unit. The voice input unit is responsible for receiving voice commands issued by humans and converting them into digital signals. The natural language processing unit parses the digital signals transmitted by the voice input unit and extracts key information in the commands (such as target area, action, etc.). The camera image capture unit uses a camera to capture images of the patrol robot's surroundings and converts them into digital images that can be processed by a computer. The image processing unit processes the digital images transmitted by the camera image capture unit and extracts feature information (such as edges, corners, textures, etc.) that is useful for path planning and positioning. The path planning unit plans the best path to the target area based on the command information extracted by the natural language processing unit and the feature information extracted by the image processing unit. The positioning and navigation unit guides the patrol robot along the planned path based on the results of the path planning unit, and performs real-time positioning and adjustment in the process.

[0014] Furthermore, the natural language processing unit analyzes the digital signal transmitted by the voice input unit to extract key information in the instruction, specifically in the following steps:

[0015] Voice Recognition:

[0016] Decode digital signals and convert them into text form, using hidden Markov models to recognize words and phrases in speech;

[0017] Natural Language Understanding:

[0018] Perform semantic analysis on the recognized text to extract key information (such as the target area "living room" and the action "turn on the light"), which involves the use of natural language processing (NLP) technologies such as syntactic analysis and semantic role labeling;

[0019] Instruction execution:

[0020] Generate corresponding control instructions based on the extracted key information.

[0021] Furthermore, the autonomous patrol and monitoring module includes a sensor unit, a camera unit, a data processing unit, a control unit and a communication unit. The sensor unit includes a laser radar, an ultrasonic sensor, an infrared thermal imaging device, etc., which is used to perceive the state of animals and the environment in real time, monitor the temperature changes of animals through infrared thermal imaging devices, and monitor the temperature and humidity and other parameters of the environment through other sensors. The camera unit includes a high-definition camera for capturing the behavior and images of animals. The data processing unit is responsible for receiving and processing data from the sensor unit and the camera unit, performing environmental perception, behavior analysis and anomaly detection. The control unit controls the movement of the patrol robot, camera angle adjustment and other states according to the analysis results of the data processing unit. The communication unit is responsible for data transmission and communication with other modules.

[0022] Furthermore, the intelligent recognition and alarm module includes a data receiving unit, a data processing unit, an intelligent recognition unit, an alarm unit and a communication unit. The data receiving unit is responsible for receiving the environmental images and the animal behavior and environmental status data provided by the visual language navigation module and the autonomous patrol and monitoring module. The data processing unit performs data processing on the environmental images and the animal behavior and environmental status data. The intelligent recognition unit uses a deep learning algorithm to process and analyze the processed data to achieve intelligent recognition of animal behavior. The alarm unit detects abnormal conditions based on the results of intelligent recognition. The communication unit is responsible for data transmission and communication with other modules to ensure timely delivery of alarm information. The environmental parameter monitoring and control module receives the real-time environmental status data provided by the autonomous patrol and monitoring module, and automatically or manually adjusts environmental equipment (such as ventilation fans, lighting equipment, etc.) based on these data.

[0023] Furthermore, the environmental parameter monitoring and control module includes an environmental status data receiving unit, a data transmission unit, a central processing unit and a central processing unit. The environmental status data receives real-time environmental status data provided by the autonomous patrol and monitoring module. The data transmission unit is responsible for transmitting the environmental status data received by the environmental status data to the central processing unit. The central processing unit processes, analyzes and stores the received data, and generates control instructions according to preset strategies. The control unit controls the operation of environmental equipment, such as ventilation fans, lighting equipment, etc., according to the instructions of the central processing unit.

[0024] A patrol robot security system based on computer vision, comprising:

[0025] Remote monitoring and management module: responsible for remotely viewing and managing the patrol robot's real-time video, environmental parameters and alarm information, supporting remote adjustment of patrol routes, monitoring parameters and alarm rules, etc., so that managers can more conveniently manage the entire security system;

[0026] Data analysis and reporting module: responsible for analyzing animal health status and environmental change trends, providing a scientific basis for management decisions. Regularly generated reports (such as animal health reports, environmental parameter reports, etc.) can help managers fully understand the status of the farm and discover and solve problems in a timely manner;

[0027] Linkage module: Linkage with the farm's automatic feeding system and environmental control system to improve overall operational efficiency. For example, when an animal is detected to be sick, the system can automatically notify the veterinarian to conduct remote diagnosis and treatment or arrange on-site treatment; when environmental parameters exceed the normal range, the system can automatically adjust environmental equipment to optimize the animal's living environment;

[0028] The remote monitoring and management module receives real-time video, environmental parameters and alarm information provided by the visual language navigation module, the autonomous patrol and monitoring module and the intelligent identification and alarm module. The remote monitoring and management module accesses the historical data and report information stored in the data analysis and reporting module. The remote monitoring and management module communicates and coordinates with the system linkage module. The data analysis and reporting module regularly receives the animal behavior and environmental status data collected by the autonomous patrol and monitoring module. The intelligent identification and alarm module sends the identified abnormal conditions and alarm information to the data analysis and reporting module for storage and analysis. The environmental parameter monitoring and control module sends the monitored environmental parameter data to the data analysis and reporting module for storage and analysis.

[0029] Furthermore, the remote monitoring and management module includes a cloud platform unit, a data transmission unit and a user interface unit. The cloud platform unit serves as the management and monitoring center and is responsible for receiving and processing the data transmitted back by the patrol robot and providing remote viewing and management functions. The data transmission unit is responsible for transmitting the data collected by the patrol robot to the cloud platform in real time to ensure the timeliness and accuracy of the data. The user interface unit provides a friendly operating interface for managers, supports remote viewing of real-time video, environmental parameters, alarm information, etc., and allows remote configuration and management.

[0030] Furthermore, the data analysis and reporting module includes a data collection unit, a data preprocessing unit, a data analysis unit, a report generation unit and an anomaly detection and early warning unit. The data collection unit is responsible for receiving data collected by the patrol robot from various sensors and equipment in the farm, including animal vital signs data, environmental data, etc. The data preprocessing unit performs preprocessing operations such as cleaning and format conversion on the collected data to ensure data quality and accuracy. The data analysis unit uses LSTM to conduct in-depth mining and analysis on the preprocessed data to identify patterns and trends in the data. The report generation unit automatically generates various reports based on the analysis results and displays key information in the form of charts, tables, etc. The anomaly detection and early warning unit monitors abnormal conditions in the data in real time and issues early warning information in a timely manner.

[0031] Furthermore, the data analysis unit constructs an analysis model through LSTM, C t-1 , H t-1 and C t , H t Respectively represent the unit state and hidden layer state at the previous moment and the unit state and hidden layer state after update; x t Indicates the current input; f t 、i t and t Respectively represent the forget gate, input gate and output gate, wherein the forget gate is filtered by the parameter σ(σ∈(0,1)) and removes irrelevant historical information, the input gate controls the current unit information, and the output layer completes the update of the hidden layer and determines the amount of hidden and cyclic information;

[0032] The LSTM transfer process is as follows:

[0033] The forget gate processes the current input information x t and the short-term memory information H of the previous moment t-1 And output the forget vector f t , the forget gate formula is f t =σ(W f x t +U f Ht-1 +b f ), where W f Indicates H t-1 and x t The weight parameter, b f Indicates bias;

[0034] Superposition x t and H t-1 , the long-term memory candidate state is updated through the Tanh function, the input gate determines the amount of long-term memory retention, and the input gate formula is i t =σ(W i ×[H t-1 ,x t ]+b i );

[0035] Abandoning useless new information and adding useful new information completes the update of the cell state. The update formula is:

[0036] The output gate outputs information for the current unit's long-term memory and short-term memory, and the formula is:

[0037] o t =σ(W0×[H t-1 ,x t ]+b0)

[0038] Y t =H t =o t ×tanh(C t ).

[0039] The present invention has the following beneficial effects:

[0040] In the present invention, through visual language navigation technology, it is possible to accurately understand voice commands issued by humans, such as "go to area A of the pig house". Through semantic analysis and syntactic parsing, it is possible to accurately extract key information in the command and plan the optimal patrol path accordingly, thus avoiding the problems of fixed paths and insufficient flexibility in traditional navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a system block diagram of a patrol robot and security system based on computer vision proposed by the present invention. DETAILED DESCRIPTION

[0042] 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.

[0043] See also Figure 1 As shown, the present invention is a patrol robot based on computer vision, comprising:

[0044] Visual language navigation module: It combines the environmental images captured by the camera and natural language processing technology to interpret human instructions and perform navigation and positioning;

[0045] Autonomous patrol and monitoring module: responsible for real-time perception of the surrounding environment, avoiding collisions, and optimizing patrol routes. At the same time, cameras and infrared thermal imaging equipment are used to monitor animal behavior and environmental conditions, such as animal activity, body temperature changes, and environmental temperature and humidity;

[0046] Intelligent identification and alarm module: responsible for intelligent identification of animal behavior, judging whether the animal is sick or frightened, etc. Once an abnormal situation is detected, the patrol robot will immediately send an alarm message;

[0047] Environmental parameter monitoring and control module: responsible for real-time monitoring of environmental parameters in the farm, such as temperature, humidity, ammonia concentration, etc. According to the monitoring results, the patrol robot automatically or manually adjusts environmental equipment, such as turning on ventilation fans and adjusting lighting, to optimize the living environment of animals;

[0048] The visual language navigation module sends the parsed command information to the autonomous patrol and monitoring module to guide the autonomous patrol and monitoring module to patrol according to the preset route. The animal behavior and environmental status data collected by the autonomous patrol and monitoring module during the patrol process are fed back to the visual language navigation module for optimizing the navigation algorithm and path planning. The intelligent recognition and alarm module receives the environmental images and animal behavior and environmental status data provided by the visual language navigation module and the autonomous patrol and monitoring module for intelligent recognition. Once an abnormal situation is identified, the intelligent recognition and alarm module will trigger an alarm and quickly locate the abnormal position through the visual language navigation module and the path information recorded by the autonomous patrol and monitoring module. The environmental parameter monitoring and control module receives the real-time environmental status data provided by the autonomous patrol and monitoring module, and automatically or manually adjusts the environmental equipment (such as ventilation fans, lighting equipment, etc.) according to these data.

[0049] In one embodiment, for the above-mentioned visual language navigation module, the visual language navigation module includes a voice input unit, a natural language processing unit, a camera image capture unit, an image processing unit, a path planning unit and a positioning and navigation unit. The voice input unit is responsible for receiving voice commands issued by humans and converting them into digital signals. The natural language processing unit parses the digital signals transmitted by the voice input unit and extracts key information in the commands (such as target area, action, etc.). The camera image capture unit uses a camera to capture images of the patrol robot's surroundings and converts them into digital images that can be processed by a computer. The image processing unit processes the digital images transmitted by the camera image capture unit and extracts feature information (such as edges, corners, textures, etc.) that is useful for path planning and positioning. The path planning unit plans the best path to the target area based on the command information extracted by the natural language processing unit and the feature information extracted by the image processing unit. The positioning and navigation unit: guides the patrol robot along the planned path according to the results of the path planning unit, and performs real-time positioning and adjustment in the process.

[0050] In one embodiment, for the above-mentioned natural language processing unit, the natural language processing unit analyzes the digital signal transmitted by the voice input unit to extract key information in the instruction, specifically, the following steps:

[0051] Voice Recognition:

[0052] Decode digital signals and convert them into text form, using hidden Markov models to recognize words and phrases in speech;

[0053] Natural Language Understanding:

[0054] Perform semantic analysis on the recognized text to extract key information (such as the target area "living room" and the action "turn on the light"), which involves the use of natural language processing (NLP) technologies such as syntactic analysis and semantic role labeling;

[0055] Instruction execution:

[0056] Generate corresponding control instructions based on the extracted key information.

[0057] In one embodiment, for the above-mentioned autonomous patrol and monitoring module, the autonomous patrol and monitoring module includes a sensor unit, a camera unit, a data processing unit, a control unit and a communication unit. The sensor unit includes a laser radar, an ultrasonic sensor, an infrared thermal imaging device, etc., which is used to perceive the state of animals and the environment in real time, monitor the temperature changes of animals through infrared thermal imaging devices, and monitor the temperature and humidity and other parameters of the environment through other sensors. The camera unit includes a high-definition camera for capturing the behavior and images of animals. The data processing unit is responsible for receiving and processing data from the sensor unit and the camera unit, and performing environmental perception, behavior analysis and anomaly detection. The control unit controls the movement of the patrol robot, camera angle adjustment and other states according to the analysis results of the data processing unit. The communication unit is responsible for data transmission and communication with other modules.

[0058] In one embodiment, for the above-mentioned intelligent recognition and alarm module, the intelligent recognition and alarm module includes a data receiving unit, a data processing unit, an intelligent recognition unit, an alarm unit and a communication unit. The data receiving unit is responsible for receiving the environmental images and animal behavior and environmental status data provided by the visual language navigation module and the autonomous patrol and monitoring module. The data processing unit processes the environmental images and animal behavior and environmental status data. The intelligent recognition unit uses a deep learning algorithm to process and analyze the processed data to achieve intelligent recognition of animal behavior. The alarm unit detects abnormal conditions based on the results of intelligent recognition. The communication unit is responsible for data transmission and communication with other modules to ensure timely delivery of alarm information. The environmental parameter monitoring and control module receives the real-time environmental status data provided by the autonomous patrol and monitoring module, and automatically or manually adjusts environmental equipment (such as ventilation fans, lighting equipment, etc.) based on these data.

[0059] In one embodiment, for the above-mentioned environmental parameter monitoring and control module, the environmental parameter monitoring and control module includes an environmental status data receiving unit, a data transmission unit, a central processing unit and a central processing unit. The environmental status data receives the real-time environmental status data provided by the autonomous patrol and monitoring module. The data transmission unit is responsible for transmitting the environmental status data received by the environmental status data to the central processing unit. The central processing unit processes, analyzes and stores the received data, and generates control instructions according to preset strategies. The control unit controls the operation of environmental equipment, such as ventilation fans, lighting equipment, etc., according to the instructions of the central processing unit.

[0060] A security system, comprising:

[0061] Remote monitoring and management module: responsible for remotely viewing and managing the patrol robot's real-time video, environmental parameters and alarm information, supporting remote adjustment of patrol routes, monitoring parameters and alarm rules, etc., so that managers can more conveniently manage the entire security system;

[0062] Data analysis and reporting module: responsible for analyzing animal health status and environmental change trends, providing a scientific basis for management decisions. Regularly generated reports (such as animal health reports, environmental parameter reports, etc.) can help managers fully understand the status of the farm and discover and solve problems in a timely manner;

[0063] Linkage module: Linkage with the farm's automatic feeding system and environmental control system to improve overall operational efficiency. For example, when an animal is detected to be sick, the system can automatically notify the veterinarian to conduct remote diagnosis and treatment or arrange on-site treatment; when environmental parameters exceed the normal range, the system can automatically adjust environmental equipment to optimize the animal's living environment;

[0064] The remote monitoring and management module receives real-time video, environmental parameters and alarm information provided by the visual language navigation module, the autonomous patrol and monitoring module and the intelligent identification and alarm module. The remote monitoring and management module accesses the historical data and report information stored in the data analysis and reporting module. The remote monitoring and management module communicates and coordinates with the system linkage module. The data analysis and reporting module regularly receives the animal behavior and environmental status data collected by the autonomous patrol and monitoring module. The intelligent identification and alarm module sends the identified abnormal conditions and alarm information to the data analysis and reporting module for storage and analysis. The environmental parameter monitoring and control module sends the monitored environmental parameter data to the data analysis and reporting module for storage and analysis

[0065] In one embodiment, for the above-mentioned remote monitoring and management module, the remote monitoring and management module includes a cloud platform unit, a data transmission unit and a user interface unit. The cloud platform unit serves as the management and monitoring center and is responsible for receiving and processing the data transmitted back by the patrol robot and providing remote viewing and management functions. The data transmission unit is responsible for transmitting the data collected by the patrol robot to the cloud platform in real time to ensure the timeliness and accuracy of the data. The user interface unit provides a friendly operating interface for management personnel, supports remote viewing of real-time video, environmental parameters, alarm information, etc., and allows remote configuration and management.

[0066] In one embodiment, for the above-mentioned data analysis and reporting module, the data analysis and reporting module includes a data collection unit, a data preprocessing unit, a data analysis unit, a report generation unit and an anomaly detection and early warning unit. The data collection unit is responsible for receiving data collected by the patrol robot from various sensors and equipment in the farm, including animal vital signs data, environmental data, etc. The data preprocessing unit performs preprocessing operations such as cleaning and format conversion on the collected data to ensure data quality and accuracy. The data analysis unit uses LSTM to conduct in-depth mining and analysis on the preprocessed data to identify patterns and trends in the data. The report generation unit automatically generates various reports based on the analysis results, and displays key information in the form of charts, tables, etc. The anomaly detection and early warning unit monitors abnormal conditions in the data in real time and issues early warning information in a timely manner.

[0067] In one embodiment, for the above data analysis unit, the data analysis unit constructs an analysis model through LSTM, C t-1 , H t-1 and C t , H t Respectively represent the unit state and hidden layer state at the previous moment and the unit state and hidden layer state after update; x t Indicates the current input; f t 、i t and t They represent the forget gate, input gate, and output gate, respectively. The forget gate is filtered by the parameter σ(σ∈(0,1)) and removes irrelevant historical information. The input gate controls the current unit information. The output layer completes the update of the hidden layer and determines the amount of hidden and cyclic information.

[0068] The LSTM transfer process is as follows:

[0069] The forget gate processes the current input information x t and the short-term memory information H of the previous moment t-1 And output the forget vector f t , the forget gate formula is f t =σ(W f x t +U f H t-1 +b f ), where W f Indicates H t-1 and x t The weight parameter, b f Indicates bias;

[0070] Superposition x t and H t-1 , the long-term memory candidate state is updated through the Tanh function, and the input gate determines the amount of long-term memory retention. The input gate formula is it =σ(W i ×[H t-1 ,x t ]+b i );

[0071] Abandoning useless new information and adding useful new information completes the update of cell state. The update formula is:

[0072]

[0073] The output gate outputs information for the current unit's long-term memory and short-term memory. The formula is:

[0074] o t =σ(W0×[H t-1 ,x t ]+b0)

[0075] Y t =H t =o t ×tanh(C t ).

[0076] 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 patrol robot based on computer vision, characterized in that: include; Visual language navigation module: It combines the environmental images captured by the camera and natural language processing technology to interpret human instructions and perform navigation and positioning; Autonomous patrol and monitoring module: responsible for real-time perception of the surrounding environment, avoiding collisions, and optimizing patrol routes, while monitoring animal behavior and environmental conditions; Intelligent identification and alarm module: responsible for intelligent identification of animal behavior and judging the animal status. Once an abnormal situation is detected, the patrol robot will immediately send an alarm message; Environmental parameter monitoring and control module: responsible for real-time monitoring of environmental parameters in the farm. According to the monitoring results, the patrol robot automatically or manually adjusts the environmental equipment; The visual language navigation module sends the parsed instruction information to the autonomous patrol and monitoring module. The animal behavior and environmental status data collected by the autonomous patrol and monitoring module during the patrol process are fed back to the visual language navigation module for optimizing the navigation algorithm and path planning. The intelligent recognition and alarm module receives the environmental images and animal behavior and environmental status data provided by the visual language navigation module and the autonomous patrol and monitoring module for intelligent recognition. Once an abnormal situation is identified, the intelligent recognition and alarm module will trigger an alarm and quickly locate the abnormal position through the visual language navigation module and the path information recorded by the autonomous patrol and monitoring module. The environmental parameter monitoring and control module receives the real-time environmental status data provided by the autonomous patrol and monitoring module.

2. A patrol robot based on computer vision according to claim 1, characterized in that: The visual language navigation module includes a voice input unit, a natural language processing unit, a camera image capture unit, an image processing unit, a path planning unit and a positioning and navigation unit. The voice input unit is responsible for receiving voice commands issued by humans and converting them into digital signals. The natural language processing unit parses the digital signals transmitted by the voice input unit and extracts key information in the commands. The camera image capture unit uses a camera to capture images of the patrol robot's surroundings and converts them into digital images that can be processed by a computer. The image processing unit processes the digital images transmitted by the camera image capture unit and extracts feature information useful for path planning and positioning. The path planning unit plans the best path to the target area based on the command information extracted by the natural language processing unit and the feature information extracted by the image processing unit. The positioning and navigation unit guides the patrol robot along the planned path based on the results of the path planning unit, and performs real-time positioning and adjustment in the process.

3. A patrol robot based on computer vision according to claim 2, characterized in that: The natural language processing unit analyzes the digital signal transmitted from the voice input unit to extract key information from the instruction, specifically in the following steps: Voice Recognition: Decode digital signals and convert them into text form, using hidden Markov models to recognize words and phrases in speech; Natural Language Understanding: Perform semantic analysis on the recognized text to extract key information; Instruction execution: Generate corresponding control instructions based on the extracted key information.

4. A patrol robot based on computer vision according to claim 1, characterized in that: The autonomous patrol and monitoring module includes a sensor unit, a camera unit, a data processing unit, a control unit and a communication unit. The sensor unit is used to perceive the state of animals and the environment in real time. The camera unit is used to capture the behavior and images of animals. The data processing unit is responsible for receiving and processing data from the sensor unit and the camera unit to perform environmental perception, behavior analysis and anomaly detection. The control unit controls the state of the patrol robot according to the analysis results of the data processing unit. The communication unit is responsible for data transmission and communication with other modules.

5. The patrol robot based on computer vision according to claim 1, characterized in that: The intelligent recognition and alarm module includes a data receiving unit, a data processing unit, an intelligent recognition unit, an alarm unit and a communication unit. The data receiving unit is responsible for receiving the environmental images and the animal's behavior and environmental status data provided by the visual language navigation module and the autonomous patrol and monitoring module. The data processing unit processes the environmental images and the animal's behavior and environmental status data. The intelligent recognition unit uses a deep learning algorithm to process and analyze the processed data to achieve intelligent recognition of animal behavior. The alarm unit detects abnormal conditions based on the results of intelligent recognition. The communication unit is responsible for data transmission and communication with other modules. The environmental parameter monitoring and control module receives the real-time environmental status data provided by the autonomous patrol and monitoring module.

6. The patrol robot based on computer vision according to claim 1, characterized in that: The environmental parameter monitoring and control module includes an environmental status data receiving unit, a data transmission unit, a central processing unit and a central processing unit. The environmental status data receives the real-time environmental status data provided by the autonomous patrol and monitoring module. The data transmission unit is responsible for transmitting the environmental status data received by the environmental status data to the central processing unit. The central processing unit processes, analyzes and stores the received data, and generates control instructions according to preset strategies. The control unit controls the operation of the environmental equipment according to the instructions of the central processing unit.

7. The security system of a patrol robot based on computer vision according to claim 1, characterized in that: include: Remote monitoring and management module: responsible for remotely viewing and managing the patrol robot's real-time video, environmental parameters and alarm information; Data analysis and reporting module: responsible for analyzing animal health status and environmental change trends, providing scientific basis for management decisions, and generating reports regularly; Linkage module: Linkage with the automatic feeding system and environmental control system of the farm; The remote monitoring and management module receives real-time video, environmental parameters and alarm information provided by the visual language navigation module, the autonomous patrol and monitoring module and the intelligent identification and alarm module. The remote monitoring and management module accesses the historical data and report information stored in the data analysis and reporting module. The remote monitoring and management module communicates and coordinates with the system linkage module. The data analysis and reporting module regularly receives the animal behavior and environmental status data collected by the autonomous patrol and monitoring module. The intelligent identification and alarm module sends the identified abnormal conditions and alarm information to the data analysis and reporting module for storage and analysis. The environmental parameter monitoring and control module sends the monitored environmental parameter data to the data analysis and reporting module for storage and analysis.

8. The computer vision-based patrol robot security system according to claim 7, characterized in that: The remote monitoring and management module includes a cloud platform unit, a data transmission unit and a user interface unit. The cloud platform unit is responsible for receiving and processing the data sent back by the patrol robot and providing remote viewing and management functions. The data transmission unit is responsible for transmitting the data collected by the patrol robot to the cloud platform in real time. The user interface unit provides a friendly operation interface for managers.

9. The security system of a patrol robot based on computer vision according to claim 7, characterized in that: The data analysis and reporting module includes a data collection unit, a data preprocessing unit, a data analysis unit, a report generation unit and an anomaly detection and early warning unit. The data collection unit is responsible for receiving data from various sensors and equipment in the farm collected by the patrol robot. The data preprocessing unit performs preprocessing operations on the collected data. The data analysis unit uses LSTM to conduct in-depth mining and analysis on the preprocessed data to identify patterns and trends in the data. The report generation unit automatically generates various reports based on the analysis results. The anomaly detection and early warning unit monitors abnormal situations in the data in real time and issues early warning information in a timely manner.

10. The security system of patrol robot based on computer vision according to claim 9, characterized in that: The data analysis unit constructs an analysis model through LSTM. t-1 , H t-1 and C t , H t Respectively represent the unit state and hidden layer state at the previous moment and the unit state and hidden layer state after update; x t Indicates the current input; f t 、i t and t Respectively represent the forget gate, input gate and output gate, wherein the forget gate is filtered by the parameter σ(σ∈(0,1)) and removes irrelevant historical information, the input gate controls the current unit information, and the output layer completes the update of the hidden layer and determines the amount of hidden and cyclic information; The LSTM transfer process is as follows: The forget gate processes the current input information x t and the short-term memory information H of the previous moment t-1 And output the forget vector f t , the forget gate formula is f t =σ(W f x t +U f H t-1 +b f ), where W f Indicates H t-1 and x t The weight parameter, b f Indicates bias; Superposition x t and H t-1 , the long-term memory candidate state is updated through the Tanh function, the input gate determines the amount of long-term memory retention, and the input gate formula is i t =σ(W i ×[H t-1 ,x t ]+b i ); Abandoning useless new information and adding useful new information completes the update of the cell state. The update formula is: The output gate outputs information for the current unit's long-term memory and short-term memory, and the formula is: about t =σ(W0×[H t-1 ,x t ]+b0) Y t =H t =o t ×tanh(C t )。

Images