An object movement monitoring system, method and device based on AI visual recognition

Abstract

The present invention proposes an object movement monitoring system, method and equipment based on AI visual recognition, which relate to the technical field of object movement monitoring. A mobile monitoring platform controls a movable camera to track and collect image data of a target moving object in real time; the mobile monitoring platform uses a deep learning model to identify the image data and obtain the position information of the target object in the image data; the mobile monitoring platform adjusts the moving speed of the movable camera according to the position information of the target object in the image data. The present invention improves the real-time, accuracy, intelligence, adaptability and scalability of the object movement monitoring equipment.

Description

Technical Field

[0001] The present invention proposes an object movement monitoring system, method and device based on AI visual recognition, which relate to the technical field of object movement monitoring, and in particular to the technical field of object movement monitoring based on AI visual recognition. Background Art

[0002] With the rapid development of science and technology, especially the significant progress of artificial intelligence (AI) and computer vision technology, the field of object movement monitoring is undergoing unprecedented changes. Traditional object movement monitoring equipment often relies on manual observation, infrared sensing or radar scanning. Although these methods can realize the movement monitoring of objects to a certain extent, they have obvious limitations in terms of real-time, accuracy, and intelligence. Traditional monitoring equipment has large delays in data collection, transmission and processing, and cannot realize real-time tracking and monitoring of object movement. Traditional monitoring equipment often relies on simple algorithms or manual operations in terms of target recognition and location determination, and is easily affected by factors such as illumination, occlusion, and changes in target morphology, resulting in low recognition accuracy. Traditional monitoring equipment can usually only perform simple data collection and display, and cannot realize intelligent analysis and processing of data, and cannot meet application requirements in complex scenarios. Summary of the invention

[0003] The present invention provides an object movement monitoring system, method and device based on AI visual recognition, which are used to solve the problems that traditional monitoring equipment has large delays in data collection, transmission and processing, and cannot realize real-time tracking and monitoring of object movement. Monitoring equipment often relies on simple algorithms or manual operations in terms of target recognition and position determination, and is easily affected by factors such as illumination, occlusion, and target morphology changes, resulting in low recognition accuracy. Usually, only simple data collection and display can be performed, and intelligent analysis and processing of data cannot be realized, and application requirements in complex scenarios cannot be met.

[0004] The present invention proposes a system, method and device for monitoring object movement based on AI visual recognition. The method for monitoring object movement based on AI visual recognition includes:

[0005] The mobile monitoring platform controls the movable camera to track and collect image data of the target moving object in real time;

[0006] The mobile monitoring platform uses a deep learning model to identify image data and obtain the location information of the target object in the image data;

[0007] The mobile monitoring platform adjusts the moving speed of the movable camera according to the position information of the target object in the image data.

[0008] Furthermore, the mobile monitoring platform controls the movable camera to track and collect image data of the target moving object in real time, including:

[0009] The movable camera collects initial image data of the target area in real time, and sends the initial image data to the mobile control platform;

[0010] After receiving the initial image data, the mobile control platform identifies the initial image data and determines whether there is a target moving object in the current initial image data;

[0011] When there is a target moving object in the initial image data, the mobile control platform sends a control instruction to the movable camera to control the movable camera to track and shoot the target moving object, and obtain image data of the target moving object;

[0012] The movable camera transmits the image data of the target moving object back to the mobile control platform.

[0013] Furthermore, the mobile monitoring platform uses a deep learning model to identify the image data and obtain the location information of the target object in the image data, including:

[0014] After receiving the image data of the target moving object transmitted back by the movable camera, the mobile control platform inputs the image data into the deep learning model;

[0015] The target object is identified through the deep learning model to obtain the position information of the target object in the image data.

[0016] Furthermore, the mobile monitoring platform adjusts the moving speed of the movable camera according to the position information of the target object in the image data, including:

[0017] The mobile monitoring platform extracts the position information of the target object in each image data;

[0018] The mobile monitoring platform obtains the moving speed compensation amount of the movable camera according to the position information of the target object in the image data; wherein the moving speed compensation amount is obtained by the following formula:

[0019]

[0020]

[0021]

[0022] Among them, ξ v Indicates the movement speed compensation amount; S 01 and S02 represent the first adjustment coefficient and the second adjustment coefficient respectively; n represents the total number of image data captured by the movable camera; S i represents the overlapping area between the target object in the i-th image data and the target object in the previous image data; S e represents the area occupied by the target object in the image data; m represents the number of image data in which the target object is in the front position compared with the target object in the previous image data; k represents the number of image data in which the target object is in the back position compared with the target object in the previous image data; S xi represents the overlapping area of ​​the target object in the i-th previous image data compared with the target object in the previous image data; S hi represents the overlapping area of ​​the target object in the i-th subsequent image data compared with the target object in the previous image data;

[0023] The mobile monitoring platform uses the mobile speed compensation to obtain the target speed; wherein the target speed is obtained by the following formula:

[0024] V=(1+ξ v )·V0

[0025] Wherein, V represents the target speed; V0 represents the tracking speed before adjustment;

[0026] The mobile monitoring platform adjusts the tracking speed of the movable camera according to the target speed.

[0027] Furthermore, the object movement monitoring system based on AI visual recognition includes:

[0028] Image data acquisition module, used for the mobile monitoring platform to control the movable camera to track and collect image data of the target moving object in real time;

[0029] Image data recognition module, used for the mobile monitoring platform to identify image data using a deep learning model and obtain the location information of the target object in the image data;

[0030] The moving speed adjustment module is used for the mobile monitoring platform to adjust the moving speed of the movable camera according to the position information of the target object in the image data.

[0031] Furthermore, the image data acquisition module includes:

[0032] An initial data acquisition module, used for the movable camera to acquire the initial image data of the target area in real time, and send the initial image data to the mobile control platform;

[0033] A movement judgment module, used for the mobile control platform to identify the initial image data after receiving the initial image data, and to judge whether there is a target moving object in the current initial image data;

[0034] A camera control module, configured to, when there is a target moving object in the initial image data, cause the mobile control platform to send a control instruction to the movable camera, control the movable camera to track and shoot the target moving object, and obtain image data of the target moving object;

[0035] The data transmission module is used for the movable camera to transmit the image data of the target moving object back to the mobile control platform.

[0036] Furthermore, the image data recognition module includes:

[0037] A model input module, used for the mobile control platform to input the image data of the target moving object transmitted back by the movable camera into the deep learning model after receiving the image data;

[0038] The position acquisition module is used to identify the target object through the deep learning model and obtain the position information of the target object in the image data.

[0039] Furthermore, the moving speed adjustment module includes:

[0040] A moving speed compensation acquisition module, used for the mobile monitoring platform to extract the position information of the target object in each image data;

[0041] The mobile monitoring platform obtains the moving speed compensation amount of the movable camera according to the position information of the target object in the image data; wherein the moving speed compensation amount is obtained by the following formula:

[0042]

[0043]

[0044]

[0045] Among them, ξ v Indicates the movement speed compensation amount; S 01 and S 02 represent the first adjustment coefficient and the second adjustment coefficient respectively; n represents the total number of image data captured by the movable camera; S i represents the overlapping area between the target object in the i-th image data and the target object in the previous image data; S erepresents the area occupied by the target object in the image data; m represents the number of image data in which the target object is in the front position compared with the target object in the previous image data; k represents the number of image data in which the target object is in the back position compared with the target object in the previous image data; S xi represents the overlapping area of ​​the target object in the i-th previous image data compared with the target object in the previous image data; S hi represents the overlapping area of ​​the target object in the i-th subsequent image data compared with the target object in the previous image data;

[0046] The target speed acquisition module is used for the mobile monitoring platform to obtain the target speed using the mobile speed compensation amount; wherein the target speed is obtained by the following formula:

[0047] V=(1+ξ v )·V0

[0048] Wherein, V represents the target speed; V0 represents the tracking speed before adjustment;

[0049] The adjustment module is used for the mobile monitoring platform to adjust the tracking speed of the movable camera according to the target speed.

[0050] Furthermore, the object movement monitoring device based on AI visual recognition includes a mobile monitoring platform and a movable camera; wherein, the mobile monitoring platform and the movable camera are connected by data signals via wireless communication.

[0051] Furthermore, the mobile monitoring platform includes a processor, wherein the processor is used to execute the object movement monitoring method based on AI visual recognition described in claims 1 to 4.

[0052] Beneficial effects of the present invention: Through the real-time tracking and acquisition of the movable camera and the fast and accurate recognition of the deep learning model, the mobile monitoring platform can realize the real-time tracking and positioning of the target object, greatly improving the real-time and accuracy of the monitoring. Due to the use of deep learning technology, the mobile monitoring platform can cope with the target object recognition tasks in various complex scenarios. At the same time, by adjusting the moving speed and direction of the camera, the platform can flexibly adapt to the movement changes of the target object to ensure a continuous and stable tracking effect. The entire workflow of the mobile monitoring platform is intelligent and automated, and the real-time tracking and positioning of the target object can be completed without human intervention. This not only reduces labor costs, but also improves work efficiency and safety. The platform can be expanded and customized according to actual needs, such as adding more hardware equipment such as cameras and sensors, or optimizing the algorithms and parameters of the deep learning model. This gives the platform a wider application prospect and higher flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A schematic diagram of an object movement monitoring method based on AI visual recognition;

[0054] Figure 2 Schematic diagram of an object movement monitoring system based on AI visual recognition. DETAILED DESCRIPTION

[0055] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0056] In one embodiment of the present invention, a system, method and device for monitoring object movement based on AI visual recognition are provided. The method for monitoring object movement based on AI visual recognition includes:

[0057] The mobile monitoring platform controls the movable camera to track and collect image data of the target moving object in real time;

[0058] The mobile monitoring platform uses a deep learning model to identify image data and obtain the location information of the target object in the image data;

[0059] The mobile monitoring platform adjusts the moving speed of the movable camera according to the position information of the target object in the image data.

[0060] The working principle of the above technical solution is as follows: the mobile monitoring platform first tracks and collects image data of the target moving object in real time through a movable camera. This involves the motion control of the camera, such as electric zoom, pan / tilt rotation, etc., to ensure that the camera can always aim at and capture the target object. The collected image data is then fed into the deep learning model for recognition. After being trained with a large amount of labeled data, the deep learning model can learn the feature representation of different objects in the image, thereby achieving high-accuracy object recognition. In the target detection task, the deep learning model can detect the position (such as bounding box coordinates) and category of the target object in the image. Based on the target object position information identified by the deep learning model, the mobile monitoring platform calculates the moving speed and direction of the target object relative to the camera. Then, the platform adjusts the moving speed of the movable camera based on this information to ensure that the camera can continuously and stably track the target object and always place it in the center or a specific position of the image.

[0061] The effect of the above technical solution is: through the real-time tracking and acquisition of the movable camera and the fast and accurate recognition of the deep learning model, the mobile monitoring platform can realize the real-time tracking and positioning of the target object, greatly improving the real-time and accuracy of the monitoring. Due to the use of deep learning technology, the mobile monitoring platform can cope with the target object recognition tasks in various complex scenarios. At the same time, by adjusting the moving speed and direction of the camera, the platform can flexibly adapt to the movement changes of the target object to ensure a continuous and stable tracking effect. The entire workflow of the mobile monitoring platform is intelligent and automated, and the real-time tracking and positioning of the target object can be completed without human intervention. This not only reduces labor costs, but also improves work efficiency and safety. The platform can be expanded and customized according to actual needs, such as adding more hardware equipment such as cameras and sensors, or optimizing the algorithms and parameters of the deep learning model. This gives the platform a wider application prospect and higher flexibility.

[0062] In one embodiment of the present invention, a mobile monitoring platform controls a movable camera to track and collect image data of a target moving object in real time, including:

[0063] The movable camera collects initial image data of the target area in real time, and sends the initial image data to the mobile control platform;

[0064] After receiving the initial image data, the mobile control platform identifies the initial image data and determines whether there is a target moving object in the current initial image data;

[0065] When there is a target moving object in the initial image data, the mobile control platform sends a control instruction to the movable camera to control the movable camera to track and shoot the target moving object, and obtain image data of the target moving object;

[0066] The movable camera transmits the image data of the target moving object back to the mobile control platform.

[0067] The working principle of the above technical solution is as follows: the movable camera is deployed in the target area to be monitored, and the initial image data of the area is collected in real time. These image data may contain static background elements and possible dynamic target moving objects. The collected initial image data is then transmitted to the mobile control platform. The transmission can be achieved through wireless communication technology (such as Wi-Fi, 4G / 5G, etc.) to ensure the real-time and efficient data. After receiving the initial image data, the mobile control platform will use a pre-trained deep learning model or other image recognition algorithm to process the image. The purpose of this step is to determine whether there is a target moving object in the current image. If the target moving object is identified in the initial image data, the mobile control platform will further analyze the position, size and movement direction of the target object in the image. Based on the detection information of the target object, the mobile control platform will calculate the parameters of the movable camera that need to be adjusted and generate corresponding control instructions. These instructions are sent to the movable camera via wireless communication. After receiving the control instruction, the movable camera will adjust its own parameters according to the instruction requirements and track and shoot the target moving object. The captured image data will be transmitted back to the mobile control platform again. After the image data of the target moving object is collected by the mobile camera, it is transmitted back to the mobile control platform in real time through wireless communication technology. The mobile control platform can further analyze, store or display the data.

[0068] The effect of the above technical solution is: through the real-time acquisition of the movable camera and the rapid processing of the mobile control platform, the system can realize the real-time tracking and positioning of the target moving object, ensuring the real-time and accuracy of the data. The entire tracking and shooting process is intelligent and automated, and can be completed without human intervention. This reduces labor costs, improves work efficiency, and reduces the possibility of human errors. The system can be flexibly configured and expanded according to actual needs. For example, more movable cameras can be added to cover a wider area, or the recognition algorithm can be optimized to improve recognition accuracy and speed. By tracking and monitoring the target moving object in real time, the system can be used in security monitoring, traffic management and other fields to provide timely and effective data support to help improve safety and reliability. The image data of the target moving object that is continuously transmitted back provides a rich data source for subsequent data analysis, decision support, etc., which helps to improve the application value and effect of the entire system.

[0069] In one embodiment of the present invention, a mobile monitoring platform uses a deep learning model to identify image data and obtain location information of a target object in the image data, including:

[0070] After receiving the image data of the target moving object transmitted back by the movable camera, the mobile control platform inputs the image data into the deep learning model;

[0071] The target object is identified through the deep learning model to obtain the position information of the target object in the image data.

[0072] The working principle of the above technical solution is as follows: the mobile control platform first receives image data from the movable camera, which has been preliminarily screened and processed and contains image information of the target moving object. The mobile control platform loads a pre-trained deep learning model. Before inputting the image data into the deep learning model, some preprocessing steps may be required, such as resizing, normalization, denoising, etc., to meet the input requirements of the model and improve the recognition effect of the model. The deep learning model processes the input image data, extracts image features through structures such as convolutional layers, pooling layers, and fully connected layers, and applies the trained weight parameters for forward propagation calculations, and finally outputs the location information of the target object in the image. The mobile control platform receives the output result of the deep learning model, that is, the location information of the target object in the image, and performs subsequent processing or display.

[0073] The effect of the above technical solution is that the deep learning model can accurately identify the target object in the image and output accurate location information after training and optimization with a large amount of data. This is very important for application scenarios that require accurate tracking and positioning of target objects. The calculation speed of the deep learning model is usually very fast, and it can process a large amount of image data in a short time and output recognition results. This enables the mobile control platform to receive and process the image data of the movable camera in real time, meeting application scenarios with high real-time requirements. The deep learning model has strong robustness to factors such as image noise, illumination changes, and occlusion, and can maintain a stable recognition effect under complex environmental conditions. This has good adaptability to unpredictable factors in practical applications. The deep learning model has good scalability and can be customized and optimized according to the needs of practical applications. For example, the recognition accuracy and performance of the model can be improved by increasing training data, adjusting model parameters, optimizing algorithms, etc. Combined with the target object recognition function of the deep learning model, the mobile control platform can realize intelligent management of the target moving object. For example, the shooting angle and focal length of the camera can be automatically adjusted according to the recognition results, or the linkage operation of other related devices can be triggered to achieve a higher level of intelligent monitoring and management.

[0074] In one embodiment of the present invention, the mobile monitoring platform adjusts the moving speed of the movable camera according to the position information of the target object in the image data, including:

[0075] The mobile monitoring platform extracts the position information of the target object in each image data;

[0076] The mobile monitoring platform obtains the moving speed compensation amount of the movable camera according to the position information of the target object in the image data; wherein the moving speed compensation amount is obtained by the following formula:

[0077]

[0078]

[0079]

[0080] Among them, ξ v Indicates the movement speed compensation amount; S 01 and S 02 represent the first adjustment coefficient and the second adjustment coefficient respectively; n represents the total number of image data captured by the movable camera; S i represents the overlapping area between the target object in the i-th image data and the target object in the previous image data; S e represents the area occupied by the target object in the image data; m represents the number of image data in which the target object is in the front position compared with the target object in the previous image data; k represents the number of image data in which the target object is in the back position compared with the target object in the previous image data; S xi represents the overlapping area of ​​the target object in the i-th previous image data compared with the target object in the previous image data; S hi represents the overlapping area of ​​the target object in the i-th subsequent image data compared with the target object in the previous image data;

[0081] The mobile monitoring platform uses the mobile speed compensation to obtain the target speed; wherein the target speed is obtained by the following formula:

[0082] V=(1+ξ v )·V0

[0083] Wherein, V represents the target speed; V0 represents the tracking speed before adjustment;

[0084] The mobile monitoring platform adjusts the tracking speed of the movable camera according to the target speed.

[0085] The working principle of the above technical solution is as follows: after the mobile monitoring platform receives the image data of the target moving object sent back by the mobile camera, it uses image processing technology or deep learning model to extract the position information of the target object in each image data, such as the coordinates of the bounding box. Based on the extracted position information, the mobile monitoring platform calculates the movement state of the target object between consecutive image data, and then calculates the movement speed compensation. This compensation takes into account multiple factors, including the overlapping area (S) of the target object in the previous and next images.i ), the area of ​​the target object in the image (S e ), and the moving state of the target object relative to the preceding and following image data (S xi and S hi ). The first adjustment coefficient (S 01 ) and the second adjustment coefficient (S 02 ) is used to adjust the weights of different factors when calculating the compensation amount. The mobile monitoring platform uses the calculated mobile speed compensation amount (ξ v ) adjusts the original tracking speed (V0) to obtain the target speed (V). Through this step, the mobile monitoring platform can dynamically adjust the tracking speed of the camera based on the actual movement state of the target object. Finally, the mobile monitoring platform converts the target speed (V) into a control instruction for the movable camera, adjusts the tracking speed of the camera, and ensures that the camera can accurately and stably track the target object.

[0086] The effect of the above technical solution is: by calculating the movement speed compensation amount in real time and adjusting the tracking speed of the camera, the mobile monitoring platform can dynamically adapt to the movement speed and direction changes of the target object to achieve a more stable and accurate tracking effect. The algorithm comprehensively considers multiple factors, including the overlapping area of ​​the target object, the area ratio, and the movement state relative to the previous and next images. It can cope with target tracking tasks in complex environments and has strong robustness. The mobile monitoring platform can receive and process image data in real time, extract the position information of the target object, and calculate the movement speed compensation amount and target speed based on this to ensure the real-time and accuracy of tracking. The entire tracking process is managed intelligently and can be completed without human intervention. This reduces labor costs, improves work efficiency, and reduces the possibility of human errors. The first adjustment coefficient (S 01 ) and the second adjustment coefficient (S 02 ) can be adjusted according to actual needs to adapt to target tracking tasks in different application scenarios. This flexibility makes the algorithm have a wider application prospect. The average ratio of the overlapping area of ​​the front image to the area occupied by the target object can be calculated. Relatively speaking, when S xi The larger the S 01 The larger the The average ratio of the overlap area of ​​the rear image to the area occupied by the target object can be calculated. Relatively speaking, when S hi The larger the S 02 The larger the The average value of the ratio of the overlapping area between the target object in the i-th image data and the target object in the previous image data to the area occupied by the target object can be calculated. Relatively speaking, when Si The larger the v The larger the S 01 With S 02 The larger the difference, the v The bigger it is, v The larger it is, the larger V is likely to be.

[0087] In one embodiment of the present invention, the object movement monitoring system based on AI visual recognition includes:

[0088] Image data acquisition module, used for the mobile monitoring platform to control the movable camera to track and collect image data of the target moving object in real time;

[0089] Image data recognition module, used for the mobile monitoring platform to identify image data using a deep learning model and obtain the location information of the target object in the image data;

[0090] The moving speed adjustment module is used for the mobile monitoring platform to adjust the moving speed of the movable camera according to the position information of the target object in the image data.

[0091] The working principle of the above technical solution is as follows: the mobile monitoring platform first tracks and collects image data of the target moving object in real time through a movable camera. This involves the motion control of the camera, such as electric zoom, pan / tilt rotation, etc., to ensure that the camera can always aim at and capture the target object. The collected image data is then fed into the deep learning model for recognition. After being trained with a large amount of labeled data, the deep learning model can learn the feature representation of different objects in the image, thereby achieving high-accuracy object recognition. In the target detection task, the deep learning model can detect the position (such as bounding box coordinates) and category of the target object in the image. Based on the target object position information identified by the deep learning model, the mobile monitoring platform calculates the moving speed and direction of the target object relative to the camera. Then, the platform adjusts the moving speed of the movable camera based on this information to ensure that the camera can continuously and stably track the target object and always place it in the center or a specific position of the image.

[0092] The effect of the above technical solution is: through the real-time tracking and acquisition of the movable camera and the fast and accurate recognition of the deep learning model, the mobile monitoring platform can realize the real-time tracking and positioning of the target object, greatly improving the real-time and accuracy of the monitoring. Due to the use of deep learning technology, the mobile monitoring platform can cope with the target object recognition tasks in various complex scenarios. At the same time, by adjusting the moving speed and direction of the camera, the platform can flexibly adapt to the movement changes of the target object to ensure a continuous and stable tracking effect. The entire workflow of the mobile monitoring platform is intelligent and automated, and the real-time tracking and positioning of the target object can be completed without human intervention. This not only reduces labor costs, but also improves work efficiency and safety. The platform can be expanded and customized according to actual needs, such as adding more hardware equipment such as cameras and sensors, or optimizing the algorithms and parameters of the deep learning model. This gives the platform a wider application prospect and higher flexibility.

[0093] In one embodiment of the present invention, the image data acquisition module includes:

[0094] An initial data acquisition module, used for the movable camera to acquire the initial image data of the target area in real time, and send the initial image data to the mobile control platform;

[0095] A movement judgment module, used for the mobile control platform to identify the initial image data after receiving the initial image data, and to judge whether there is a target moving object in the current initial image data;

[0096] A camera control module, configured to, when there is a target moving object in the initial image data, cause the mobile control platform to send a control instruction to the movable camera, control the movable camera to track and shoot the target moving object, and obtain image data of the target moving object;

[0097] The data transmission module is used for the movable camera to transmit the image data of the target moving object back to the mobile control platform.

[0098] The working principle of the above technical solution is as follows: the movable camera is deployed in the target area to be monitored, and the initial image data of the area is collected in real time. These image data may contain static background elements and possible dynamic target moving objects. The collected initial image data is then transmitted to the mobile control platform. The transmission can be achieved through wireless communication technology (such as Wi-Fi, 4G / 5G, etc.) to ensure the real-time and efficient data. After receiving the initial image data, the mobile control platform will use a pre-trained deep learning model or other image recognition algorithm to process the image. The purpose of this step is to determine whether there is a target moving object in the current image. If the target moving object is identified in the initial image data, the mobile control platform will further analyze the position, size and movement direction of the target object in the image. Based on the detection information of the target object, the mobile control platform will calculate the parameters of the movable camera that need to be adjusted and generate corresponding control instructions. These instructions are sent to the movable camera via wireless communication. After receiving the control instruction, the movable camera will adjust its own parameters according to the instruction requirements and track and shoot the target moving object. The captured image data will be transmitted back to the mobile control platform again. After the image data of the target moving object is collected by the mobile camera, it is transmitted back to the mobile control platform in real time through wireless communication technology. The mobile control platform can further analyze, store or display the data.

[0099] The effect of the above technical solution is: through the real-time acquisition of the movable camera and the rapid processing of the mobile control platform, the system can realize the real-time tracking and positioning of the target moving object, ensuring the real-time and accuracy of the data. The entire tracking and shooting process is intelligent and automated, and can be completed without human intervention. This reduces labor costs, improves work efficiency, and reduces the possibility of human errors. The system can be flexibly configured and expanded according to actual needs. For example, more movable cameras can be added to cover a wider area, or the recognition algorithm can be optimized to improve recognition accuracy and speed. By tracking and monitoring the target moving object in real time, the system can be used in security monitoring, traffic management and other fields to provide timely and effective data support to help improve safety and reliability. The image data of the target moving object that is continuously transmitted back provides a rich data source for subsequent data analysis, decision support, etc., which helps to improve the application value and effect of the entire system.

[0100] In one embodiment of the present invention, the image data recognition module includes:

[0101] A model input module, used for the mobile control platform to input the image data of the target moving object transmitted back by the movable camera into the deep learning model after receiving the image data;

[0102] The position acquisition module is used to identify the target object through the deep learning model and obtain the position information of the target object in the image data.

[0103] The working principle of the above technical solution is as follows: the mobile control platform first receives image data from the movable camera, which has been preliminarily screened and processed and contains image information of the target moving object. The mobile control platform loads a pre-trained deep learning model. Before inputting the image data into the deep learning model, some preprocessing steps may be required, such as resizing, normalization, denoising, etc., to meet the input requirements of the model and improve the recognition effect of the model. The deep learning model processes the input image data, extracts image features through structures such as convolutional layers, pooling layers, and fully connected layers, and applies the trained weight parameters for forward propagation calculations, and finally outputs the location information of the target object in the image. The mobile control platform receives the output result of the deep learning model, that is, the location information of the target object in the image, and performs subsequent processing or display.

[0104] The effect of the above technical solution is that the deep learning model can accurately identify the target object in the image and output accurate location information after training and optimization with a large amount of data. This is very important for application scenarios that require accurate tracking and positioning of target objects. The calculation speed of the deep learning model is usually very fast, and it can process a large amount of image data in a short time and output recognition results. This enables the mobile control platform to receive and process the image data of the movable camera in real time, meeting application scenarios with high real-time requirements. The deep learning model has strong robustness to factors such as image noise, illumination changes, and occlusion, and can maintain a stable recognition effect under complex environmental conditions. This has good adaptability to unpredictable factors in practical applications. The deep learning model has good scalability and can be customized and optimized according to the needs of practical applications. For example, the recognition accuracy and performance of the model can be improved by increasing training data, adjusting model parameters, optimizing algorithms, etc. Combined with the target object recognition function of the deep learning model, the mobile control platform can realize intelligent management of the target moving object. For example, the shooting angle and focal length of the camera can be automatically adjusted according to the recognition results, or the linkage operation of other related devices can be triggered to achieve a higher level of intelligent monitoring and management.

[0105] In one embodiment of the present invention, the moving speed adjustment module includes:

[0106] A moving speed compensation acquisition module, used for the mobile monitoring platform to extract the position information of the target object in each image data;

[0107] The mobile monitoring platform obtains the moving speed compensation amount of the movable camera according to the position information of the target object in the image data; wherein the moving speed compensation amount is obtained by the following formula:

[0108]

[0109]

[0110]

[0111] Among them, ξ v Indicates the movement speed compensation amount; S 01 and S 02 represent the first adjustment coefficient and the second adjustment coefficient respectively; n represents the total number of image data captured by the movable camera; S i represents the overlapping area between the target object in the i-th image data and the target object in the previous image data; S e represents the area occupied by the target object in the image data; m represents the number of image data in which the target object is in the front position compared with the target object in the previous image data; k represents the number of image data in which the target object is in the back position compared with the target object in the previous image data; S xi represents the overlapping area of ​​the target object in the i-th previous image data compared with the target object in the previous image data; S hi represents the overlapping area of ​​the target object in the i-th subsequent image data compared with the target object in the previous image data;

[0112] The target speed acquisition module is used for the mobile monitoring platform to obtain the target speed using the mobile speed compensation amount; wherein the target speed is obtained by the following formula:

[0113] V=(1+ξ v )·V0

[0114] Wherein, V represents the target speed; V0 represents the tracking speed before adjustment;

[0115] The adjustment module is used for the mobile monitoring platform to adjust the tracking speed of the movable camera according to the target speed.

[0116] The working principle of the above technical solution is as follows: after the mobile monitoring platform receives the image data of the target moving object sent back by the mobile camera, it uses image processing technology or deep learning model to extract the position information of the target object in each image data, such as the coordinates of the bounding box. Based on the extracted position information, the mobile monitoring platform calculates the movement state of the target object between consecutive image data, and then calculates the movement speed compensation. This compensation takes into account multiple factors, including the overlapping area (S) of the target object in the previous and next images. i ), the area of ​​the target object in the image (S e ), and the moving state of the target object relative to the preceding and following image data (S xi and S hi ). The first adjustment coefficient (S 01 ) and the second adjustment coefficient (S 02 ) is used to adjust the weights of different factors when calculating the compensation amount. The mobile monitoring platform uses the calculated mobile speed compensation amount (ξ v ) adjusts the original tracking speed (V0) to obtain the target speed (V). Through this step, the mobile monitoring platform can dynamically adjust the tracking speed of the camera based on the actual movement state of the target object. Finally, the mobile monitoring platform converts the target speed (V) into a control instruction for the movable camera, adjusts the tracking speed of the camera, and ensures that the camera can accurately and stably track the target object.

[0117] The effect of the above technical solution is: by calculating the movement speed compensation amount in real time and adjusting the tracking speed of the camera, the mobile monitoring platform can dynamically adapt to the movement speed and direction changes of the target object to achieve a more stable and accurate tracking effect. The algorithm comprehensively considers multiple factors, including the overlapping area of ​​the target object, the area ratio, and the movement state relative to the previous and next images. It can cope with target tracking tasks in complex environments and has strong robustness. The mobile monitoring platform can receive and process image data in real time, extract the position information of the target object, and calculate the movement speed compensation amount and target speed based on this to ensure the real-time and accuracy of tracking. The entire tracking process is managed intelligently and can be completed without human intervention. This reduces labor costs, improves work efficiency, and reduces the possibility of human errors. The first adjustment coefficient (S 01 ) and the second adjustment coefficient (S 02 ) can be adjusted according to actual needs to adapt to target tracking tasks in different application scenarios. This flexibility makes the algorithm have a wider application prospect.

[0118] In one embodiment of the present invention, the object movement monitoring device based on AI visual recognition comprises a mobile monitoring platform and a movable camera; wherein the mobile monitoring platform and the movable camera are connected by wireless communication for data signals. The mobile monitoring platform comprises a processor, wherein the processor is used to execute the object movement monitoring method based on AI visual recognition according to claims 1 to 4.

[0119] The working principle of the above technical solution is as follows: the movable camera collects image data of the target area in real time, and sends the collected initial image data to the mobile monitoring platform through wireless communication methods (such as Wi-Fi, 4G / 5G, etc.). The wireless communication between the mobile monitoring platform and the movable camera ensures the real-time transmission of data. This ensures that the mobile monitoring platform can process and analyze the image data from the movable camera in a timely manner. After receiving the image data, the processor of the mobile monitoring platform will execute the object movement monitoring method based on AI visual recognition. The method includes the steps of image recognition, target detection, calculation of moving speed compensation, and determination of target speed. The image data is processed by a deep learning model to identify the target object in the image. The position information of the target object in the image, such as the coordinates of the bounding box, is extracted. Based on the position change of the target object between continuous image data, the moving speed compensation is calculated to reflect the actual moving speed and direction of the target object. The original tracking speed is adjusted by the moving speed compensation to obtain the target speed. According to the calculated target speed, the mobile monitoring platform generates a control instruction and sends it to the movable camera through wireless communication. After receiving the control command, the mobile camera adjusts its own shooting parameters (such as focal length, angle, moving speed, etc.) according to the command to track and shoot the target object. The mobile camera transmits the image data of the target object captured by the mobile camera back to the mobile monitoring platform through wireless communication for subsequent analysis or storage.

[0120] The effect of the above technical solution is: through wireless communication technology and AI visual recognition method, the device can track and monitor the movement of the target object in real time, and accurately obtain the location information of the target object. The device realizes intelligent management, can automatically identify and track the target object, and dynamically adjust the camera shooting parameters according to the movement state of the target object. This reduces labor costs and improves work efficiency. The AI ​​visual recognition method has strong robustness and adaptability, and can cope with target tracking tasks in complex environments, such as lighting changes, occlusion, and changes in the shape of the target object. The equipment can be flexibly configured and adjusted according to actual needs, such as increasing or decreasing the number of movable cameras, adjusting the parameters of the AI ​​visual recognition method, etc. The image data of the target object that is continuously transmitted back provides a rich data source for subsequent data analysis, decision support, etc., which helps to improve the application value and effect of the entire system.

[0121] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for object movement monitoring based on AI visual recognition, characterized in that: The object movement monitoring method based on AI visual recognition includes: The mobile monitoring platform controls the movable camera to track and collect image data of the target moving object in real time; The mobile monitoring platform uses a deep learning model to identify image data and obtain the location information of the target object in the image data; The mobile monitoring platform adjusts the moving speed of the movable camera according to the position information of the target object in the image data; The mobile monitoring platform adjusts the moving speed of the movable camera according to the position information of the target object in the image data, including: The mobile monitoring platform extracts the position information of the target object in each image data; The mobile monitoring platform obtains the moving speed compensation amount of the movable camera according to the position information of the target object in the image data; wherein the moving speed compensation amount is obtained by the following formula: Among them, ξ v Indicates the movement speed compensation amount; S 01 and S 02 represent the first adjustment coefficient and the second adjustment coefficient respectively; n represents the total number of image data captured by the movable camera; S i represents the overlapping area between the target object in the i-th image data and the target object in the previous image data; S e represents the area occupied by the target object in the image data; m represents the number of image data in which the target object is in the front position compared with the target object in the previous image data; k represents the number of image data in which the target object is in the back position compared with the target object in the previous image data; S xi represents the overlapping area of ​​the target object in the i-th previous image data compared with the target object in the previous image data; S hi represents the overlapping area of ​​the target object in the i-th subsequent image data compared with the target object in the previous image data; The mobile monitoring platform uses the mobile speed compensation to obtain the target speed; wherein the target speed is obtained by the following formula: V=(1+ξ v )·V0 Wherein, V represents the target speed; V0 represents the tracking speed before adjustment; The mobile monitoring platform adjusts the tracking speed of the movable camera according to the target speed.

2. The object movement monitoring method based on AI visual recognition according to claim 1 is characterized in that: The mobile monitoring platform controls the movable camera to track and collect image data of the target moving object in real time, including: The movable camera collects initial image data of the target area in real time, and sends the initial image data to the mobile control platform; After receiving the initial image data, the mobile control platform identifies the initial image data and determines whether there is a target moving object in the current initial image data; When there is a target moving object in the initial image data, the mobile control platform sends a control instruction to the movable camera to control the movable camera to track and shoot the target moving object, and obtain image data of the target moving object; The movable camera transmits the image data of the target moving object back to the mobile control platform.

3. The object movement monitoring method based on AI visual recognition according to claim 2 is characterized in that: The mobile monitoring platform uses deep learning models to identify image data and obtain the location information of the target object in the image data, including: After receiving the image data of the target moving object transmitted back by the movable camera, the mobile control platform inputs the image data into the deep learning model; The target object is identified through the deep learning model to obtain the position information of the target object in the image data.

4. An object movement monitoring system based on AI visual recognition, characterized in that: The object movement monitoring system based on AI visual recognition includes: Image data acquisition module, used for the mobile monitoring platform to control the movable camera to track and collect image data of the target moving object in real time; Image data recognition module, used for the mobile monitoring platform to identify image data using a deep learning model and obtain the location information of the target object in the image data; A moving speed adjustment module, used for the mobile monitoring platform to adjust the moving speed of the movable camera according to the position information of the target object in the image data; The moving speed adjustment module includes: A moving speed compensation acquisition module, used for the mobile monitoring platform to extract the position information of the target object in each image data; The mobile monitoring platform obtains the moving speed compensation amount of the movable camera according to the position information of the target object in the image data; wherein the moving speed compensation amount is obtained by the following formula: Among them, ξ v Indicates the movement speed compensation amount; S 01 and S 02 represent the first adjustment coefficient and the second adjustment coefficient respectively; n represents the total number of image data captured by the movable camera; S i represents the overlapping area between the target object in the i-th image data and the target object in the previous image data; S e represents the area occupied by the target object in the image data; m represents the number of image data in which the target object is in the front position compared with the target object in the previous image data; k represents the number of image data in which the target object is in the back position compared with the target object in the previous image data; S xi represents the overlapping area of ​​the target object in the i-th previous image data compared with the target object in the previous image data; S hi represents the overlapping area of ​​the target object in the i-th subsequent image data compared with the target object in the previous image data; The target speed acquisition module is used for the mobile monitoring platform to obtain the target speed using the mobile speed compensation amount; wherein the target speed is obtained by the following formula: V=(1+ξ v )·V0 Wherein, V represents the target speed; V0 represents the tracking speed before adjustment; The adjustment module is used for the mobile monitoring platform to adjust the tracking speed of the movable camera according to the target speed.

5. The object movement monitoring system based on AI visual recognition according to claim 4 is characterized in that: The image data acquisition module includes: An initial data acquisition module, used for the movable camera to acquire the initial image data of the target area in real time, and send the initial image data to the mobile control platform; A movement judgment module, used for the mobile control platform to identify the initial image data after receiving the initial image data, and to judge whether there is a target moving object in the current initial image data; A camera control module, configured to, when there is a target moving object in the initial image data, cause the mobile control platform to send a control instruction to the movable camera, control the movable camera to track and shoot the target moving object, and obtain image data of the target moving object; The data transmission module is used for the movable camera to transmit the image data of the target moving object back to the mobile control platform.

6. The object movement monitoring system based on AI visual recognition according to claim 5 is characterized in that: The image data recognition module includes: A model input module, used for the mobile control platform to input the image data of the target moving object transmitted back by the movable camera into the deep learning model after receiving the image data; The position acquisition module is used to identify the target object through the deep learning model and obtain the position information of the target object in the image data.

7. An object movement monitoring device based on AI visual recognition is used to execute the object movement monitoring method according to any one of claims 1 to 3, characterized in that: The object movement monitoring device based on AI visual recognition includes a mobile monitoring platform and a movable camera; wherein the mobile monitoring platform and the movable camera are connected by data signals via wireless communication.

8. The object movement monitoring device based on AI visual recognition according to claim 7 is characterized in that: The mobile monitoring platform includes a processor, wherein the processor is used to execute the object movement monitoring method based on AI visual recognition as described in claims 1 to 3.

Images