Moving target remote monitoring system based on FPGA

Abstract

The invention relates to an FPGA-based moving target remote monitoring system, which relates to the technical field of motion monitoring, and comprises an image acquisition and processing module, a motion detection module, a classification module, a target tracking module, a self-adaptive adjustment module, a perception cooperation module and an emotion and intention recognition module. According to the moving target remote monitoring system based on the FPGA, various sensors are integrated, environmental changes are sensed in real time, and parameters and algorithms of the system are automatically adjusted according to environmental conditions, so that the system can keep efficient monitoring performance in various complex environments such as low illumination, rainy days and foggy days, the accuracy and stability of target detection and tracking are ensured, and the system is suitable for large-scale popularization and application. Compared with a single sensing mode, multi-sensor fusion significantly improves the robustness and flexibility of the system and adapts to more diversified application scenes, the sensing cooperation module realizes fusion of multiple sensing modes such as vision, audio and vibration, and multi-dimensional comprehensive sensing can be performed on a moving target.

Description

Technical Field

[0001] The invention relates to the technical field of motion monitoring, in particular to a moving target remote monitoring system based on FPGA. Background Art

[0002] The moving target remote monitoring system is an automated system that monitors moving targets (such as people, vehicles, animals, etc.) in a specific area in real time through technical means such as sensors, cameras, communication networks and intelligent algorithms. It can automatically detect, track, identify and analyze the behavior of moving targets, and transmit relevant information to the remote monitoring center or user terminal for further processing and decision-making. Such systems are widely used in many fields, including but not limited to: security monitoring, traffic management, smart cities, industrial automation, military and defense, and wildlife monitoring.

[0003] A Chinese patent with publication (announcement) number: CN109375177B discloses a moving target detection method for an airport scene surveillance radar system, comprising: obtaining target echo information of the i-th frame after processing by the radar front end; performing false point trace filtering and channel selection processing on the target echo information of the i-th frame; constructing a two-dimensional image of the i-th frame according to the target echo information of the i-th frame after channel selection and performing target edge extraction; performing inter-frame difference processing according to the two-dimensional image of the i-th frame and the two-dimensional image of the i-1-th frame; matching processing is performed based on the target echo information of the i-th frame belonging to the moving target obtained after the inter-frame difference processing and the target echo information of the i-th frame belonging to the same target after the target edge extraction, to obtain The target echo information of the i-th frame belongs to the same moving target; the above application mainly relies on radar echo information for target detection and tracking, lacks real-time perception and adjustment of environmental changes (such as light, weather, noise, etc.). Although the detection effect can be improved by adjusting the image enhancement algorithm under low light conditions, the performance of the system may be affected in complex and changeable environments (such as rainy days, foggy days, strong winds, etc.). Secondly, it mainly relies on radar echo information for target detection and tracking. Although it can effectively detect moving targets, it lacks the fusion of other perception methods (such as sound, vibration, etc.), which makes the system in some scenarios (such as low signal-to-noise ratio environments or complex backgrounds) may not be able to accurately distinguish between targets and background noise. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a remote monitoring system for moving targets based on FPGA, which has the advantages of improving the accuracy of target detection and classification, and solves the problem of limitations of a single perception method in complex environments.

[0005] To achieve the above-mentioned object, the present invention provides the following technical solutions: a remote monitoring system for moving targets based on FPGA, comprising an image acquisition and processing module, a motion detection module, a classification module and a target tracking module, the monitoring system also comprising an adaptive adjustment module, a perception and coordination module and an emotion and intention recognition module;

[0006] The image acquisition and processing module is used to obtain video streams and image data from the camera and image sensor and pre-process the acquired images. The motion detection module is used to detect the motion area in the image and distinguish between background and foreground objects. The classification module identifies and classifies the detected moving targets. The target tracking module continuously tracks the identified targets and records their motion trajectories.

[0007] in,

[0008] Adaptive environmental perception and adjustment module: It senses environmental changes in real time and automatically adjusts system parameters and algorithms according to environmental conditions to optimize detection and tracking effects;

[0009] Perception collaboration module: By integrating multiple perception methods, it realizes multi-dimensional perception of moving targets and makes collaborative decisions based on multimodal information.

[0010] Emotion and intention recognition module: Analyzes the target's behavior, facial expressions, body language, etc. to infer the target's emotional state and intentions.

[0011] Furthermore, the adaptive environmental perception adjustment module includes but is not limited to a camera, a temperature sensor, a humidity sensor, a light intensity sensor, a wind speed sensor and a noise sensor. The adaptive environmental perception adjustment module extracts features from the collected multimodal data, and the features include but are not limited to visual features, temperature features, humidity features, light intensity features, wind speed features and noise features.

[0012] Furthermore, under low light conditions, the adaptive environmental perception adjustment module enables infrared fill light and adjusts image enhancement; in rainy or foggy days, the system adjusts the edge detection algorithm to improve clarity.

[0013] Furthermore, the system adjusts the edge detection algorithm to include expanding the grayscale range of the image and enhancing the contrast of the image. The formula is as follows:

[0014]

[0015] Among them, H out (k) is the cumulative histogram of the output image, ni is the number of pixels with gray value i in the input image, N is the total number of all pixels in the image, and k is the gray level;

[0016] The system adjusts the edge detection algorithm and also adjusts the brightness and contrast of the image to enhance the details of the low-light area. The formula is as follows:

[0017] I out (x, y) = I in (x, y) γ ,

[0018] Among them, I in (x, y) is the pixel value of the input image, I out (x, y) is the pixel value of the output image and γ is the gamma value.

[0019] Furthermore, the system adjusts the edge detection algorithm to also simulate the adaptability of the human eye to changes in illumination to enhance the local contrast of the image, and the formula is as follows:

[0020] R(x,y)=log(I(x,y))-log(L(x,y)),

[0021] Among them, I(x, y) is the pixel value of the input image, L(x, y) is the illumination component, and R(x, y) is the reflection component.

[0022] Furthermore, the perception collaboration module integrates a microphone array, a vibration sensor and a perception device to obtain non-visual information such as sound and vibration. When the image acquisition and processing module detects a target, the sound sensor further confirms whether the target makes a specific sound.

[0023] Furthermore, the emotion and intention recognition module uses a lightweight deep learning model to analyze the target's facial expressions, postures, and movements to infer their emotional state. When the system detects that someone stares at an area for a long time or repeatedly approaches an object, it infers that they may have the intention of stealing or destroying it, and issues an alarm in advance. The FPGA realizes real-time emotion recognition through a hardware accelerator.

[0024] Furthermore, the deep learning model includes a facial expression recognition model, a behavior analysis and intention prediction model, an audio emotion recognition model and a multimodal fusion model. The facial expression recognition model is used to extract facial features from images or video frames and recognize facial expressions; the behavior analysis and intention prediction model extracts static features of the target and captures the dynamic behavior of the target to predict the target's intention; the audio emotion recognition model extracts emotional features from audio signals and predicts the speaker's emotional state; the multimodal fusion model directly fuses data of different modalities in the feature extraction stage to form a comprehensive multimodal feature representation.

[0025] Furthermore, the image acquisition and processing module is connected to the camera using interfaces such as USB, HDMI, MIP I-CSI, etc., supports image input with different resolutions and frame rates, quickly transfers image data to the memory or processing unit through DMA, i.e. direct memory access, removes noise through a convolution filter, and grays the image and performs histogram equalization.

[0026] Furthermore, the motion detection algorithm of the motion detection module includes but is not limited to optical flow method, background modeling and frame difference method, and the tracking algorithm of the target tracking module includes but is not limited to Kalman filter, particle filter and SORT algorithm.

[0027] Compared with the prior art, the technical solution of this application has the following beneficial effects:

[0028] 1. This FPGA-based moving target remote monitoring system integrates multiple sensors to perceive environmental changes in real time and automatically adjusts system parameters and algorithms according to environmental conditions, enabling the system to maintain efficient monitoring performance in various complex environments such as low light, rainy days, foggy days, etc., ensuring the accuracy and stability of target detection and tracking. Compared with a single perception method, multi-sensor fusion significantly improves the robustness and flexibility of the system and adapts to more diverse application scenarios.

[0029] 2. The FPGA-based remote monitoring system for moving targets, the perception and coordination module realizes the integration of multiple perception methods such as vision, audio, and vibration. It can comprehensively perceive moving targets in multiple dimensions and make collaborative decisions based on multimodal information. It not only improves the accuracy of target classification and behavior analysis, but also effectively reduces false detections and missed detections. In addition, the emotion and intention recognition module analyzes the target's behavior, facial expressions, and body language through a deep learning model, infers its emotional state and potential intentions, and provides intelligent early warning functions to help predict and prevent potential security threats in advance, greatly enhancing the intelligence level and application scope of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a schematic diagram of the system framework structure of the present invention. DETAILED DESCRIPTION

[0031] 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.

[0032] See also Figure 1,The FPGA-based moving target remote monitoring system in this embodiment includes an image acquisition and processing module, a motion detection module, a classification module and a target tracking module, and the monitoring system also includes an adaptive adjustment module, a perception and coordination module and an emotion and intention recognition module;

[0033] The image acquisition and processing module is used to obtain video streams and image data from cameras and image sensors and pre-process the acquired images. The motion detection module is used to detect the moving area in the image and distinguish between background and foreground objects. The classification module identifies and classifies the detected moving targets. The target tracking module continuously tracks the identified targets and records their motion trajectories.

[0034] in,

[0035] Adaptive environmental perception and adjustment module: It senses environmental changes in real time and automatically adjusts system parameters and algorithms according to environmental conditions to optimize detection and tracking effects;

[0036] Perception collaboration module: By integrating multiple perception methods, it realizes multi-dimensional perception of moving targets and makes collaborative decisions based on multimodal information.

[0037] Emotion and intention recognition module: Analyzes the target's behavior, facial expressions, body language, etc. to infer the target's emotional state and intentions.

[0038] In this embodiment, the adaptive environmental perception adjustment module includes but is not limited to a camera, a temperature sensor, a humidity sensor, a light intensity sensor, a wind speed sensor and a noise sensor. The adaptive environmental perception adjustment module extracts features from the collected multimodal data, and the features include but are not limited to visual features, temperature features, humidity features, light intensity features, wind speed features and noise features.

[0039] It should be noted that the adaptive environmental perception and adjustment module integrates a variety of sensors, including but not limited to cameras, temperature sensors, humidity sensors, light intensity sensors, wind speed sensors and noise sensors. The module extracts features from the collected multimodal data. The extracted features include but are not limited to visual features, temperature features, humidity features, light intensity features, wind speed features and noise features. By perceiving environmental changes in real time and analyzing these multimodal features, the module can automatically adjust the system parameters and algorithms to optimize target detection and tracking effects. For example, in low light conditions, the module enables infrared fill light and adjusts the image enhancement algorithm to ensure that the target is clearly visible; in rainy or foggy days, the system adjusts the edge detection algorithm to improve image clarity. This comprehensive environmental perception and adaptive adjustment capability significantly improves the robustness and flexibility of the system, ensuring that it can maintain efficient monitoring performance in various complex environments and enhancing its wide adaptability to application scenarios.

[0040] In this embodiment, under low light conditions, the adaptive environmental perception adjustment module enables infrared fill light and adjusts image enhancement; in rainy or foggy days, the system adjusts the edge detection algorithm to improve clarity.

[0041] It should be noted that the adaptive environmental perception adjustment module can automatically optimize the performance of the system under different environmental conditions to ensure efficient target detection and tracking. Under low light conditions, the module enables infrared fill light and adjusts the image enhancement algorithm to expand the grayscale range of the image and enhance the contrast to ensure that the target is still clearly visible in a dark environment. On rainy or foggy days, the system will automatically adjust the edge detection algorithm, optimize the image processing parameters, improve the image clarity and detail retention, and effectively deal with the impact of bad weather on visual perception. This intelligent adaptive adjustment not only improves the robustness and flexibility of the system, but also ensures efficient monitoring performance in various complex environments, enhancing the system's ability to adapt to application scenarios.

[0042] In this embodiment, the system adjusts the edge detection algorithm to include expanding the grayscale range of the image and enhancing the contrast of the image. The formula is as follows:

[0043]

[0044] Among them, H out (k) is the cumulative histogram of the output image, ni is the number of pixels with gray value i in the input image, N is the total number of all pixels in the image, and k is the gray level;

[0045] The system adjusts the edge detection algorithm and also adjusts the brightness and contrast of the image to enhance the details in low-light areas. The formula is as follows:

[0046] I out (x, y) = I in (x, y) γ ,

[0047] Among them, I in (x, y) is the pixel value of the input image, I out (x, y) is the pixel value of the output image and γ is the gamma value.

[0048] In this embodiment, the system adjusts the edge detection algorithm to also simulate the adaptability of the human eye to changes in illumination to enhance the local contrast of the image. The formula is as follows:

[0049] R(x,y)=log(I(x,y))-log(L(x,y)),

[0050] Among them, I(x, y) is the pixel value of the input image, L(x, y) is the illumination component, and R(x, y) is the reflection component.

[0051] In this embodiment, the perception collaboration module integrates a microphone array, a vibration sensor and a perception device to obtain non-visual information such as sound and vibration. When the image acquisition and processing module detects a target, the sound sensor in this embodiment confirms whether the target makes a specific sound.

[0052] It should be noted that the perception collaboration module acquires non-visual information such as sound and vibration by integrating microphone arrays, vibration sensors and other perception devices, thereby achieving multi-dimensional and comprehensive perception of the target. When the image acquisition and processing module detects a target, the sound sensor can confirm in real time whether the target makes specific sounds (such as car horns, footsteps, etc.), and further verify the identity and behavior of the target. This fusion of multimodal information not only improves the accuracy and robustness of target detection, but also reduces false detections and missed detections, ensuring the reliability and intelligence level of the system. Through collaborative work, the perception collaboration module can provide more accurate target recognition and behavior analysis in complex environments, enhancing the overall performance of the system and its adaptability to application scenarios.

[0053] In this embodiment, the emotion and intention recognition module uses a lightweight deep learning model to analyze the target's facial expressions, postures, and movements to infer its emotional state. When the system detects that someone stares at an area for a long time or repeatedly approaches an object, it infers that they may have the intention of stealing or destroying it, and issues an alarm in advance. The FPGA realizes real-time emotion recognition through a hardware accelerator.

[0054] It should be noted that the emotion and intention recognition module uses a lightweight deep learning model, which can efficiently analyze the target's facial expressions, postures and movements, and accurately infer its emotional state. When the system detects that someone is staring at an area for a long time or repeatedly approaches an object, the module can intelligently judge that they may have the intention of stealing or destroying, and issue an alarm in advance to provide intelligent security warnings. In addition, with the support of FPGA hardware accelerators, the module realizes real-time emotion recognition, ensuring the system's rapid response and high efficiency. It is suitable for various application scenarios that require instant decision-making and reaction. It not only improves the intelligence level of the system, but also enhances the security prevention capabilities and effectively prevents potential security threats.

[0055] In this embodiment, the deep learning model includes a facial expression recognition model, a behavior analysis and intention prediction model, an audio emotion recognition model and a multimodal fusion model. The facial expression recognition model is used to extract facial features from images or video frames and recognize facial expressions; the behavior analysis and intention prediction model extracts the static features of the target and captures the dynamic behavior of the target to predict the target's intention; the audio emotion recognition model extracts emotional features from audio signals and predicts the speaker's emotional state; the multimodal fusion model directly fuses data from different modalities in the feature extraction stage to form a comprehensive multimodal feature representation.

[0056] It should be noted that the deep learning model includes facial expression recognition model, behavior analysis and intention prediction model, audio emotion recognition model and multimodal fusion model, which can comprehensively perceive and analyze the target from multiple dimensions. The facial expression recognition model extracts facial features from images or video frames and accurately identifies expressions. The behavior analysis and intention prediction model accurately predicts the target's intention by extracting the static features of the target and capturing its dynamic behavior. The audio emotion recognition model extracts emotional features from audio signals and predicts the speaker's emotional state. In addition, the multimodal fusion model directly fuses data from multiple modalities such as vision and audio in the feature extraction stage to form a comprehensive multimodal feature representation, which significantly improves the system's perception ability and decision-making accuracy. This multi-level and multi-dimensional intelligent analysis capability enables the system to provide more comprehensive and accurate target monitoring and early warning functions in complex scenarios.

[0057] In this embodiment, the image acquisition and processing module uses USB, HDMI, MIP I-CSI and other interfaces to connect to the camera, supports image input of different resolutions and frame rates, and quickly transfers image data to the memory or processing unit through DMA, i.e. direct memory access. At the same time, it removes noise through a convolution filter, and grayscales and histograms the image.

[0058] It should be noted that the image acquisition and processing module is connected to the camera through a variety of interfaces such as USB, HDMI, M IPI-CSI, etc., and supports image input with different resolutions and frame rates, ensuring the flexibility and compatibility of the system. The module uses DMA (direct memory access) technology to achieve fast transmission of image data, significantly improving data processing efficiency. At the same time, the built-in convolution filter effectively removes noise to ensure image quality, and further optimizes the image through grayscale and histogram equalization, enhancing the accuracy and reliability of subsequent processing, which not only improves the performance of the system, but also ensures stable performance in various complex environments.

[0059] In this embodiment, the motion detection algorithm of the motion detection module includes but is not limited to optical flow method, background modeling and frame difference method, and the tracking algorithm of the target tracking module includes but is not limited to Kalman filter, particle filter and SORT algorithm.

[0060] It should be noted that by integrating a variety of advanced algorithms and technologies, the accuracy and robustness of target detection and tracking have been significantly improved. The motion detection module adopts a variety of motion detection algorithms including optical flow method, background modeling and frame difference method, which can flexibly respond to the needs of motion target detection in different scenarios and ensure high-precision motion area identification. The target tracking module uses advanced tracking algorithms such as Kalman filtering, particle filtering and SORT algorithm to achieve real-time and stable target tracking, and effectively deal with problems such as target occlusion and disappearance, ensuring the continuity and reliability of the target trajectory. This combination not only improves the intelligence level of the system, but also enhances its adaptability in complex environments and is suitable for a variety of application scenarios.

[0061] The advantages of the above embodiments are: it has the advantage of multi-sensor fusion, can integrate multiple perception methods such as vision, audio, vibration, temperature, humidity, light intensity, etc., to achieve comprehensive and multi-dimensional perception of the environment and targets, and significantly improve the accuracy and robustness of target detection and classification. Through the adaptive adjustment module, the system can perceive environmental changes in real time and automatically optimize parameters to ensure efficient monitoring performance in complex environments such as low light, rainy days, and foggy days. In addition, the system's perception collaboration module realizes the fusion and collaborative decision-making of multimodal information, reduces false detection and missed detection, and enhances the continuity and stability of target tracking. The emotion and intention recognition module analyzes the target's behavior, facial expressions and body language through a deep learning model, infers its emotional state and potential intentions, and provides intelligent early warning functions to prevent potential security threats in advance. Overall, it not only improves the intelligence level of monitoring, but also greatly enhances the capabilities of security warning and emergency response, and is suitable for efficient monitoring needs in a variety of complex scenarios.

[0062] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0063] 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. The FPGA-based moving target remote monitoring system is characterized by: It includes an image acquisition and processing module, a motion detection module, a classification module and a target tracking module, and the monitoring system also includes an adaptive adjustment module, a perception and coordination module and an emotion and intention recognition module; The image acquisition and processing module is used to obtain video streams and image data from the camera and image sensor and pre-process the acquired images. The motion detection module is used to detect the motion area in the image and distinguish between background and foreground objects. The classification module identifies and classifies the detected moving targets. The target tracking module continuously tracks the identified targets and records their motion trajectories. in, Adaptive environmental perception and adjustment module: It senses environmental changes in real time and automatically adjusts system parameters and algorithms according to environmental conditions to optimize detection and tracking effects; Perception and coordination module: By integrating multiple perception methods, it realizes multi-dimensional perception of moving targets and makes collaborative decisions based on multi-modal information; Emotion and intention recognition module: Analyzes the target's behavior, facial expressions, body language, etc. to infer the target's emotional state and intentions.

2. The FPGA-based moving target remote monitoring system according to claim 1 is characterized in that: The adaptive environment perception adjustment module includes but is not limited to a camera, a temperature sensor, a humidity sensor, a light intensity sensor, a wind speed sensor and a noise sensor. The adaptive environment perception adjustment module extracts features from the collected multimodal data, and the features include but are not limited to visual features, temperature features, humidity features, light intensity features, wind speed features and noise features.

3. The FPGA-based moving target remote monitoring system according to claim 1 is characterized in that: In low-light conditions, the adaptive environmental perception adjustment module enables infrared fill light and adjusts image enhancement; in rainy or foggy days, the system adjusts the edge detection algorithm to improve clarity.

4. The FPGA-based moving target remote monitoring system according to claim 3 is characterized in that: The system adjusts the edge detection algorithm to include expanding the grayscale range of the image and enhancing the contrast of the image. The formula is as follows: Among them, H out (k) is the cumulative histogram of the output image, ni is the number of pixels with gray value i in the input image, N is the total number of all pixels in the image, and k is the gray level; The system adjusts the edge detection algorithm and also adjusts the brightness and contrast of the image to enhance the details of the low-light area. The formula is as follows: I out (x,y)=I in (x,y) γ , Among them, I in (x, y) is the pixel value of the input image, I out (x, y) is the pixel value of the output image and γ is the gamma value.

5. The FPGA-based moving target remote monitoring system according to claim 3 is characterized in that: The system adjusts the edge detection algorithm and also includes simulating the adaptability of the human eye to changes in illumination to enhance the local contrast of the image. The formula is as follows: R(x,y)=log(I(x,y))-log(L(x,y)), Among them, I(x, y) is the pixel value of the input image, L(x, y) is the illumination component, and R(x, y) is the reflection component.

6. The FPGA-based moving target remote monitoring system according to claim 1, characterized in that: The perception collaboration module integrates a microphone array, a vibration sensor and a perception device to obtain non-visual information such as sound and vibration. When the image acquisition and processing module detects a target, the sound sensor further confirms whether the target makes a specific sound.

7. The FPGA-based moving target remote monitoring system according to claim 1 is characterized in that: The emotion and intention recognition module uses a lightweight deep learning model to analyze the target's facial expressions, postures, and movements to infer its emotional state. When the system detects that someone stares at an area for a long time or repeatedly approaches an object, it infers that they may have the intention of stealing or destroying it, and issues an alarm in advance. The FPGA realizes real-time emotion recognition through a hardware accelerator.

8. The FPGA-based moving target remote monitoring system according to claim 7 is characterized in that: The deep learning model includes a facial expression recognition model, a behavior analysis and intention prediction model, an audio emotion recognition model, and a multimodal fusion model. The facial expression recognition model is used to extract facial features from images or video frames and recognize facial expressions; the behavior analysis and intention prediction model extracts static features of the target and captures the dynamic behavior of the target to predict the target's intention; The audio emotion recognition model extracts emotional features from audio signals and predicts the speaker's emotional state; the multimodal fusion model directly fuses data of different modalities in the feature extraction stage to form a comprehensive multimodal feature representation.

9. The FPGA-based moving target remote monitoring system according to claim 1, characterized in that: The image acquisition and processing module uses USB, HDMI, MIPI-CSI and other interfaces to connect to the camera, supports image input with different resolutions and frame rates, and quickly transfers image data to the memory or processing unit through DMA, i.e. direct memory access. At the same time, it removes noise through convolution filters, and grayscales and histograms the images.

10. The FPGA-based moving target remote monitoring system according to claim 1, characterized in that: The motion detection algorithm of the motion detection module includes but is not limited to optical flow method, background modeling and frame difference method, and the tracking algorithm of the target tracking module includes but is not limited to Kalman filter, particle filter and SORT algorithm.

Images