Swimming pool drowning video monitoring system based on laser radar

Abstract

The invention relates to the field of swimming pool drowning video monitoring, in particular to a swimming pool drowning video monitoring system based on a laser radar, which comprises a data acquisition module, a laser radar module, a video camera module, an alarm output module and a database, and is characterized in that the data acquisition module builds a swimmer platform and acquires data; the laser radar module acquires target point data by taking a pool wall as a reference, constructs a three-dimensional image, calculates a head distance difference value, tangent slopes of four limbs and an abdomen and an attitude anomaly index, and judges and sends a first early warning signal; the video camera module analyzes human body posture and trajectory changes after receiving the signal, predicts a drowning probability by using a convolutional neural network drowning model, and generates a second early warning signal; after the alarm output module receives the second early warning signal, the drowner is positioned, a rescue route is planned, and a large-screen display and a lifeguard bracelet are notified; the swimming pool drowning monitoring system integrates multiple modules to realize accurate monitoring and quick response of the swimming pool drowning condition, and effectively guarantees the safety of the swimming pool.

Description

Technical Field

[0001] The present invention relates to the field of swimming pool drowning video monitoring, and specifically to a swimming pool drowning video monitoring system based on lidar. Background Technique

[0002] In the development process of swimming pool drowning monitoring technology, traditional monitoring means have gradually revealed serious limitations; pure video monitoring technology has achieved automated monitoring to a certain extent, but faces many insurmountable problems; for example, the refraction, reflection, and fluctuation phenomena of water will interfere with the quality of video images, resulting in blurred details of swimmers' movements, increasing the difficulty of accurately identifying drowning behaviors, and greatly reducing the accuracy of pure video monitoring systems;

[0003] The emergence of lidar technology has brought new opportunities for swimming pool drowning monitoring. In a swimming pool environment, lidar can continuously scan target objects, obtain a large amount of target point data including distance and reflection intensity, and then construct a three-dimensional image of the swimming pool, providing depth information and spatial position data for precise monitoring, effectively making up for the deficiency of video monitoring in depth perception; the RGB camera plays an indispensable role in collecting visual information in the monitoring system. It can capture high-definition images in the swimming pool, record the movement trajectories of swimmers, and provide intuitive data support for the system from a visual perspective;

[0004] However, there are still some problems with single lidar or video camera technology in practical applications; although lidar has advantages in distance and speed measurement, its ability to recognize some complex human body posture details and subtle movement changes is relatively weak. When the video camera group faces water interference and complex backgrounds, the accuracy of image analysis needs to be improved;

[0005] Deeply integrating lidar and video camera groups to construct a brand-new swimming pool drowning video monitoring system based on lidar has extremely important practical significance and application value. Summary of the Invention

[0006] To solve the technical problems proposed in the above background technique, the present invention provides a swimming pool drowning video monitoring system based on lidar.

[0007] The object of the present invention can be achieved by the following technical solutions:

[0008] The present invention is a swimming pool drowning video monitoring system based on lidar, including a data acquisition module, a lidar module, a video camera module, an alarm output module, and a database.

[0009] The data acquisition module is used to establish a swimmer platform, install lidar sensors and RGB camera groups in the swimming pool, and collect the basic information data of the people in the swimming pool and the motion data in the pool in real time, and send them to the lidar module and the video camera module respectively. The specific process is as follows:

[0010] A platform establishment unit and a collection unit are set in the data acquisition module;

[0011] The platform establishment unit is used to build a data platform for swimmers. Swimmers upload their personal information through the platform. Specifically: first, a swimmer login interface is established through the display screen. The swimmer login interface includes weight number input, number of swimming certificates input, number of monthly swimming times input, single swimming distance input, and single swimming time input. When the swimmer completes the input of the login interface, a user interface corresponding to the swimmer is established and a new identity number of the swimmer is generated and sent to the database for storage;

[0012] Analyze the basic data information uploaded by each swimmer in turn. Mark the selected data of weight, number of swimming certificates, number of monthly swimming times, single swimming distance, and single swimming time as stw1, stw2, stw3, stw4, and stw5 respectively, and substitute them into the formula for calculation after normalization processing Obtain the comprehensive swimming ability evaluation value of each swimmer. Among them, Q1, Q2, Q3, Q4, and Q5 are the weight coefficients corresponding to weight, number of swimming certificates, number of monthly swimming times, single swimming distance, and single swimming time. i represents the number of each swimmer, and n represents the number of each swimmer. Synchronously send the comprehensive swimming ability evaluation value corresponding to each swimmer to the database for storage;

[0013] The collection unit divides the deep pool and the shallow pool into several swimming areas in equal proportion respectively, installs RGB cameras according to the number of swimming areas, and one RGB camera corresponds to collecting the data information of each swimmer in one swimming area, and sends the obtained data information to the video camera module; install a lidar sensor directly above the center position of the swimming pool to collect the target point data of each swimmer in real time and send it to the lidar module.

[0014] The lidar module constructs a three-dimensional image of the swimming pool by emitting near-infrared light signals and measuring the target point data, identifies the parts of the swimmers in the swimming pool, calculates the difference value, slope, and attitude anomaly index, and determines the first warning signal of the swimmer. The specific process is as follows:

[0015] Set the pool wall as the reference point. The lidar emits near-infrared light signals in real time, which contact each target object and return to the lidar. The target objects include swimmers, floating equipment, and the pool wall. Based on this, a large amount of target point data in the pool is obtained. The target point data includes the distance and reflection intensity of the target object. Filter the target point data to remove noise points and abnormal points; Set the position of the lidar as the origin of the reference coordinate system, and construct a three-dimensional image of the pool with the processed target point data;

[0016] In real time, identify the positions of the head target points, the positions of the limbs target points, and the positions of the abdominal target points of each swimmer in the three-dimensional image of the pool. Connect the position of the head target point and the water surface position to obtain the actual head distance value between the head of each swimmer and the water surface. Extract the standard head distance value in the database, and take the difference between the actual head distance value and the standard head distance value to obtain the head distance difference;

[0017] Take the acquisition time point as the abscissa, and use the positions of the limbs target points and the positions of the abdominal target points as the ordinates to construct a two-dimensional rectangular coordinate system. Then, input the positions of the limbs target points and the positions of the abdominal target points into the coordinate system in the order of the acquisition time point. Mark the positions of the limbs target points and the positions of the abdominal target points in the coordinate system as the limb movement points and the abdominal movement points. Use smooth curves to connect the limb movement points and the abdominal movement points in sequence to obtain the curve graphs of the limb movement points and the abdominal movement points changing with time. Make tangents to the curves at the limb movement points and the abdominal movement points respectively to obtain the limb movement tangents and the abdominal movement tangents. At the same time, calculate the slopes of the limb movement tangents and the abdominal movement tangents respectively to obtain the limb movement tangent slope and the abdominal movement tangent slope;

[0018] Mark the head distance difference, the limb movement tangent slope, and the abdominal movement tangent slope as GY, GR, and GE respectively, and substitute them into the formula for calculation after normalization processing The posture abnormality index GH of the swimmer, where J1, J2, and J3 are the weight coefficients corresponding to the head distance difference, the limb movement tangent slope, and the abdominal movement tangent slope; Extract the preset posture abnormality range in the database. If the posture abnormality index is less than the minimum value of the posture abnormality range, mark this swimmer as a non-dangerous person. If the posture abnormality index is within the posture abnormality range, mark this swimmer as a person to be analyzed. If the posture abnormality index is greater than the maximum value of the posture abnormality range, mark this swimmer as a dangerous person, and generate a first warning signal and send it to the video camera module. The first warning signal includes the position information of the dangerous person and the person to be analyzed.

[0019] The video camera module analyzes the human body posture and trajectory changes of the swimmer in the video captured by the RGB camera group, determines whether the movement of the swimmer is abnormal according to the preset drowning model, and generates a second warning signal accordingly. The specific process is as follows:

[0020] When the video camera module receives the first warning signal, it obtains the position information of the dangerous person and the person to be analyzed through the RGB camera group, marks the dangerous person with an orange frame and the person to be analyzed with a yellow frame in the real-time video, and sequentially identifies the real-time states of the dangerous person and the person to be analyzed. Specifically: First, obtain the video image of the dangerous person in the swimming pool, decompose the video image frame by frame into continuous video frames, mark any two adjacent position points of the dangerous person in the continuous video frames as P1(x1, y1) and P2(x2, y2), and then obtain the time interval between these two adjacent position points as t1, and calculate using the formula to obtain the trajectory speed CW of the dangerous person;

[0021] Set the parametric equations of the trajectory as f = f(t2) and g = g(t2), and then calculate using the formula to obtain the trajectory curvature CR of the dangerous person, where f 1 (t2), f 2 (t2), g 1 (t2) and g 2 (t2) are the first-order and second-order derivatives of f = f(t2) and g = g(t2);

[0022] Sequentially identify the coordinate positions of the adjacent joint points of the dangerous person, connect the coordinates of the positions of three adjacent joint points to form a joint triangle, including the joint triangle formed by connecting the positions of the shoulder, elbow, and wrist joint points. Taking this joint triangle as the center, obtain the displacement values of the joint triangle in the vertical direction in the continuous video frames, and calculate the average value of these displacement values to obtain the average displacement value;

[0023] Select a convolutional neural network of deep learning to construct a drowning model. Specifically:

[0024] A convolutional neural network is used as the architecture of the drowning model, which is provided with a convolutional layer, a pooling layer and a fully connected layer. The sum of the characteristic quantities of the trajectory speed, trajectory curvature, comprehensive evaluation value of swimming ability and average displacement value of the dangerous person is used as the number of nodes in the convolutional layer. The drowning case data and normal swimming behavior data in the database are extracted and integrated into a drowning data set. The drowning data set is used to train the drowning model. The drowning data set is divided into a training set and a test set, and 80% of the drowning data set is used as the training set and 20% as the test set. The trajectory speed, trajectory curvature, comprehensive evaluation value of swimming ability and average displacement value of any dangerous person are used as the input characteristic quantities of the convolutional layer, and the drowning probability prediction value is used as the output characteristic value of the fully connected layer. The drowning probability prediction value is calculated through the backpropagation algorithm and iterated sequentially until the drowning probability prediction value is obtained. The preset drowning probability prediction range in the database is extracted. If the drowning probability prediction value is greater than the maximum value of the preset drowning probability prediction range, the dangerous person is marked with a red frame, and a rescue instruction is generated and sent to the alarm output module. If the drowning probability prediction value is less than the minimum value of the preset drowning probability prediction range, the dangerous person is marked with a blue frame, and the mark of the dangerous person is removed; and so on. The trajectory speed, trajectory curvature and average displacement value of the person to be analyzed are input into the drowning model to obtain the drowning probability prediction value of the person to be analyzed. If it exceeds the preset threshold, a second warning signal is generated until all dangerous persons and persons to be analyzed are marked with blue frames. It should be noted that being marked with a blue frame means that the swimming state of the swimmer is normal.

[0025] According to the second warning signal received, the alarm output module obtains the position information of the corresponding dangerous person and the video image of the drowning person and sends them to the large screen display, and vibrates the smart bracelet of the lifeguard. The specific process is as follows:

[0026] Upon receiving the second warning signal, the drowning time of the drowning person is timed at the same time. Taking the position of the drowning person as the center point, the position information of the lifeguards within the radius is obtained, and the drowning person's position and each lifeguard are connected to obtain a rescue route. The rescue routes are sorted from large to small, and the lifeguard corresponding to the smallest rescue route is marked as the preferred person. The video image and position information of the drowning person are sent to the large screen display, and the smart bracelet of the preferred person is vibrated at the same time; if the preferred person has been assigned a rescue task, the smart bracelet of the second-ranked lifeguard is vibrated until the rescue is completed.

[0027] Compared with the prior art, the beneficial effects of the present invention are as follows: Precise drowning monitoring ability: By emitting near-infrared light signals through the lidar module and measuring target point data to construct a three-dimensional image of the swimming pool, it can accurately identify various parts of the swimmers in the pool, calculate key indicators such as the head distance difference, the tangent slopes of the limbs and abdomen, and the abnormal posture index, so as to accurately determine the dangerous state of the swimmers and send out the first warning signal in a timely manner; The video camera module uses the convolutional neural network drowning model of deep learning to analyze multi-dimensional data such as the trajectory speed, trajectory curvature, comprehensive evaluation value of swimming ability, and average displacement value of the swimmers, further improving the accuracy of drowning situation judgment and effectively reducing the risks of false alarms and missed alarms.

[0028] Efficient rescue response mechanism: After receiving the second warning signal, the alarm output module can quickly obtain the location information and video images of the drowning person and send them to the large-screen display, enabling the management personnel and relevant staff to intuitively understand the on-site situation; Quickly determine the positions of the lifeguards within the radius through intelligent algorithms, plan the rescue routes and screen out the preferred personnel, and vibrate their smart bracelets to remind them, ensuring that the rescue operation can be quickly launched; If the preferred personnel already have tasks, it can automatically switch to the lifeguards ranked lower, guaranteeing the timeliness and efficiency of the rescue and increasing the chance of survival of the drowning person. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. The following drawings are not deliberately drawn to scale in actual size, and the focus is on showing the gist of the present invention.

[0030] Figure 1 It is the principle block diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts also belong to the scope of protection of the present invention.

[0032] Please refer to Figure 1 As shown, the present invention is a swimming pool drowning video monitoring system based on lidar, including a data acquisition module, a lidar module, a video camera module, an alarm output module, and a database.

[0033] The data acquisition module is used to establish a swimmer platform, install lidar sensors and RGB camera groups in the swimming pool, and collect the basic information data of the people in the swimming pool and the motion data in the pool in real time, and send them to the lidar module and the video camera module respectively. The specific process is as follows:

[0034] A platform establishment unit and a collection unit are set in the data acquisition module;

[0035] The platform establishment unit is used to build a data platform for swimmers. Swimmers upload their personal information through the platform. Specifically: first, a swimmer login interface is established through the display screen. The swimmer login interface includes weight number input, number of swimming certificates input, number of monthly swimming times input, single - time swimming distance input, and single - time swimming time input. When the swimmer completes the input of the login interface, a user interface corresponding to the swimmer is established and a new identity number of the swimmer is generated and sent to the database for storage;

[0036] Parse the basic data information uploaded by each swimmer in turn. Mark the selected data of weight, number of swimming certificates, number of monthly swimming times, single - time swimming distance, and single - time swimming time as stw1, stw2, stw3, stw4, and stw5 respectively, and substitute them into the formula for calculation after normalization processing Obtain the comprehensive swimming ability evaluation value of each swimmer. Among them, Q1, Q2, Q3, Q4, and Q5 are the weight coefficients corresponding to weight, number of swimming certificates, number of monthly swimming times, single - time swimming distance, and single - time swimming time respectively. i represents the number of each swimmer, and n represents the number of all swimmers. Synchronously send the comprehensive swimming ability evaluation value corresponding to each swimmer to the database for storage;

[0037] The collection unit equally divides the deep pool and the shallow pool into several swimming areas respectively, installs RGB cameras according to the number of swimming areas, and one RGB camera collects the data information of each swimmer in one swimming area and sends the obtained data information to the video camera module; install a lidar sensor directly above the center position of the swimming pool to collect the target point data of each swimmer in real time and send it to the lidar module.

[0038] The lidar module emits near - infrared light signals and measures the target point data, constructs a three - dimensional image of the swimming pool, identifies the parts of the swimmers in the pool, calculates the difference value, slope, and attitude anomaly index, and determines the first warning signal of the swimmer. The specific process is as follows:

[0039] Set the pool wall as the reference point. The lidar emits near-infrared light signals in real time, which contact each target object and return to the lidar. The target objects include swimmers, floating equipment, and the pool wall. Based on this, a large amount of target point data in the pool is obtained. The target point data includes the distance and reflection intensity of the target object. Filter the target point data to remove noise points and outliers; Set the lidar position as the origin of the reference coordinate system, and construct a three-dimensional image of the pool with the processed target point data;

[0040] In real time, identify the positions of the head target points, the positions of the limbs target points, and the positions of the abdominal target points of each swimmer in the three-dimensional image of the pool. Connect the position of the head target point and the water surface position to obtain the actual head distance value between the head of each swimmer and the water surface. Extract the standard head distance value in the database, and calculate the difference between the actual head distance value and the standard head distance value to obtain the head distance difference;

[0041] Take the acquisition time point as the abscissa, and use the positions of the limb target points and the abdominal target points as the ordinates to construct a two-dimensional rectangular coordinate system. Then, input the positions of the limb target points and the abdominal target points into the coordinate system in the order of the acquisition time point. Mark the positions of the limb target points and the abdominal target points in the coordinate system as limb movement points and abdominal movement points. Use smooth curves to connect the limb movement points and the abdominal movement points in sequence to obtain the curve graphs of the limb movement points and the abdominal movement points changing with time. Make tangents to the curves at the limb movement points and the abdominal movement points respectively to obtain the limb movement tangents and the abdominal movement tangents. At the same time, calculate the slopes of the limb movement tangents and the abdominal movement tangents respectively to obtain the limb movement tangent slope and the abdominal movement tangent slope;

[0042] Mark the head distance difference, the limb movement tangent slope, and the abdominal movement tangent slope as GY, GR, and GE respectively, and substitute them into the formula for calculation after normalization processing The posture abnormality index GH of the swimmer, where J1, J2, and J3 are the weight coefficients corresponding to the head distance difference, the limb movement tangent slope, and the abdominal movement tangent slope; Extract the preset posture abnormality range in the database. If the posture abnormality index is less than the minimum value of the posture abnormality range, mark this swimmer as a non-dangerous person. If the posture abnormality index is within the posture abnormality range, mark this swimmer as a person to be analyzed. If the posture abnormality index is greater than the maximum value of the posture abnormality range, mark this swimmer as a dangerous person, and generate a first warning signal and send it to the video camera module. The first warning signal includes the position information of the dangerous person and the person to be analyzed.

[0043] The video camera module analyzes the human posture and trajectory changes of the swimmer in the video captured by the RGB camera group, determines whether the movement of the swimmer is abnormal according to the preset drowning model, and generates a second warning signal accordingly. The specific process is as follows:

[0044] The video camera module receives the first warning signal, obtains the position information of the dangerous person and the person to be analyzed through the RGB camera group, marks the dangerous person with an orange frame and the person to be analyzed with a yellow frame in the real-time video, and sequentially identifies the real-time states of the dangerous person and the person to be analyzed. Specifically: First, obtain the video image of the dangerous person in the swimming pool, decompose the video image frame by frame into continuous video frames, mark any two adjacent position points of the dangerous person in the continuous video frames as P1(x1, y1) and P2(x2, y2), then obtain the time interval between these two adjacent position points as t1, and calculate using the formula to obtain the trajectory speed CW of the dangerous person;

[0045] Set the parametric equations of the trajectory as f = f(t2) and g = g(t2), then calculate using the formula to obtain the trajectory curvature CR of the dangerous person, where f 1 (t2), f 2 (t2), g 1 (t2) and g 2 (t2) are the first and second derivatives of f = f(t2) and g = g(t2);

[0046] Sequentially identify the coordinate positions of the adjacent joint points of the dangerous person, connect the coordinates of the positions of three adjacent joint points to form a joint triangle, including the joint triangle formed by connecting the positions of the shoulder, elbow and wrist joint points. Taking this joint triangle as the center, obtain the displacement values of the joint triangle in the vertical direction in the continuous video frames, and calculate the average value of these displacement values to obtain the average displacement value;

[0047] Select a convolutional neural network of deep learning to construct a drowning model. Specifically:

[0048] The convolutional neural network is adopted as the architecture of the drowning model, in which there is a convolutional layer, a pooling layer and a fully connected layer. The sum of the characteristic quantities of the trajectory speed, trajectory curvature, comprehensive evaluation value of swimming ability and average displacement value of the dangerous person is used as the number of nodes in the convolutional layer. The drowning case data and normal swimming behavior data in the database are extracted and integrated into a drowning data set. The drowning model is trained with the drowning data set. The drowning data set is divided into a training set and a test set, and 80% of the drowning data set is used as the training set and 20% as the test set. The trajectory speed, trajectory curvature, comprehensive evaluation value of swimming ability and average displacement value of any dangerous person are used as the input characteristic quantities of the convolutional layer, and the drowning probability prediction value is used as the output characteristic value of the fully connected layer; the drowning probability prediction value is calculated through the backpropagation algorithm and iterated sequentially until the drowning probability prediction value is obtained; the preset drowning probability prediction range in the database is extracted. If the drowning probability prediction value is greater than the maximum value of the preset drowning probability prediction range, the dangerous person is marked with a red frame, and a rescue instruction is generated and sent to the alarm output module. If the drowning probability prediction value is less than the minimum value of the preset drowning probability prediction range, the dangerous person is marked with a blue frame, and the marking of the dangerous person is cancelled; and so on. The trajectory speed, trajectory curvature and average displacement value of the person to be analyzed are input into the drowning model to obtain the drowning probability prediction value of the person to be analyzed. If it exceeds the preset threshold, a second warning signal is generated until all dangerous persons and persons to be analyzed are marked with blue frames. It should be noted that being marked with a blue frame indicates that the swimmer's motion state is normal.

[0049] According to the second warning signal received, the alarm output module obtains the position information of the corresponding dangerous person and the video image of the drowning person and sends them to the large screen display, and vibrates the smart bracelet of the lifeguard. The specific process is as follows:

[0050] Upon receiving the second warning signal, the drowning time of the drowning person is timed at the same time. With the position of the drowning person as the center point, the position information of the lifeguards within the radius is obtained, and the drowning person's position and each lifeguard are connected to obtain a rescue route. The rescue routes are sorted from large to small, and the lifeguard corresponding to the smallest rescue route is marked as the preferred person. The video image and position information of the drowning person are sent to the large screen display, and the smart bracelet of the preferred person is vibrated at the same time; if the preferred person has been assigned a rescue task, the smart bracelet of the second-ranked lifeguard is vibrated until the rescue is completed.

[0051] The foregoing is a description of the present invention and should not be construed as limiting thereof. Although several exemplary embodiments of the present invention have been described, those skilled in the art will readily appreciate that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of the present invention. Accordingly, all such modifications are intended to be included within the scope of the present invention as defined by the claims. It should be understood that the foregoing is a description of the present invention and should not be considered limited to the specific embodiments disclosed, and modifications to the disclosed embodiments as well as other embodiments are intended to be included within the scope of the appended claims. The present invention is defined by the claims and their equivalents.

Claims

1. A lidar-based pool drowning video monitoring system, comprising a data acquisition module, a video camera module, an alarm output module, and a database, characterized in that, It also includes a lidar module; Based on the video captured by the RGB camera group, the video camera module analyzes the human body posture and trajectory changes of the swimmers in the video, determines whether the movement of the swimmers is abnormal according to a preset drowning model, and generates a second warning signal accordingly; According to the received second warning signal, the alarm output module obtains the position information of the corresponding dangerous person and the video image of the drowning person and sends them to the large-screen display, and vibrates the smart bracelet of the lifeguard; The lidar module constructs a three-dimensional image of the swimming pool by emitting near-infrared light signals and measuring the target point data, and identifies the parts of the swimmers in the swimming pool to calculate the difference and slope. Specifically: Set the pool wall as the reference point. The lidar emits near-infrared light signals in real time to contact each target object and returns to the lidar, obtaining a large amount of target point data in the swimming pool. Filter the target point data to remove noise points and abnormal points; Use the position of the lidar as the origin of the reference coordinate system, and construct a three-dimensional image of the swimming pool with the processed target point data; In real time, identify the positions of the head target points, the positions of the limbs target points, and the positions of the abdominal target points of each swimmer in the three-dimensional image of the swimming pool. Connect the position of the head target point and the water surface position to obtain the actual head distance value between the head of each swimmer and the water surface. Extract the standard head distance value in the database, and subtract the actual head distance value from the standard head distance value to obtain the head distance difference; Taking the acquisition time point as the abscissa, and the positions of the limbs target points and the abdominal target points as the ordinate, construct a two-dimensional rectangular coordinate system, and input the positions of the limbs target points and the abdominal target points into the coordinate system in the order of the acquisition time point. Mark the positions of the limbs target points and the abdominal target points in the coordinate system as the limb movement points and the abdominal movement points. Use a smooth curve to connect the limb movement points and the abdominal movement points in sequence to obtain the curve graphs of the limb movement points and the abdominal movement points changing with time. Make tangents to the curves at the limb movement points and the abdominal movement points respectively to obtain the limb movement tangents and the abdominal movement tangents. At the same time, calculate the slopes of the limb movement tangents and the abdominal movement tangents respectively to obtain the limb movement tangent slope and the abdominal movement tangent slope.

2. The lidar-based pool drowning video monitoring system according to claim 1, wherein The lidar module calculates and obtains the posture abnormality index according to the obtained data, and determines the first warning signal of the swimmer. Specifically: Mark the head distance difference, the limb movement tangent slope, and the abdominal movement tangent slope as GY, GR, and GE respectively, and substitute them into the formula for calculation after normalization processing. The posture abnormality index GH of the swimmer, where J1, J2, and J3 are the weight coefficients corresponding to the head distance difference, the limb movement tangent slope, and the abdominal movement tangent slope; extract the preset posture abnormality range in the database. If the posture abnormality index is less than the minimum value of the posture abnormality range, mark the swimmer as a non-dangerous person. If the posture abnormality index is within the posture abnormality range, mark the swimmer as a person to be analyzed. If the posture abnormality index is greater than the maximum value of the posture abnormality range, mark the swimmer as a dangerous person and generate a first warning signal to be sent to the video camera module. The first warning signal includes the location information of the dangerous person and the person to be analyzed.

3. The lidar-based pool drowning video monitoring system according to claim 1, wherein, Based on the video captured by the RGB camera group, the video camera module analyzes the trajectory changes of the swimmers in the video, and obtains the trajectory speed and trajectory curvature of the dangerous person. The specific process is as follows: Upon receiving the first warning signal, obtain the location information of dangerous personnel and personnel to be analyzed through the RGB camera group, and mark the dangerous personnel with an orange frame and the personnel to be analyzed with a yellow frame in the live video. Then, identify the real-time status of the dangerous personnel and the personnel to be analyzed in sequence, specifically as follows: First, obtain the video footage of the dangerous personnel in the pool, decompose the video footage frame by frame into consecutive video frames, mark any two adjacent position points of the dangerous personnel in the consecutive video frames as P1(x1, y1) and P2(x2, y2), and then obtain the time interval between these two adjacent position points as t1. Calculate using the formula to obtain the trajectory speed CW of the dangerous personnel; Set the parametric equations of the trajectory as \(f = f(t_2)\) and \(g = g(t_2)\), then calculate using the formula to obtain the trajectory curvature \(CR\) of the dangerous person, where \(f\) 1 (t_2), \(f\) 2 (t_2), \(g\) 1 (t_2) and \(g\) 2 (t_2) are the first and second derivatives of \(f = f(t_2)\) and \(g = g(t_2)\).

4. The lidar-based pool drowning video monitoring system according to claim 3, wherein, The video camera module analyzes the human body posture and obtains the average displacement value. Specifically: Judge whether the movement of the swimmer is abnormal according to the preset drowning model, and generate a second warning signal accordingly. Identify the coordinate positions of the adjacent joint points of the dangerous person in sequence, and connect the coordinates of the positions of three adjacent joint points to obtain a joint triangle, including the joint triangle formed by connecting the positions of the shoulder, elbow, and wrist joint points. Taking this joint triangle as the center, obtain the displacement values of the joint triangle in the vertical direction in consecutive video frames, and calculate the mean value of the displacement values to obtain the average displacement value.

5. The lidar-based pool drowning video monitoring system according to claim 3, wherein, The video camera module selects a convolutional neural network of deep learning to construct a drowning model and obtains a drowning probability prediction value, specifically as follows: A convolutional neural network is used as the architecture of the drowning model, which is provided with a convolutional layer, a pooling layer and a fully connected layer. The sum of the characteristic quantities of the trajectory speed, trajectory curvature, comprehensive evaluation value of swimming ability and average displacement value of the dangerous person is used as the number of nodes in the convolutional layer. The drowning case data and normal swimming behavior data in the database are extracted and integrated into a drowning data set. The drowning model is trained with the drowning data set. The drowning data set is divided into a training set and a test set, and 80% of the drowning data set is used as the training set and 20% as the test set. The trajectory speed, trajectory curvature, comprehensive evaluation value of swimming ability and average displacement value of any dangerous person are used as the input characteristic quantities of the convolutional layer, and the drowning probability prediction value is used as the output characteristic value of the fully connected layer. The drowning probability prediction value is calculated and iterated successively through the backpropagation algorithm until the drowning probability prediction value is obtained. The preset drowning probability prediction range in the database is extracted. If the drowning probability prediction value is greater than the maximum value of the preset drowning probability prediction range, the dangerous person is marked with a red frame, and a rescue instruction is generated and sent to the alarm output module. If the drowning probability prediction value is less than the minimum value of the preset drowning probability prediction range, the dangerous person is marked with a blue frame, and the marking of the dangerous person is cancelled; And so on, the trajectory speed, trajectory curvature and average displacement value of the person to be analyzed are input into the drowning model to obtain the drowning probability prediction value of the person to be analyzed. If the preset threshold is exceeded, a second warning signal is generated until all dangerous persons and persons to be analyzed are marked with blue frames.

6. The lidar-based pool drowning video monitoring system according to claim 1, wherein The alarm output module, according to the received second warning signal, obtains the position information of the corresponding dangerous person and the video image of the drowning person and sends them to the large-screen display, and vibrates the smart bracelet of the lifeguard. The specific process is as follows: Receiving the second warning signal, at the same time, timing the drowning time of the drowning person, taking the position of the drowning person as the center point, obtaining the position information of the lifeguards within the radius, and connecting the position of the drowning person with each lifeguard to obtain a rescue route. The rescue routes are sorted from large to small, and the lifeguard corresponding to the smallest rescue route is marked as the preferred person. The video image and position information of the drowning person are sent to the large-screen display, and at the same time, the smart bracelet of the preferred person is vibrated. If the preferred person has been assigned a rescue task, the smart bracelet of the second-ranked lifeguard is vibrated until the rescue is completed.

7. The lidar-based pool drowning video monitoring system according to claim 1, characterized in that, The data acquisition module is used to establish a swimmer platform and obtain the comprehensive evaluation value of the swimming ability of each swimmer. The specific process is as follows: The platform establishment unit is used to build a data platform for swimmers. Swimmers upload their personal information through the platform. Specifically: First, a swimmer login interface is established through the display screen. The swimmer login interface includes weight number input, number of swimming certificates input, number of monthly swimming times input, single-swim distance input, and single-swim time input. When the swimmer completes the input on the login interface, a user interface corresponding to the swimmer is established and a new identity number of the swimmer is generated and sent to the database for storage; Parse the basic data information uploaded by each swimmer in sequence. For the selected data of weight, number of swimming certificates, monthly swimming times, single - swim distance, and single - swim time, mark weight, number of swimming certificates, monthly swimming times, single - swim distance, and single - swim time as stw1, stw2, stw3, stw4, and stw5 respectively, and after normalizing them, substitute them into the formula for calculation Obtain the comprehensive evaluation value of each swimmer's swimming ability. Among them, Q1, Q2, Q3, Q4, and Q5 are the weight coefficients corresponding to weight, number of swimming certificates, monthly swimming times, single - swim distance, and single - swim time respectively. i represents the number of each swimmer, and n represents the number of all swimmers. Synchronously send the comprehensive evaluation value of each swimmer's swimming ability to the database for storage.

8. The lidar-based pool drowning video monitoring system according to claim 7, characterized in that, The acquisition unit installs lidar sensors and RGB camera groups in the swimming pool to collect the basic information data of the people in the swimming pool and the motion data in the swimming pool in real time, and sends them to the lidar module and the video camera module respectively. Specifically: The deep pool and the shallow pool are respectively divided into several swimming areas in equal proportion. RGB cameras are installed according to the number of swimming areas. One RGB camera collects the data information of each swimmer in one swimming area, and sends the obtained data information to the video camera module; A lidar sensor is installed directly above the center position of the swimming pool to collect the target point data of each swimmer in real time and send it to the lidar module.

Images