Charging space safety monitoring method and device based on multi-sensor cooperation

Abstract

The application discloses a kind of based on the safety monitoring method and device of charging parking stall of multi-sensor cooperation, and it is related to charging safety technical field.The method comprises: the multi-sensor data of charging parking stall vehicle is collected, and three-dimensional vehicle model with temperature attribute is established;Vehicle key area is segmented using image recognition, and multidimensional feature vector is extracted;Charging parking stall safety evaluation model is constructed;Improved alpha evolution algorithm is used to iteratively optimize model parameters, convergence accuracy is improved by introducing chaotic initialization and adaptive base vector mechanism, and the optimal charging parking stall safety evaluation model is obtained;Based on the optimal charging parking stall safety evaluation model, real-time monitoring is executed, and hierarchical intervention is executed according to risk value.The application solves the problem of single monitoring means limitation and parameter subjective setting through multi-source data fusion and model parameter adaptive optimization, effectively reduces the false negative rate and false positive rate, and realizes the accurate identification and early warning of charging safety hidden danger.

Description

Technical Field

[0001] This invention relates to the field of charging safety technology, and in particular to a method and device for safety monitoring of charging parking spaces based on multi-sensor collaboration. Background Technology

[0002] With the widespread adoption of new energy vehicles, the number of charging stations has increased dramatically. However, fires involving new energy vehicles during charging are frequent, posing a significant threat to public safety. Traditional charging safety monitoring primarily relies on internal data provided by the battery management system, such as voltage, current, and temperature. However, battery management system data may suffer from communication delays, data drift, or sensor failures, and it is difficult to cover safety risks arising from non-battery internal faults, such as vehicle chassis collisions or ignition by external fire sources.

[0003] Existing technologies also include solutions that utilize external sensors for monitoring, such as cameras or smoke detectors. However, single sensors often have limitations: cameras are susceptible to lighting conditions, smoke detectors have response lag, and ordinary temperature sensors struggle to accurately locate heat sources.

[0004] Therefore, there is an urgent need for a safety monitoring method that can integrate data from multiple sensors and accurately model and assess the condition of key parts of a vehicle to improve the accuracy and timeliness of early warnings. Summary of the Invention

[0005] This invention provides a method and device for safety monitoring of charging parking spaces based on multi-sensor collaboration. This invention solves the problems of limited monitoring methods and low early warning accuracy in existing technologies.

[0006] In a first aspect, embodiments of the present invention provide a method for safety monitoring of charging parking spaces based on multi-sensor collaboration, the method comprising: Using sensor data acquisition equipment deployed in charging parking spaces, multi-sensor data of vehicles located in charging parking spaces are collected, and a three-dimensional vehicle model is established through a unified spatial coordinate system. Using an image recognition model, the key vehicle regions in the 3D vehicle model are segmented and defined as Regions of Interest (ROIs). Based on multi-sensor data, multi-dimensional feature vectors of the ROI regions are extracted. Based on the multidimensional feature vector, an initial safety assessment model for charging parking spaces is constructed. The initial safety assessment model for charging parking spaces is set with independent variables, dependent variables, and fixed quantities. The independent variables are multidimensional feature vectors, the dependent variables are safety risk values, and the fixed quantities are the model parameter set. The fixed parameters of the initial charging parking space safety assessment model are used as the optimization objectives. An improved alpha evolution algorithm is used to iteratively optimize the parameters in the parameter space to obtain the optimal model parameter set. The initial charging parking space safety assessment model is then optimized to obtain the optimal charging parking space safety assessment model. Based on the optimal safety assessment model for charging parking spaces, safety monitoring of charging parking spaces is carried out, and graded interventions are implemented according to the latest safety risk values ​​generated in real time.

[0007] The technical solution provided in this application has at least the following beneficial effects: By employing a multi-sensor collaboration involving LiDAR, thermal imagers, industrial cameras, smoke sensors, and vehicle vibration sensors, a 3D vehicle model with temperature attributes was constructed, enabling comprehensive and multi-dimensional state perception of charging vehicles. Image recognition and direct-pass filtering algorithms were used to segment ROI regions (such as the chassis battery pack area), allowing for precise focusing on high-risk areas and eliminating interference from non-critical regions, thus improving the targeting of feature extraction. A charging space safety assessment model incorporating weights and decision thresholds was built, and an improved Alpha Evolutionary Algorithm was innovatively introduced to optimize model parameters. This improved Alpha Evolutionary Algorithm utilizes Logistic mapping to initialize the population, increasing population diversity. By constructing adaptive basis vectors and an adaptive step-size update mechanism, it balances global search and local exploitation capabilities, effectively preventing the algorithm from getting trapped in local optima and improving optimization efficiency and accuracy. The optimized charging space safety assessment model can more accurately calculate safety risk values. Combined with a tiered intervention strategy, it achieves a smooth transition from early warning to severe alarms, significantly improving the safety management level of charging spaces.

[0008] In one alternative implementation, the sensor data acquisition device includes a lidar, a thermal imager, an industrial camera, a smoke sensor, and a vehicle vibration sensor. The multi-sensor data includes lidar point cloud data, thermal imaging temperature data, visible light image data, smoke detection data, and vehicle vibration signals.

[0009] In one alternative implementation, sensor data acquisition devices deployed in charging parking spaces are used to collect multi-sensor data of the vehicle located in the charging parking space, and a three-dimensional vehicle model is established through a unified spatial coordinate system, including: On the cloud server, a unified spatial coordinate system is set based on the ground center of the charging parking space, and the external parameter matrix of the lidar coordinate system and the world coordinate system of the lidar in the sensor data acquisition device is set. Using sensor data acquisition devices deployed in charging parking spaces, multi-sensor data of vehicles located in charging parking spaces is collected, pre-processed using an edge gateway, and the pre-processed multi-sensor data is uploaded to a cloud server. On the cloud server, based on the extrinsic parameter matrix, the preprocessed lidar point cloud data in the preprocessed multi-sensor data is transformed to the world coordinate system to obtain the transformed lidar point cloud data. Based on the extrinsic parameter matrix and the transformed lidar point cloud data, a pinhole camera model is used to construct the mapping relationship between the preprocessed thermal imaging temperature data and the transformed lidar point cloud data in the preprocessed multi-sensor data. Based on the mapping relationship, the temperature value in the coordinates of each pixel in the preprocessed thermal imaging temperature data is assigned to the corresponding 3D point in the converted LiDAR point cloud data, thus obtaining a 3D vehicle model with temperature attributes.

[0010] In one alternative implementation, an image recognition model is used to segment key vehicle regions in the 3D vehicle model, defined as Regions of Interest (ROIs). Based on multi-sensor data, multi-dimensional feature vectors of the ROIs are extracted, including: The preprocessed visible light image data from the preprocessed multi-sensor data is input into a pre-trained image recognition model to generate the vehicle type. Based on the vehicle type, a height range is set for the key vehicle areas, which include the vehicle chassis battery pack area and / or charging interface area. Based on the height range, a pass-through filtering algorithm is used to segment the point cloud set of the key vehicle region in the 3D vehicle model, and the segmented point cloud set is defined as the ROI region. Based on the preprocessed multi-sensor data, multi-dimensional feature vectors of the ROI region are extracted. These multi-dimensional feature vectors include temperature rise rate gradient features, vehicle posture deformation features, vibration spectrum energy features, and smoke concentration features.

[0011] In one alternative implementation, a multidimensional feature vector of the ROI region is extracted based on multi-sensor data, including: Calculate the rate of change of the highest temperature value of the model's three-dimensional points in the ROI region within the time window to obtain the temperature rise rate gradient characteristics. Based on the ROI region, the mean height change of the center point cloud of the four wheels of the vehicle is calculated to obtain the vehicle attitude deformation characteristics. Fast Fourier Transform is performed on the preprocessed vehicle vibration signal in the preprocessed multi-sensor data, and the energy spectral density in the low-frequency band is extracted to obtain the vibration spectrum energy characteristics. Read the preprocessed smoke detection data from the preprocessed multi-sensor data and perform normalization to obtain smoke concentration characteristics; By integrating the temperature rise rate gradient features, vehicle posture deformation features, vibration spectrum energy features, and smoke concentration features, a multidimensional feature vector of the ROI region is obtained.

[0012] In one optional implementation, the formula for the charging parking space safety assessment model is: In the formula, To be based on multidimensional feature vectors and model parameter set The safety risk value obtained from the safety assessment of charging parking spaces is the dependent variable in the safety assessment of charging parking spaces. This is the model parameter set, which is a fixed quantity during the safety assessment of charging parking spaces; It is a multidimensional feature vector, which serves as the independent variable in the safety assessment of charging parking spaces; For the first in the model parameter set k The weight parameters of the features; k For parameter indication; The first in the multidimensional feature vector k feature; For the first in the model parameter set k Decision threshold for features.

[0013] In one optional implementation, the fixed parameters of the initial charging parking space safety assessment model are used as the optimization objective. An improved alpha evolution algorithm is used to iteratively optimize within the parameter space to obtain the optimal model parameter set. This optimizes the initial charging parking space safety assessment model, resulting in the optimal charging parking space safety assessment model, which includes: The fixed quantities of the initial charging parking space safety assessment model are used as the optimization target, and the vector formed by the model parameter set is encoded as the spatial position of the individual in the improved alpha evolution algorithm. A fitness function based on a simulated environment is constructed, wherein the simulated environment is constructed from several historical multidimensional feature vector samples with real labels; The chaotic sequence is generated using the Logistic mapping, and then mapped to the parameter space of individuals in the improved Alpha Evolution algorithm to obtain the initial population. Using the fitness function, calculate the fitness value of each initial individual in the initial population, and update the individual with the best fitness value as the optimal individual. An evolutionary matrix is ​​generated by sampling with replacement from the initial population, and adaptive basis vectors are generated based on the evolutionary matrix. The initial population is updated based on the adaptive basis vectors to obtain the updated population. Using the fitness function, calculate the fitness value of each updated individual in the updated population, and update the updated individual with the best fitness value as the best individual; The position of the population is repeatedly updated. When the number of iterations reaches the maximum number of iterations or the fitness value of the best individual meets the requirements, the iterative optimization of the population is terminated, and the spatial position of the best individual is output. The spatial location of the optimal individual is decoded to obtain the optimal model parameter set. Based on the optimal model parameter set, the fixed quantities of the initial charging parking space safety assessment model are optimized to obtain the optimal charging parking space safety assessment model.

[0014] In one alternative implementation, the fitness function is formulated as follows: In the formula, For individuals X The fitness values ​​of the alternative charging parking space safety assessment models in the simulation environment are set according to the corresponding alternative model parameter set. To determine the correct number of alarms; This represents the number of missed reports; This represents the number of false alarms. To correctly identify the normal number of times; The base is ; For fitness weighting coefficients; Alarm response weights; This represents the alarm response time; max is the maximum value symbol.

[0015] In one alternative implementation, based on an optimal charging space safety assessment model, safety monitoring of the charging space is performed, and tiered interventions are implemented according to the latest safety risk values ​​generated in real time, including: Using sensor data acquisition devices deployed in charging parking spaces, the latest multi-sensor data of vehicles located in the charging parking spaces is collected, pre-processed using an edge gateway, and the latest pre-processed multi-sensor data is uploaded to the cloud server. On the cloud server, the latest preprocessed visible light image data from the latest preprocessed multi-sensor data is input into the image recognition model to generate the latest vehicle type. Based on the latest vehicle type, the latest ROI region of the vehicle is directly located, and based on the latest preprocessed multi-sensor data, the latest multi-dimensional feature vector of the latest ROI region is extracted. The latest multidimensional feature vector is input into the optimal charging parking space safety assessment model to conduct a safety assessment of the charging parking space and obtain the latest safety risk value. Based on the latest safety risk values, implement tiered interventions.

[0016] Secondly, embodiments of the present invention provide a charging parking space safety monitoring device based on multi-sensor collaboration, used to implement a charging parking space safety monitoring method, the device comprising: The data acquisition and 3D modeling unit is used to collect multi-sensor data of vehicles located in charging spaces using sensor data acquisition equipment deployed in charging spaces, and to establish a 3D vehicle model through a unified spatial coordinate system. The image recognition and region segmentation unit is used to segment key vehicle regions in a 3D vehicle model using an image recognition model, defining them as Regions of Interest (ROIs), and extracting multidimensional feature vectors of the ROIs based on multi-sensor data. The evaluation model building unit is used to construct an initial safety evaluation model for charging parking spaces based on multi-dimensional feature vectors. The model parameter optimization unit is used to take the fixed quantities of the initial charging parking space safety assessment model as the optimization target, and use the improved alpha evolution algorithm to iteratively optimize in the parameter space to obtain the optimal model parameter set, and optimize the initial charging parking space safety assessment model to obtain the optimal charging parking space safety assessment model. The safety monitoring unit is used to perform safety monitoring of charging spaces based on the optimal safety assessment model for charging spaces, and to perform graded interventions based on the latest safety risk values ​​generated in real time.

[0017] A third aspect of this invention provides an electronic device, which includes: At least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by at least one processor, such that the at least one processor can perform the method proposed in the first aspect of the present invention.

[0018] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in the first aspect of the present invention. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the electronic device structure of the hardware operating environment involved in the embodiments of the present invention; Figure 2 This is a flowchart illustrating the steps of a charging parking space safety monitoring method based on multi-sensor collaboration provided in an embodiment of the present invention. Figure 3 This is a functional unit diagram of a charging parking space safety monitoring device based on multi-sensor collaboration provided in an embodiment of the present invention. Detailed Implementation

[0020] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0021] The present invention will be further described below with reference to the accompanying drawings.

[0022] Reference Figure 1 , Figure 1 This is a schematic diagram of the electronic device structure of the hardware operating environment involved in the embodiments of the present invention.

[0023] like Figure 1 As shown, the electronic device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0024] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0025] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a data storage module, a network communication module, a user interface module, and an electronic program for a charging parking space safety monitoring device based on multi-sensor collaboration.

[0026] exist Figure 1In the electronic device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the electronic device of the present invention can be set in the electronic device. The electronic device calls the electronic program of the charging parking space safety monitoring device based on multi-sensor collaboration stored in the memory 1005 through the processor 1001, and executes the charging parking space safety monitoring method based on multi-sensor collaboration provided in the embodiment of the present invention.

[0027] Reference Figure 2 The present invention provides a method for safety monitoring of charging parking spaces based on multi-sensor collaboration, the method comprising: S201: Using sensor data acquisition equipment deployed in charging parking spaces, collect multi-sensor data of vehicles located in charging parking spaces, and establish a three-dimensional vehicle model through a unified spatial coordinate system. S202: Using an image recognition model, the key vehicle regions in the 3D vehicle model are segmented and defined as Region of Interest (ROI). Based on multi-sensor data, multi-dimensional feature vectors of the ROI regions are extracted. S203: Based on the multidimensional feature vector, an initial safety assessment model for charging parking spaces is constructed. The initial safety assessment model for charging parking spaces is set with independent variables, dependent variables, and fixed quantities. The independent variables are multidimensional feature vectors, the dependent variables are safety risk values, and the fixed quantities are model parameter sets. S204: Using the fixed parameters of the initial charging parking space safety assessment model as the target to be optimized, the improved alpha evolution algorithm is used to iteratively optimize in the parameter space to obtain the optimal model parameter set, and the initial charging parking space safety assessment model is optimized to obtain the optimal charging parking space safety assessment model. S205: Based on the optimal safety assessment model for charging parking spaces, perform safety monitoring of charging parking spaces and implement graded interventions according to the latest safety risk values ​​generated in real time.

[0028] The technical solution provided in this application has at least the following beneficial effects: By employing a multi-sensor collaboration involving LiDAR, thermal imagers, industrial cameras, smoke sensors, and vehicle vibration sensors, a 3D vehicle model with temperature attributes was constructed, enabling comprehensive and multi-dimensional state perception of charging vehicles. Image recognition and direct-pass filtering algorithms were used to segment ROI regions (such as the chassis battery pack area), allowing for precise focusing on high-risk areas and eliminating interference from non-critical regions, thus improving the targeting of feature extraction. A charging space safety assessment model incorporating weights and decision thresholds was built, and an improved Alpha Evolutionary Algorithm was innovatively introduced to optimize model parameters. This improved Alpha Evolutionary Algorithm utilizes Logistic mapping to initialize the population, increasing population diversity. By constructing adaptive basis vectors and an adaptive step-size update mechanism, it balances global search and local exploitation capabilities, effectively preventing the algorithm from getting trapped in local optima and improving optimization efficiency and accuracy. The optimized charging space safety assessment model can more accurately calculate safety risk values. Combined with a tiered intervention strategy, it achieves a smooth transition from early warning to severe alarms, significantly improving the safety management level of charging spaces.

[0029] In one alternative implementation, the sensor data acquisition device includes a lidar, a thermal imager, an industrial camera, a smoke sensor, and a vehicle vibration sensor. The multi-sensor data includes lidar point cloud data, thermal imaging temperature data, visible light image data, smoke detection data, and vehicle vibration signals.

[0030] In one alternative implementation, sensor data acquisition devices deployed in charging parking spaces are used to collect multi-sensor data of the vehicle located in the charging parking space, and a three-dimensional vehicle model is established through a unified spatial coordinate system, including: S2011: On the cloud server, based on the center of the charging parking space, a unified spatial coordinate system is set, and the external parameter matrix of the lidar coordinate system and the world coordinate system in the sensor data acquisition equipment is set. ,in, Let be a rotation matrix. It is a translation vector; In this embodiment, a lidar is installed above or to the side of the charging parking space to acquire the point cloud of the vehicle's external contour; a thermal imager is installed to acquire the surface temperature distribution; an industrial camera is installed for visible light imaging; and a smoke sensor and a vehicle vibration sensor are installed on or near the parking space floor. Establish a world coordinate system (e.g., with the center of the parking space ground as the origin, the X-axis pointing in the length direction of the parking space, and the Z-axis pointing vertically upward), and obtain the rotation matrix and translation vector between the lidar coordinate system and the world coordinate system through calibration algorithms (such as checkerboard calibration method or point cloud registration method); S2012: Use sensor data acquisition equipment deployed in charging parking spaces to collect multi-sensor data of vehicles located in charging parking spaces, use edge gateways for preprocessing, and upload the preprocessed multi-sensor data to cloud servers. In this embodiment, the preprocessing includes: LiDAR point cloud data: outlier filtering and downsampling; Thermal imaging temperature data: Non-uniformity correction and noise filtering are performed; Visible light image data: Denoising and white balance adjustment are performed; Smoke detection data: Filtered and noise-reduced; Vehicle vibration signal: DC component removal processing is performed; S2013: On the cloud server, based on the extrinsic parameter matrix, the preprocessed LiDAR point cloud data from the preprocessed multi-sensor data is transformed to the world coordinate system to obtain the transformed LiDAR point cloud data. The formula is: In the formula, The first part of the converted lidar point cloud data j Three-dimensional point; j The point indicator is used to ensure that all vehicle point clouds are in the same three-dimensional space, which facilitates subsequent spatial segmentation and analysis. S2014: Based on the extrinsic parameter matrix and the transformed LiDAR point cloud data, using a pinhole camera model, construct the mapping relationship between the preprocessed thermal imaging temperature data and the transformed LiDAR point cloud data in the preprocessed multi-sensor data. The formula is: In the formula, Scale factor; For the first j Pixel coordinates; This is the intrinsic parameter matrix of the thermal imager; S2015: Based on the mapping relationship, the temperature value of each pixel coordinate in the preprocessed thermal imaging temperature data is assigned to the corresponding 3D point in the transformed LiDAR point cloud data, resulting in a 3D vehicle model with temperature attributes. The formula is: In the formula, A three-dimensional vehicle model; For the first j Three-dimensional points of the model; For the first j The coordinates of the three-dimensional points in the model correspond to the first point in the transformed LiDAR point cloud data. j Coordinates of a 3D point; For the first jThe temperature values ​​of the three-dimensional points in the model correspond to the first point in the preprocessed thermal imaging temperature data. j Temperature value at pixel coordinates; For the first j The LiDAR reflection intensity of three-dimensional points in the model; the model not only includes the vehicle's geometric shape, but also carries surface thermal distribution information.

[0031] In one alternative implementation, an image recognition model is used to segment key vehicle regions in the 3D vehicle model, defined as Regions of Interest (ROIs). Based on multi-sensor data, multi-dimensional feature vectors of the ROIs are extracted, including: S2021: Input the preprocessed visible light image data from the preprocessed multi-sensor data into a pre-trained image recognition model (such as YOLO series or residual network, etc.) to generate the vehicle type (such as compact SUV, large bus, etc.). S022: Set the height range based on the vehicle's key areas preset according to the vehicle type. The critical areas of the vehicle include the vehicle chassis battery pack area (a high-risk area for thermal runaway) and / or the charging interface area (which is prone to overheating due to poor contact). S2023: Based on the height range, a direct-pass filtering algorithm is used to segment the point cloud set of the key vehicle regions in the 3D vehicle model, and the segmented point cloud set is defined as the ROI region, with the formula as follows: In the formula, Region of Interest (ROI); For the first j Height values ​​of 3D points in the model; These are the minimum and maximum values ​​for the height range of the vehicle's critical areas. S2024: Based on the preprocessed multi-sensor data, extract the multi-dimensional feature vector of the ROI region. The multi-dimensional feature vector includes temperature rise rate gradient features, vehicle posture deformation features, vibration spectrum energy features, and smoke concentration features.

[0032] In one alternative implementation, a multidimensional feature vector of the ROI region is extracted based on multi-sensor data, including: S20241: Calculate the rate of change of the highest temperature value of the model's three-dimensional points in the ROI region within the time window to obtain the temperature rise rate gradient feature. The formula is: In the formula, The temperature rise rate gradient characteristic; ROI region At any moment The highest temperature; For time indication; The time window length; early stages of a fire are often accompanied by localized temperature rises, and the rate of temperature rise is a better indicator of the trend of thermal runaway than a single temperature value. S20242: Based on the ROI region, calculate the mean height change of the center point cloud of the four wheels of the vehicle to obtain the vehicle attitude deformation characteristics. The formula is as follows: In the formula, Vehicle attitude deformation characteristics; For the ROI region, the first Wheel area at time The average height; This is the initial moment when the vehicle is positioned. This is an indicator for the wheel area; battery thermal runaway may cause bulging, resulting in chassis deformation or vehicle tilting. It quantifies minute changes in vehicle posture with high sensitivity. S20243: Perform a Fast Fourier Transform on the preprocessed vehicle vibration signal from the preprocessed multi-sensor data, and extract the energy spectral density in the low-frequency band to obtain the vibration spectrum energy characteristics. The formula is as follows: In the formula, The vibrational spectrum energy characteristics; The frequency spectrum function of the preprocessed vehicle vibration signal is obtained by performing a fast Fourier transform on the preprocessed vehicle vibration signal. For frequency variables; These are the lower and upper limits of the low-frequency band; For frequency micro-elements; the intensification of internal chemical reactions in the battery or the opening of the vent valve may cause vibrations at specific frequencies. The energy spectral density of specific low-frequency bands (corresponding to the frequency of damage to the internal structure of the battery) is extracted as a precursor feature of mechanical failure. S20244: Read the preprocessed smoke detection data from the preprocessed multi-sensor data, and perform normalization to obtain the smoke concentration characteristics. The formula is: In the formula, Characteristic of smoke concentration; This is the smoke detection data after preprocessing; This serves as a benchmark for smoke detection data; This represents the maximum value of the smoke detection data; smoke is a direct indicator of fire. The smoke sensor values ​​are read and normalized to eliminate differences in sensor dimensions. S20245: Integrating temperature rise rate gradient features, vehicle attitude deformation features, vibration spectrum energy features, and smoke concentration features, a multidimensional feature vector of the ROI region is obtained. .

[0033] In one optional implementation, the formula for the charging parking space safety assessment model is: In the formula, To be based on multidimensional feature vectors and model parameter set The safety risk value obtained from the safety assessment of charging parking spaces is the dependent variable in the safety assessment of charging parking spaces. This is the model parameter set, which is a fixed quantity during the safety assessment of charging parking spaces; It is a multidimensional feature vector, which serves as the independent variable in the safety assessment of charging parking spaces; For the first in the model parameter set k The weight parameters of the features; k For parameter indication; The first in the multidimensional feature vector k feature; For the first in the model parameter set k Decision threshold for features.

[0034] In one optional implementation, the fixed parameters of the initial charging parking space safety assessment model are used as the optimization objective. An improved alpha evolution algorithm is used to iteratively optimize within the parameter space to obtain the optimal model parameter set. This optimizes the initial charging parking space safety assessment model, resulting in the optimal charging parking space safety assessment model, which includes: S2041: The fixed quantities of the initial charging parking space safety assessment model are used as the optimization target, and the vector formed by the model parameter set is encoded as the spatial position of the individual in the improved alpha evolution algorithm. S2042: Construct a fitness function based on a simulated environment, wherein the simulated environment is constructed by setting a number of historical multidimensional feature vector samples with real labels of "safe", "early failure" and "critical failure"; S2043: Use Logistic mapping to generate chaotic sequences, and map the chaotic sequences to the parameter space of individuals in the improved Alpha Evolution algorithm to obtain the initial population; The formula is: In the formula, For the first n+ 1. n There are several chaotic variables whose values ​​range from [0, 1]. The stability coefficient is typically 4. This sequence is ergodic and random, ensuring that the initial population is uniformly distributed in the solution space, avoiding getting trapped in local optima, which is superior to traditional random initialization. n Indicator of chaotic variables; In the formula, For the initial population, the first i An initial individual; For the first i One chaotic variable; These are the upper and lower limits of the parameter space; S2044: Using the fitness function, calculate the fitness value of each initial individual in the initial population, and update the individual with the best fitness value as the optimal individual; S2045: Generate an evolutionary matrix by sampling with replacement from the initial population, and generate adaptive basis vectors based on the evolutionary matrix, using the following formula: In the formula, For the first t The adaptive basis vectors for the next iteration; The historical optimal directions for the diagonal matrix path and the weighted matrix path; The learning rate for the evolutionary path is the learning rate for the diagonal matrix path and the weighted matrix path; This is a diagonal extraction function; This is the diagonal parameter matrix extracted from the evolution matrix; A random number between (0, 1); For from the first t The first one randomly selected from the population in the second iteration Individual, the initial individual in the first iteration; For individual indicators; K The total number of random samples; These are the parameter weighting coefficients; t This represents the current iteration number; In the formula, For the randomly selected number The initial fitness value of an individual; In the formula, This is the threshold for the number of iterations; S2046: Based on the adaptive basis vectors, the initial population is updated to obtain the updated population, as shown in the formula: In the formula, For the first t+ 1 ,t In the updated population of the next iteration i The updated individual, in the first iteration, For the initial individual; For the firstt The adaptive step size for the next iteration; for; For the first t The convergence factor of the next iteration; For the first t Two distinct individuals are randomly selected from the evolutionary matrix in the next iteration; In the formula, These are the upper and lower limits of the parameter space; From an individual perspective; S2047: Using the fitness function, calculate the fitness value of each updated individual in the updated population, and update the updated individual with the best fitness value as the best individual; S2048: Repeatedly update the position of the population. When the number of iterations reaches the maximum number of iterations or the fitness value of the best individual meets the requirements, terminate the iterative optimization of the population and output the spatial position of the best individual. S2049: Decode the spatial location of the optimal individual to obtain the optimal model parameter set, and optimize the fixed quantities of the initial charging parking space safety assessment model based on the optimal model parameter set to obtain the optimal charging parking space safety assessment model.

[0035] In one alternative implementation, the fitness function is formulated as follows: In the formula, For individuals X The fitness values ​​of the alternative charging parking space safety assessment models in the simulation environment are set according to the corresponding alternative model parameter set. To determine the correct number of alarms; This represents the number of missed reports; This represents the number of false alarms. To correctly identify the normal number of times; The base is ; For fitness weighting coefficients; Alarm response weights; This represents the alarm response time; max is the maximum value symbol.

[0036] In one alternative implementation, based on an optimal charging space safety assessment model, safety monitoring of the charging space is performed, and tiered interventions are implemented according to the latest safety risk values ​​generated in real time, including: S2051: Use sensor data acquisition equipment deployed in charging parking spaces to collect the latest multi-sensor data of vehicles located in charging parking spaces, use an edge gateway for preprocessing, and upload the latest preprocessed multi-sensor data to the cloud server. S2052: On the cloud server, the latest preprocessed visible light image data from the latest preprocessed multi-sensor data is input into the image recognition model to generate the latest vehicle type. S2053: Based on the latest vehicle type, directly locate the latest ROI region of the vehicle, and extract the latest multidimensional feature vector of the latest ROI region based on the latest preprocessed multi-sensor data. S2054: Input the latest multidimensional feature vector into the optimal charging parking space safety assessment model to conduct a charging parking space safety assessment and obtain the latest safety risk value; S2055: Implement tiered interventions based on the latest safety risk values; In this embodiment, if the latest safety risk value is less than 0, the vehicle in the charging parking space is determined to be safe, and charging is maintained. If the latest safety risk value is less than the risk threshold and greater than or equal to 0, the vehicle in the charging space is determined to be at low risk. An early alarm needs to be issued, the charging rate reduced, and the early alarm information is pushed to the monitoring center or the vehicle user. If the latest safety risk value is greater than or equal to the risk threshold, the vehicle in the charging space is determined to be at high risk, and a serious alarm needs to be triggered. This involves cutting off the charging power, activating the fire alarm and audible and visual alarm, and pushing the serious alarm information to the monitoring center or the vehicle user.

[0037] This invention also provides a charging parking space safety monitoring device 300 based on multi-sensor collaboration, see reference. Figure 3 The device may include the following units: The data acquisition and 3D modeling unit 301 is used to collect multi-sensor data of the vehicle located in the charging parking space using sensor data acquisition equipment deployed in the charging parking space, and to establish a 3D vehicle model through a unified spatial coordinate system. The image recognition and region segmentation unit 302 is used to segment the key vehicle region in the three-dimensional vehicle model using an image recognition model, define it as the ROI region, and extract the multi-dimensional feature vector of the ROI region based on multi-sensor data. The evaluation model building unit 303 is used to build an initial safety evaluation model for charging parking spaces based on multidimensional feature vectors. The model parameter optimization unit 304 is used to take the fixed quantities of the initial charging parking space safety assessment model as the optimization target, and use the improved alpha evolution algorithm to iteratively optimize in the parameter space to obtain the optimal model parameter set, and optimize the initial charging parking space safety assessment model to obtain the optimal charging parking space safety assessment model. The safety monitoring unit 305 is used to perform safety monitoring of charging parking spaces based on the optimal safety assessment model for charging parking spaces, and to perform graded interventions based on the latest safety risk values ​​generated in real time.

[0038] Based on the same inventive concept, another embodiment of the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. Memory, used to store computer programs; When the processor executes the program stored in the memory, it implements the multi-sensor collaborative charging space safety monitoring method of the present invention.

[0039] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EI) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of representation, only one thick line is used in the diagram, but this does not indicate that there is only one bus or one type of bus. The communication interface is used for communication between the aforementioned terminal and other devices. The memory can include Random Access Memory (RAM), or non-volatile memory, such as at least one disk storage device. Optionally, the memory can also be at least one storage device located remotely from the aforementioned processor.

[0040] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0041] Furthermore, to achieve the above objectives, embodiments of the present invention also propose a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the charging parking space safety monitoring method based on multi-sensor collaboration of the embodiments of the present invention.

[0042] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable hardware devices (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0043] The embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (apparatus), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0044] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0045] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0046] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. "And / or" indicates that either one or both can be chosen. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes the element.

[0047] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for safety monitoring of charging parking spaces based on multi-sensor collaboration, characterized in that, The method includes: Using sensor data acquisition equipment deployed in charging parking spaces, multi-sensor data of vehicles located in charging parking spaces are collected, and a three-dimensional vehicle model is established through a unified spatial coordinate system. Using an image recognition model, the key vehicle regions in the 3D vehicle model are segmented and defined as Regions of Interest (ROIs). Based on multi-sensor data, multi-dimensional feature vectors of the ROI regions are extracted. Based on the multidimensional feature vector, an initial safety assessment model for charging parking spaces is constructed. The initial safety assessment model for charging parking spaces is set with independent variables, dependent variables, and fixed quantities. The independent variables are multidimensional feature vectors, the dependent variables are safety risk values, and the fixed quantities are the model parameter set. The fixed parameters of the initial charging parking space safety assessment model are used as the optimization objectives. An improved alpha evolution algorithm is used to iteratively optimize the parameters in the parameter space to obtain the optimal model parameter set. The initial charging parking space safety assessment model is then optimized to obtain the optimal charging parking space safety assessment model. Based on the optimal safety assessment model for charging parking spaces, safety monitoring of charging parking spaces is carried out, and graded interventions are implemented according to the latest safety risk values ​​generated in real time.

2. The charging parking space safety monitoring method based on multi-sensor collaboration according to claim 1, characterized in that, The sensor data acquisition equipment includes lidar, thermal imager, industrial camera, smoke sensor, and vehicle vibration sensor. The multi-sensor data includes lidar point cloud data, thermal imaging temperature data, visible light image data, smoke detection data, and vehicle vibration signals.

3. The charging parking space safety monitoring method based on multi-sensor collaboration according to claim 2, characterized in that, Using sensor data acquisition equipment deployed in charging parking spaces, multi-sensor data of vehicles located in the charging spaces is collected, and a three-dimensional vehicle model is established using a unified spatial coordinate system, including: On the cloud server, a unified spatial coordinate system is set based on the ground center of the charging parking space, and the external parameter matrix of the lidar coordinate system and the world coordinate system of the lidar in the sensor data acquisition device is set. Using sensor data acquisition devices deployed in charging parking spaces, multi-sensor data of vehicles located in charging parking spaces is collected, pre-processed using an edge gateway, and the pre-processed multi-sensor data is uploaded to a cloud server. On the cloud server, based on the extrinsic parameter matrix, the preprocessed lidar point cloud data in the preprocessed multi-sensor data is transformed to the world coordinate system to obtain the transformed lidar point cloud data. Based on the extrinsic parameter matrix and the transformed lidar point cloud data, a mapping relationship between the preprocessed thermal imaging temperature data and the transformed lidar point cloud data in the preprocessed multi-sensor data is constructed using a pinhole camera model. Based on the mapping relationship, the temperature value in the coordinates of each pixel in the preprocessed thermal imaging temperature data is assigned to the corresponding 3D point in the converted LiDAR point cloud data, thus obtaining a 3D vehicle model with temperature attributes.

4. The charging parking space safety monitoring method based on multi-sensor collaboration according to claim 3, characterized in that, Using an image recognition model, key vehicle regions in a 3D vehicle model are segmented and defined as Regions of Interest (ROIs). Based on multi-sensor data, multi-dimensional feature vectors of the ROIs are extracted, including: The preprocessed visible light image data from the preprocessed multi-sensor data is input into a pre-trained image recognition model to generate the vehicle type. Based on the vehicle type, a height range is set for the key vehicle areas, which include the vehicle chassis battery pack area and / or charging interface area. Based on the height range, a pass-through filtering algorithm is used to segment the point cloud set of the key vehicle region in the 3D vehicle model, and the segmented point cloud set is defined as the ROI region. Based on the preprocessed multi-sensor data, multi-dimensional feature vectors of the ROI region are extracted. These multi-dimensional feature vectors include temperature rise rate gradient features, vehicle posture deformation features, vibration spectrum energy features, and smoke concentration features.

5. The charging parking space safety monitoring method based on multi-sensor collaboration according to claim 4, characterized in that, Based on multi-sensor data, multi-dimensional feature vectors of the ROI region are extracted, including: Calculate the rate of change of the highest temperature value of the model's three-dimensional points in the ROI region within the time window to obtain the temperature rise rate gradient characteristics. Based on the ROI region, the mean height change of the center point cloud of the four wheels of the vehicle is calculated to obtain the vehicle attitude deformation characteristics. Fast Fourier Transform is performed on the preprocessed vehicle vibration signal in the preprocessed multi-sensor data, and the energy spectral density in the low-frequency band is extracted to obtain the vibration spectrum energy characteristics. Read the preprocessed smoke detection data from the preprocessed multi-sensor data and perform normalization to obtain smoke concentration characteristics; By integrating the temperature rise rate gradient features, vehicle posture deformation features, vibration spectrum energy features, and smoke concentration features, a multidimensional feature vector of the ROI region is obtained.

6. The charging parking space safety monitoring method based on multi-sensor collaboration according to claim 5, characterized in that, The formula for the safety assessment model of the charging parking space is: In the formula, To be based on multidimensional feature vectors and model parameter set The safety risk value obtained from the safety assessment of charging parking spaces is the dependent variable in the safety assessment of charging parking spaces. This is the model parameter set, which is a fixed quantity during the safety assessment of charging parking spaces; It is a multidimensional feature vector, which serves as the independent variable in the safety assessment of charging parking spaces; For the first in the model parameter set k The weight parameters of the features; k For parameter indication; The first in the multidimensional feature vector k feature; For the first in the model parameter set k Decision threshold for features.

7. The charging parking space safety monitoring method based on multi-sensor collaboration according to claim 6, characterized in that, Using the fixed parameters of the initial charging parking space safety assessment model as the optimization objective, an improved alpha evolution algorithm is used to iteratively optimize within the parameter space to obtain the optimal model parameter set. This optimizes the initial charging parking space safety assessment model, resulting in the optimal charging parking space safety assessment model, which includes: The fixed quantities of the initial charging parking space safety assessment model are used as the optimization target, and the vector formed by the model parameter set is encoded as the spatial position of the individual in the improved alpha evolution algorithm. A fitness function based on a simulated environment is constructed, wherein the simulated environment is constructed from several historical multidimensional feature vector samples with real labels; The chaotic sequence is generated using the Logistic mapping, and then mapped to the parameter space of individuals in the improved Alpha Evolution algorithm to obtain the initial population. Using the fitness function, calculate the fitness value of each initial individual in the initial population, and update the individual with the best fitness value as the optimal individual. An evolutionary matrix is ​​generated by sampling with replacement from the initial population, and adaptive basis vectors are generated based on the evolutionary matrix. The initial population is updated based on the adaptive basis vectors to obtain the updated population. Using the fitness function, calculate the fitness value of each updated individual in the updated population, and update the updated individual with the best fitness value as the best individual; The position of the population is repeatedly updated. When the number of iterations reaches the maximum number of iterations or the fitness value of the best individual meets the requirements, the iterative optimization of the population is terminated, and the spatial position of the best individual is output. The spatial location of the optimal individual is decoded to obtain the optimal model parameter set. Based on the optimal model parameter set, the fixed quantities of the initial charging parking space safety assessment model are optimized to obtain the optimal charging parking space safety assessment model.

8. The charging parking space safety monitoring method based on multi-sensor collaboration according to claim 7, characterized in that, The formula for the fitness function is: In the formula, For individuals X The fitness values ​​of the alternative charging parking space safety assessment models in the simulation environment are set according to the corresponding alternative model parameter set. To determine the correct number of alarms; This represents the number of missed reports; This represents the number of false alarms. To correctly identify the normal number of times; The base is ; For fitness weighting coefficients; Alarm response weights; This represents the alarm response time; max is the maximum value symbol.

9. The charging parking space safety monitoring method based on multi-sensor collaboration according to claim 8, characterized in that, Based on the optimal safety assessment model for charging parking spaces, safety monitoring of charging parking spaces is performed, and tiered interventions are implemented according to the latest safety risk values ​​generated in real time, including: Using sensor data acquisition devices deployed in charging parking spaces, the latest multi-sensor data of vehicles located in the charging parking spaces is collected, pre-processed using an edge gateway, and the latest pre-processed multi-sensor data is uploaded to the cloud server. On the cloud server, the latest preprocessed visible light image data from the latest preprocessed multi-sensor data is input into the image recognition model to generate the latest vehicle type. Based on the latest vehicle type, the latest ROI region of the vehicle is directly located, and based on the latest preprocessed multi-sensor data, the latest multi-dimensional feature vector of the latest ROI region is extracted. The latest multidimensional feature vector is input into the optimal charging parking space safety assessment model to conduct a safety assessment of the charging parking space and obtain the latest safety risk value. Based on the latest safety risk values, implement tiered interventions.

10. A charging parking space safety monitoring device based on multi-sensor collaboration, used to implement the charging parking space safety monitoring method as described in any one of claims 1-9, characterized in that, The device includes: The data acquisition and 3D modeling unit is used to collect multi-sensor data of vehicles located in charging spaces using sensor data acquisition equipment deployed in charging spaces, and to establish a 3D vehicle model through a unified spatial coordinate system. The image recognition and region segmentation unit is used to segment key vehicle regions in a 3D vehicle model using an image recognition model, defining them as Regions of Interest (ROIs), and extracting multidimensional feature vectors of the ROIs based on multi-sensor data. The evaluation model building unit is used to construct an initial safety evaluation model for charging parking spaces based on multi-dimensional feature vectors. The model parameter optimization unit is used to take the fixed quantities of the initial charging parking space safety assessment model as the optimization target, and use the improved alpha evolution algorithm to iteratively optimize in the parameter space to obtain the optimal model parameter set, and optimize the initial charging parking space safety assessment model to obtain the optimal charging parking space safety assessment model. The safety monitoring unit is used to perform safety monitoring of charging spaces based on the optimal safety assessment model for charging spaces, and to perform graded interventions based on the latest safety risk values ​​generated in real time.

Images