Unmanned aerial vehicle patrol intelligent early warning method and system for subway security area construction

Abstract

The invention discloses an unmanned aerial vehicle patrol intelligent early warning method and system for subway security area construction, and relates to the technical field of unmanned aerial vehicles, and the method comprises the following steps: S10, planning a patrol route path and flight parameters of an unmanned aerial vehicle; s20, using the unmanned aerial vehicle to collect a patrol video along the set route; s30, determining the position of a subway security area in the patrol video and drawing a boundary line of the subway security area; s40, detecting risk construction equipment in the subway security area in the boundary line by using the target recognition model, and performing screenshot on the patrol video according to the detected risk construction equipment; and S50, acquiring GPS information of the screenshot, generating early warning information in combination with the screenshot, the GPS information of the screenshot and the service data, and sending the early warning information to a worker for processing. The beneficial effect of the invention is that a firm and reliable guarantee is provided for effective supervision of a subway security area.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle technology, and more specifically, to an unmanned aerial vehicle inspection intelligent early warning method and system for subway security area construction. Background Art

[0002] As an important livelihood project in big cities, the safe operation of subways depends not only on engineering construction and operation guarantee, but also on many external factors. In order to avoid the impact of external operations such as construction and exploration around the track and protect the safe use of rail transit structures, the construction department will set up subway safety protection zones within a specific range around the rail transit lines and related structures, and inspect and investigate high-risk illegal construction such as well drilling, pipe jacking, and earth excavation in the security zone. However, traditional inspections mainly rely on manual inspections, which are inefficient and labor-intensive. Moreover, it is not easy for personnel and vehicles to reach some complex areas, resulting in many difficulties and loopholes in the inspection work.

[0003] In today's era of rapid development of the security industry and computer vision technology, in order to ensure the safety of the subway and prevent unauthorized construction activities from damaging the rail transit structure, the use of fixed cameras to monitor in the subway security area has become a mainstream method. However, this method has some obvious shortcomings in practical applications. First, from the perspective of monitoring coverage, the subway security area is complex, with not only undulating terrain, but also dense distribution of various types of building facilities. Fixed cameras are limited by their installation position and viewing angle and cannot take into account all directions. In particular, subway lines often extend for several kilometers, and their security areas also extend accordingly. If all cameras are to be covered, it is not only costly, but many areas do not have the conditions for installation. Secondly, from the perspective of target recognition, occlusion and overlap between different targets are extremely common, which undoubtedly increases the difficulty of timely discovery of safety hazards.

[0004] To this end, the present invention provides an intelligent early warning method and system for drone inspections in subway security area construction, which can not only conduct all-round inspections of subway security areas, but also increase the probability of discovering risky construction equipment, and quickly obtain the location information of risky construction equipment and send it to staff in a timely manner, thereby providing a solid and reliable guarantee for the effective supervision of subway security areas. Summary of the invention

[0005] In order to overcome the shortcomings of the prior art, the present invention provides an intelligent early warning method and system for drone inspection of subway security area construction, which can not only conduct an all-round inspection of the subway security area, but also improve the probability of discovering risky construction equipment, and can also quickly obtain the location information of risky construction equipment and send it to the staff in time, so as to provide a solid and reliable guarantee for the effective supervision of the subway security area.

[0006] The technical solution adopted by the present invention to solve the technical problem is: an intelligent early warning method for unmanned aerial vehicle inspection of subway security area construction, the improvement of which is that the unmanned aerial vehicle inspection method for subway security area construction includes the following steps: S10, planning the inspection route and flight parameters of the drone; S20, using drones to collect patrol videos along a set route; S30, determining the location of the subway security area in the patrol video and drawing a boundary line of the subway security area; S40, using the target recognition model to detect risky construction equipment in the subway security area within the boundary line, and taking screenshots of the inspection video according to the detected risky construction equipment; S50, obtaining the GPS information of the screenshot, and combining the screenshot, the GPS information of the screenshot and the business data to generate warning information and send it to the staff for processing.

[0007] Furthermore, in step S10, the specific method of planning the inspection route of the drone is: Obtain an accurate map of the subway line to be inspected, including the route, station locations, and surrounding terrain; Determine the scope of the subway security area at different sections of the subway line: set the boundary line of the subway security area L meters outside the center line of the subway track, where the value range of L is between 30-70; Based on the subway line to be inspected, the drone's flight path is set just above the center line of the subway track.

[0008] Furthermore, in step 10, the flight parameters of the drone include flight altitude and flight speed, wherein the flight altitude of the drone is set between 70m-100m to ensure that the video shot by the drone can completely cover the subway security area on both sides of the subway line; the flight speed of the drone is set between 5m / s-10m / s to meet the inspection accuracy.

[0009] Furthermore, in step S20, the drone needs to collect GPS data of the drone in real time during the process of collecting the patrol video, and the GPS data includes longitude, latitude, altitude and timestamp.

[0010] Furthermore, in step S30, the specific method of determining the location of the subway security area in the patrol video and drawing the boundary line of the subway security area is: S301, drawing a virtual straight line perpendicular to the horizontal direction and passing through the center point of the image in the inspection video as a center line, wherein the center point of the image is a projection of the position of the drone; S302, according to the distance Pd between the boundary line of the subway security area and the center line of the image, determine the position of the boundary line of the subway security area in the patrol video, and draw the boundary line of the subway security area in the patrol video; wherein the calculation expression of Pd is: ; In the formula, Indicates the actual distance between the subway security area boundary and the subway track centerline; Indicates the width of the patrol video image; Indicates the height of the patrol video image; Indicates the actual distance of the diagonal line in the patrol video image. The calculation expression is: ; In the formula, Indicates the height of the drone relative to the ground; Indicates the field of view of the drone, that is, the maximum angle range that the drone can shoot. The calculation expression is: , where Indicates the size of the image sensor of the drone’s camera; Indicates the focal length of the lens.

[0011] Furthermore, in step S40, before using the target recognition model to detect risky construction equipment in the subway security area within the boundary line, the trained target recognition model needs to be deployed on the drone, wherein the specific training method of the target recognition model is: S401 uses drones to collect construction equipment data in various construction scenes at different time periods every day, including various overhead shooting heights and angles, to build an aerial perspective dataset mainly consisting of excavators, pile drivers, drilling machines, and cranes; S402, designing a target recognition model based on construction equipment in the subway security area by improving the target detection model yolov11; S403, loading the data set through the target recognition model designed in S402 for multiple rounds of training, and performing performance evaluation on the target recognition model after each round of training, and selecting the optimal model based on the evaluation results to deploy on the UAV.

[0012] Furthermore, in step S402, the specific improvement method of the target detection model yolov11 is: enhancing data through geometric transformation and color space transformation at the input end; in the backbone network, replacing the conventional convolution kernel in the C3K2 module with the adaptive rotation convolution kernel ARconv2d to extract target features of different directions in the image to output feature maps of different directions; in the detection head part, a rotation angle path is added to each detection head to extract features related to the rotation angle.

[0013] Furthermore, in step S50, the obtaining of the GPS information of the screenshot and combining the screenshot, the GPS information of the screenshot and the business data to generate warning information and send it to the staff for processing specifically includes the following steps: Determine the longitude, latitude and altitude of the drone in the screenshot from the GPS information data collected by the drone using the screenshot timestamp; Design a data structure to store screenshots, GPS information of screenshots, and business data, and convert the data structure into JSON format to generate warning information; The warning information is reported to the remote control platform of the drone through the network communication protocol, and the remote control platform of the drone pushes the warning information to the terminal device of the staff for processing.

[0014] An intelligent early warning system for unmanned aerial vehicle inspections in subway security area construction is applied to the above-mentioned intelligent early warning method for unmanned aerial vehicle inspections in subway security area construction, and the improvement thereof is that the intelligent early warning system for unmanned aerial vehicle inspections in subway security area construction comprises: The flight mission planning module is used to set the inspection route and flight parameters of the drone; Video acquisition module, used to collect patrol videos in real time along the route; The subway security area identification and marking module is used to determine the subway security area and draw the boundary line of the subway security area; The risky construction equipment identification and detection module is used to detect risky construction equipment within the subway security area in the patrol video using the target recognition model; The risky construction equipment extraction module is used to take screenshots of the locations of risky construction equipment in the inspection video and extract the GPS information of the screenshots; The early warning module is used to generate early warning information by combining screenshots, GPS information of screenshots and business data and send it to staff for processing.

[0015] The beneficial effects of the present invention are as follows: the present scheme firstly utilizes the unmanned aerial vehicle to collect inspection videos along the planned flight path according to the planned flight parameters, so as to ensure that the viewing angle of the inspection video shot by the unmanned aerial vehicle can completely cover the subway security areas on both sides of the subway track area and can meet the inspection accuracy; then, by determining the position of the subway security area in the inspection video and drawing the boundary line of the subway security area, the subway security area can be accurately located in the inspection video; then, the target recognition model is used to detect the risk construction equipment in the subway security area within the boundary line, and the inspection video is screenshotted according to the detected risk construction equipment, so as to improve the detection efficiency of the risk construction equipment; finally, by obtaining the GPS information of the screenshot, and combining the screenshot, the GPS information of the screenshot and the business data to generate an early warning information and send it to the staff for processing, the staff can be helped to quickly obtain the location information of the risk construction equipment for timely processing; therefore, the present invention can not only conduct a comprehensive inspection of the subway security area, but also improve the probability of discovering the risk construction equipment, and can also quickly obtain the location information of the risk construction equipment and send it to the staff in time, so as to provide a solid and reliable guarantee for the effective supervision of the subway security area. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of an intelligent early warning method for unmanned aerial vehicle inspection in subway security area construction according to the present invention; Figure 2 It is a block diagram of an intelligent early warning system for UAV inspection in subway security area construction according to the present invention; Figure 3 A hardware structure diagram of an electronic device is shown in an exemplary embodiment; Figure 4 A block diagram of an electronic device is shown in accordance with an exemplary embodiment; Figure 5 Added rotation angle path diagram for the detection head; Figure 6 This is the structural diagram of the improved asymmetric detection head. DETAILED DESCRIPTION

[0017] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0018] The following will clearly and completely describe the concept, specific structure and technical effects of the present invention in combination with the embodiments and drawings, so as to fully understand the purpose, characteristics and effects of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, other embodiments obtained by technicians in this field without creative work are all within the scope of protection of the present invention. In addition, all the connection / connection relationships involved in the patent do not refer to the direct connection of components, but refer to the formation of a better connection structure by adding or reducing connection accessories according to the specific implementation situation. The various technical features in the invention can be combined interchangeably without conflicting with each other.

[0019] Reference Figure 1 As shown, the present invention discloses an intelligent early warning method for unmanned aerial vehicle inspection of subway security area construction, and the unmanned aerial vehicle inspection method for subway security area construction includes the following steps: S10, planning the inspection route and flight parameters of the drone; in this embodiment, during the specific application of the drone, the drone's remote control platform sets the drone's inspection route and the drone's flight parameters. After the setting is completed, the remote control platform controls the drone to perform the inspection task; specifically, the specific method of planning the drone's inspection route is: Obtain an accurate map of the subway line to be inspected, including the route, station locations, and surrounding terrain; Determine the scope of the subway security area at different sections of the subway line: set the boundary line of the subway security area L meters outside the center line of the subway track, where the value range of L is between 30-70; Based on the subway line to be inspected, the drone's flight path is set just above the center line of the subway track; The flight parameters of the drone include flight altitude and flight speed. The flight altitude of the drone is set between 70m-100m to ensure that the drone video can fully cover the subway security area on both sides of the subway line; the flight speed of the drone is set between 5m / s-10m / s to meet the inspection accuracy; It should be noted that in the process of patrol route planning, the accurate map of the subway line to be patrolled is the basis for planning the drone patrol route. The existence of map data enables the remote control platform to accurately understand the actual layout of the subway line, help formulate the most appropriate patrol route, and provide necessary geographic information support for subsequent route planning, ensuring that the drone patrol mission can cover all areas that need to be inspected and reduce omissions; the determination of the scope of the subway security area is to ensure that the patrol drone can cover the security areas on both sides of the subway line, which is helpful to discover potential safety hazards. In addition, due to the different geographical locations of different sections of the subway line, the actual distance L between the subway track center and the subway security area needs to be set after field investigation of the geographical environment of the subway line. Adaptive adjustment is made; the flight path of the drone is set directly above the center line of the subway track, which means that the inspection path of the drone will fly vertically above the subway line. This setting ensures the symmetry of the drone's flight path, allowing it to fully cover the subway line and the subway security areas on both sides, avoiding missing any possible monitoring blind spots; in the process of planning the drone's inspection route, the flight altitude of the drone is set between 70m-100m, so that the drone can ensure a suitable observation range and clear vision during the inspection mission, and effectively avoid collisions with obstacles; the flight speed of the drone is generally maintained between 5m / s-10m / s, and the flight speed can be appropriately reduced in areas that need to be observed in order to conduct more detailed inspections; S20, using the drone to collect patrol videos along the set route; in addition, the drone needs to collect GPS data of the drone in real time during the process of collecting patrol videos, and the GPS data includes longitude, latitude, altitude and timestamp; It should be noted that, in this embodiment, the drone is equipped with a high-definition camera. In the process of performing the inspection task, the drone collects the inspection video in a vertical overhead shooting manner, and according to the set flight altitude, the picture taken by the inspection video often exceeds the actual subway security area, so as to completely cover the subway security area on both sides of the track without leaving blind spots, thereby ensuring a full range of inspections of the subway security area; in addition, since the drone itself is equipped with a GPS module, in the process of the drone collecting the inspection video, the drone will collect its own GPS information in real time through the GPS module, and the GPS information includes longitude, latitude, altitude and timestamp; the collection of GPS information can record the navigation track and position of the drone in real time, and combined with the collected video data, it can accurately provide the video with geographical location information, so as to provide assistance for the subsequent analysis and positioning of the video data; S30, determining the position of the subway security area in the patrol video and drawing the boundary line of the subway security area; in this embodiment, since the collected patrol video screen exceeds the actual subway security area, it is necessary to exclude the area outside the subway security area before detecting the risky construction equipment to prevent false detection; therefore, this embodiment determines the scope of the subway security area in the patrol video screen by drawing the security area boundary line in the patrol video image to ensure the accuracy of subsequent risky construction equipment detection; specifically, the specific method of determining the position of the subway security area in the patrol video and drawing the boundary line of the subway security area is: S301, drawing a virtual straight line perpendicular to the horizontal direction and passing through the center point of the image in the inspection video as a center line, wherein the center point of the image is a projection of the position of the drone; S302, according to the distance Pd between the boundary line of the subway security area and the center line of the image, determine the position of the boundary line of the subway security area in the patrol video, and draw the boundary line of the subway security area in the patrol video; wherein the calculation expression of Pd is: ; In the formula, Indicates the actual distance between the subway security area boundary and the subway track centerline; Indicates the width of the patrol video image; Indicates the height of the patrol video image; Indicates the actual distance of the diagonal line in the patrol video image. The calculation expression is: ; In the formula, Indicates the height of the drone relative to the ground; Indicates the field of view of the drone, that is, the maximum angle range that the drone can shoot. The calculation expression is: , where Indicates the size of the image sensor of the drone’s camera; Indicates the focal length of the lens; S40, using the target recognition model to detect risky construction equipment in the subway security area within the boundary line, and taking screenshots of the inspection video according to the detected risky construction equipment; in this embodiment, the subway security area in the inspection video is analyzed by the target recognition model to identify the construction equipment in the subway security area that may pose a threat to the safety of the subway structure; therefore, before using the target recognition model to detect risky construction equipment in the subway security area within the boundary line, it is necessary to first deploy the trained target recognition model to the drone, wherein the specific training method of the target recognition model is: S401, using drones to collect construction equipment data in various construction scenes at different time periods every day, including various overhead shooting heights and angles, to build an aerial perspective data set mainly composed of excavators, pile drivers, drilling machines and cranes; among them, excavators, pile drivers, drilling machines and cranes are the key construction equipment that subway security areas pay the most attention to; S402, designing a target recognition model based on construction equipment in the subway security area by improving the target detection model yolov11; the specific improvement method of the target detection model yolov11 is: The data is enhanced through geometric transformation and color space transformation at the input end. Specifically, the improvement of the input end increases the probability of mosaic mixing from 50% to 70%, increasing the frequency of small objects. The probability of rotation enhancement is increased from 15% to 50% to ensure that the model can adapt to objects at different angles. In the backbone network, the adaptive rotation convolution kernel ARconv2d is used to replace the conventional convolution kernel in the C3K2 module to extract target features of different directions in the image, so as to output feature maps of different directions; specifically, during training, for each conventional convolution , a rotation angle will be predicted , first, according to the predicted rotation angle Construct a rotation matrix , then, through the affine network to generate the function right Operate to generate a sampling grid after affine transformation , and then according to the rotated network coordinates, Liu Yong's network sampling function For the original convolution Sampling is performed to obtain the rotation convolution kernel , the calculation process is as follows: , ; ;Finally, the rotated convolution kernel is used to convolve with the input feature map to obtain the output feature map; In the detection head part, a rotation angle path is added to each detection head to extract features related to the rotation angle. Specifically, for feature maps of different scales output by the backbone network, yolov11 is equipped with detection heads of corresponding scales. Each detection head has a classification path and a regression path. On this basis, a rotation angle path is added to each detection head, which is specifically used to extract features related to the rotation angle. It should be noted that, in this embodiment, referring to Figure 5-Figure 6 As shown, Figure 5is the path diagram of the rotation angle. In the figure, ARConv represents the combination of ARconv2d+bn+silu, where ARconv2d represents a convolution layer with a rotation convolution kernel, bn represents a batch normalization layer, and silu represents an activation function; conv2d is a conventional convolution layer, where the channel is set to 1, which is used to output the predicted angle; these detection heads are usually composed of multiple convolution layers and adopt a symmetrical structure, that is, the detection head structure of each scale is the same. Since the drone is far away from the shooting target, it is necessary to focus on the detection head corresponding to the feature map P3 (sampling 8 times) for small targets. In order to improve the feature extraction capability, this embodiment adopts Figure 6 The design of the asymmetric structure shown in the figure keeps the classification path of the detection head corresponding to the feature map P3 unchanged, and makes the following improvements to its regression path: adding a CA (Coordinate Attention) attention module between the two Conv layers to enhance the small target detection capability. Specifically, the CA (Coordinate Attention) attention module first performs global pooling in the horizontal and vertical directions on the feature map p (the feature map p is the intermediate output result of the feature map P3 regression path) output by the first layer Conv of the feature map P3 regression path, and generates feature vectors in these two directions. and , then concatenate the features in these two directions, and perform convolution on the concatenated features to generate intermediate features ; Separate the intermediate features into horizontal features and vertical features , then, yes and Perform convolution operations separately to generate attention weights and ; The generated attention weights , The output feature map of the CA layer is obtained by element-by-element multiplication with the input feature map p. In addition, in order to reduce the computational complexity caused by the introduction of the CA layer, the Ghost operation is used in the second Conv in the regression path of the feature map P3 for lightweight improvement. First, a part of the feature map is generated by conventional convolution, and then a redundant feature map is generated by a cheap operation (linear transformation). Then, these two types of feature maps are spliced ​​to keep the same number of output channels as the original and input into the subsequent network. For the detection heads of the feature maps P4 and P5, except for adding the rotation angle path, the other paths keep the original structure unchanged. S403, loading the data set through the target recognition model designed in S402 for multiple rounds of training, and evaluating the performance of the target recognition model after each round of training, and selecting the optimal model to be deployed on the drone according to the evaluation results; it should be noted that during the training process, in order to speed up the convergence speed, a pre-trained model based on public data of the drone can be used for initialization, and hyperparameters such as the learning rate, batch size, and number of training rounds can be set at the same time; in addition, the target recognition model reads data from the training set, calculates the loss value through forward propagation, and uses back propagation to update the model parameters, and after each training round, the validation set is used to evaluate the model performance; S50, obtaining the GPS information of the screenshot, and combining the screenshot, the GPS information of the screenshot, and the business data to generate warning information and send it to the staff for processing; specifically, obtaining the GPS information of the screenshot, and combining the screenshot, the GPS information of the screenshot, and the business data to generate warning information and send it to the staff for processing specifically includes the following steps: Determine the longitude, latitude and altitude of the drone in the screenshot from the GPS information data collected by the drone using the screenshot timestamp; Design a data structure to store screenshots, GPS information of the screenshots, and business data, and convert the data structure into JSON format to generate warning information; wherein the business data includes data related to the identified risky construction equipment, such as equipment ID and event description; The warning information is reported to the remote control platform of the drone through the network communication protocol, and the remote control platform of the drone pushes the warning information to the terminal device of the staff for processing.

[0020] It should be noted that, in this embodiment, by packaging the screenshot, the GPS information of the screenshot, and the business data together to generate early warning information and sending it to the staff, it can help the staff quickly confirm the equipment information and location information of the risky construction equipment and deal with it in a timely manner, so as to provide solid and reliable guarantees for the effective supervision of the subway security area.

[0021] Reference Figure 2 As shown, the present invention also discloses an intelligent early warning system 600 for unmanned aerial vehicle inspection for subway security area construction, which is applied to an intelligent early warning method for unmanned aerial vehicle inspection for subway security area construction as described in the above embodiment. The intelligent early warning system 600 for unmanned aerial vehicle inspection for subway security area construction includes: The flight mission planning module 601 is used to set the inspection route and flight parameters of the UAV; The video acquisition module 602 is used to collect patrol videos in real time along the route path; The subway security area identification and marking module 603 is used to determine the subway security area and draw the boundary line of the subway security area; The risky construction equipment identification and detection module 604 is used to detect risky construction equipment within the subway security area in the patrol video using the target recognition model; The risky construction equipment extraction module 605 is used to take a screenshot of the location of the risky construction equipment in the inspection video and extract the GPS information of the screenshot; The warning module 606 is used to generate warning information by combining the screenshot, the GPS information of the screenshot and the business data and send it to the staff for processing.

[0022] It should be noted that the intelligent early warning strategy for drone inspection in subway security area construction provided in the above embodiment is only illustrated by the division of the above functional modules. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the intelligent early warning system 600 for drone inspection in subway security area construction will be divided into different functional modules to complete all or part of the functions described above. In addition, the drone inspection intelligent early warning system 600 for subway security area construction provided in the above-mentioned embodiment and the embodiment of the drone inspection intelligent early warning method for subway security area construction belong to the same concept, and the specific way in which each module performs the operation has been described in detail in the method embodiment and will not be repeated here.

[0023] Figure 3 The present invention schematically illustrates the structure of an electronic device according to an exemplary embodiment.

[0024] It should be noted that the electronic device is only an example adapted to the present invention and cannot be considered to provide any limitation on the scope of use of the present invention. The electronic device cannot be interpreted as requiring dependence on or having Figure 3 One or more components of exemplary electronic device 2000 are shown.

[0025] The hardware structure of the electronic device 2000 may vary greatly due to differences in configuration or performance, such as Figure 3 As shown, the electronic device 2000 includes: a power supply 210 , an interface 230 , at least one memory 250 , and at least one central processing unit (CPU) 270 .

[0026] Specifically, the power supply 210 is used to provide operating voltage for each hardware device on the electronic device 2000 .

[0027] The interface 230 includes at least one wired or wireless network interface 231 for interacting with external devices. Of course, in other examples adapted by the present invention, the interface 230 may further include at least one serial-to-parallel conversion interface 233, at least one input-output interface 235, and at least one USB interface 237, etc. Figure 3 As shown, this is not intended to be a specific limitation.

[0028] The memory 250 is a carrier for storing resources, which may be a read-only memory, a random access memory, a disk or an optical disk, etc. The resources stored thereon include an operating system 251, an application 253 and data 255, etc. The storage method may be temporary storage or permanent storage.

[0029] Among them, the operating system 251 is used to manage and control various hardware devices and application programs 253 on the electronic device 2000 to enable the central processor 270 to calculate and process the massive data 255 in the memory 250. It can be WindowsServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

[0030] The application program 253 is a computer-readable instruction that performs at least one specific task based on the operating system 251, and may include at least one module ( Figure 3 (not shown), each module may respectively include computer-readable instructions for the electronic device 2000. For example, the device for assessing task priority may be regarded as an application 253 deployed on the electronic device 2000.

[0031] The data 255 may be signal information, etc., stored in the memory 250 .

[0032] The central processor 270 may include one or more processors, and is configured to communicate with the memory 250 through at least one communication bus to read the computer-readable instructions stored in the memory 250, thereby realizing the operation and processing of the mass data 255 in the memory 250. For example, the intelligent early warning method for drone inspection of subway security area construction is completed by the central processor 270 reading a series of computer-readable instructions stored in the memory 250.

[0033] In addition, the present invention can also be implemented by hardware circuits or hardware circuits combined with software. Therefore, the implementation of the present invention is not limited to any specific hardware circuits, software, or a combination of the two.

[0034] See also Figure 4 In an embodiment of the present invention, an electronic device 4000 is provided. The electronic device 4000 may include: a desktop computer, a laptop computer, a server, etc. with sensor recognition capability.

[0035] exist Figure 4 In the embodiment, the electronic device 4000 includes at least one processor 4001 and at least one memory 4003.

[0036] The data interaction between the processor 4001 and the memory 4003 can be realized through at least one communication bus 4002. The communication bus 4002 may include a path for transmitting data between the processor 4001 and the memory 4003. The communication bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The communication bus 4002 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0037] Optionally, the electronic device 4000 may further include a transceiver 4004, which may be used for data interaction between the electronic device and other electronic devices, such as data transmission and / or data reception, etc. It should be noted that in actual applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present invention.

[0038] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of the present invention. Processor 4001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0039] The memory 4003 can be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program instructions or codes in the form of instructions or data structures and can be accessed by the electronic device 4000, but is not limited to this.

[0040] Computer-readable instructions are stored in the memory 4003 , and the processor 4001 can read the computer-readable instructions stored in the memory 4003 through the communication bus 4002 .

[0041] The computer-readable instructions are executed by one or more processors 4001 to implement the intelligent early warning method for drone inspection of subway security area construction in the above-mentioned embodiments.

[0042] In addition, a storage medium is provided in an embodiment of the present invention, on which computer-readable instructions are stored, and the computer-readable instructions are executed by one or more processors to implement the above-mentioned drone inspection intelligent early warning method for subway security area construction.

[0043] Compared with the prior art, the present invention has the following beneficial effects: first, by using a drone to collect inspection videos along a planned flight path according to planned flight parameters, it can be ensured that the viewing angle of the inspection video shot by the drone can completely cover the subway security areas on both sides of the subway track area and can meet the inspection accuracy; then, by determining the position of the subway security area in the inspection video and drawing the boundary line of the subway security area, the subway security area can be accurately located in the inspection video; then, the target recognition model is used to detect risky construction equipment in the subway security area within the boundary line, and the inspection video is screenshotted according to the detected risky construction equipment, so as to improve the detection efficiency of the risky construction equipment; finally, by obtaining the GPS information of the screenshot, and combining the screenshot, the GPS information of the screenshot and the business data to generate an early warning information and send it to the staff for processing, it can help the staff quickly obtain the location information of the risky construction equipment for timely processing; therefore, the present invention can not only conduct a comprehensive inspection of the subway security area, but also improve the probability of discovering risky construction equipment, and can also quickly obtain the location information of the risky construction equipment and send it to the staff in time, so as to provide a solid and reliable guarantee for the effective supervision of the subway security area.

[0044] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. An intelligent early warning method for drone inspection in subway security area construction, characterized in that: The drone inspection method for subway security area construction includes the following steps: S10, planning the inspection route and flight parameters of the drone; S20, using drones to collect patrol videos along a set route; S30, determining the location of the subway security area in the patrol video and drawing a boundary line of the subway security area; S40, using the target recognition model to detect risky construction equipment in the subway security area within the boundary line, and taking screenshots of the inspection video according to the detected risky construction equipment; S50, obtaining the GPS information of the screenshot, and combining the screenshot, the GPS information of the screenshot and the business data to generate warning information and send it to the staff for processing.

2. According to claim 1, a drone inspection intelligent early warning method for subway security area construction is characterized in that: In step S10, the specific method of planning the inspection route of the drone is: Obtain an accurate map of the subway line to be inspected, including the route, station locations, and surrounding terrain; Determine the scope of the subway security area at different sections of the subway line: set the boundary line of the subway security area L meters outside the center line of the subway track, where the value range of L is between 30-70; Based on the subway line to be inspected, the drone's flight path is set just above the center line of the subway track.

3. The intelligent early warning method for UAV inspection in subway security area construction according to claim 2 is characterized in that: In step 10, the flight parameters of the drone include flight altitude and flight speed, wherein the flight altitude of the drone is set between 70m-100m to ensure that the video shot by the drone can completely cover the subway security area on both sides of the subway line; the flight speed of the drone is set between 5m / s-10m / s to meet the inspection accuracy.

4. The intelligent early warning method for UAV inspection in subway security area construction according to claim 3 is characterized in that: In step S20, while the drone is collecting patrol videos, it is also necessary to collect GPS data of the drone in real time, where the GPS data includes longitude, latitude, altitude and timestamp.

5. The intelligent early warning method for unmanned aerial vehicle inspection in subway security area construction according to claim 4 is characterized in that: In step S30, the specific method of determining the location of the subway security area in the patrol video and drawing the boundary line of the subway security area is: S301, drawing a virtual straight line perpendicular to the horizontal direction and passing through the center point of the image in the inspection video as a center line, wherein the center point of the image is a projection of the location of the drone; S302, according to the distance Pd between the boundary line of the subway security area and the center line of the image, determine the position of the boundary line of the subway security area in the patrol video, and draw the boundary line of the subway security area in the patrol video; wherein the calculation expression of Pd is: ; In the formula, Indicates the actual distance between the subway security area boundary and the subway track centerline; Indicates the width of the patrol video image; Indicates the height of the patrol video image; Indicates the actual distance of the diagonal line in the patrol video image. The calculation expression is: ; In the formula, Indicates the height of the drone relative to the ground; Indicates the field of view of the drone, that is, the maximum angle range that the drone can shoot. The calculation expression is: , where Indicates the size of the image sensor of the drone’s camera; Indicates the focal length of the lens.

6. The intelligent early warning method for UAV inspection in subway security area construction according to claim 5 is characterized in that: In step S40, before using the target recognition model to detect risky construction equipment in the subway security area within the boundary line, it is necessary to first deploy the trained target recognition model to the drone, wherein the specific training method of the target recognition model is: S401 uses drones to collect construction equipment data in various construction scenes at different time periods every day, including various overhead shooting heights and angles, to build an aerial perspective dataset mainly consisting of excavators, pile drivers, drilling machines, and cranes; S402, designing a target recognition model based on construction equipment in the subway security area by improving the target detection model yolov11; S403, loading the data set through the target recognition model designed in S402 for multiple rounds of training, and performing performance evaluation on the target recognition model after each round of training, and selecting the optimal model based on the evaluation results to deploy on the UAV.

7. The intelligent early warning method for UAV inspection in subway security area construction according to claim 6 is characterized in that: In step S402, the specific improvement method of the target detection model yolov11 is: enhancing data through geometric transformation and color space transformation at the input end; in the backbone network, replacing the conventional convolution kernel in the C3K2 module with the adaptive rotation convolution kernel ARconv2d to extract target features of different directions in the image to output feature maps of different directions; in the detection head part, a rotation angle path is added to each detection head to extract features related to the rotation angle.

8. The intelligent early warning method for UAV inspection in subway security area construction according to claim 7 is characterized in that: In step S50, the process of obtaining the GPS information of the screenshot and generating warning information by combining the screenshot, the GPS information of the screenshot and the business data and sending the warning information to the staff for processing specifically includes the following steps: Determine the longitude, latitude and altitude of the drone in the screenshot from the GPS information data collected by the drone using the screenshot timestamp; Design a data structure to store screenshots, GPS information of screenshots, and business data, and convert the data structure into JSON format to generate warning information; The warning information is reported to the remote control platform of the drone through the network communication protocol, and the remote control platform of the drone pushes the warning information to the terminal device of the staff for processing.

9. An intelligent early warning system for unmanned aerial vehicle inspections in subway security area construction, applied to an intelligent early warning method for unmanned aerial vehicle inspections in subway security area construction as described in any one of claims 1 to 8, characterized in that: The UAV inspection intelligent early warning system for subway security area construction includes: The flight mission planning module is used to set the inspection route and flight parameters of the drone; Video acquisition module, used to collect patrol videos in real time along the route; The subway security area identification and marking module is used to determine the subway security area and draw the boundary line of the subway security area; The risky construction equipment identification and detection module is used to detect risky construction equipment within the subway security area in the patrol video using the target recognition model; The risky construction equipment extraction module is used to take screenshots of the locations of risky construction equipment in the inspection video and extract the GPS information of the screenshots; The early warning module is used to generate early warning information by combining screenshots, GPS information of screenshots and business data and send it to staff for processing.

Images