Highway intelligent accident detection processing method and system based on unmanned aerial vehicle

Abstract

The invention provides an intelligent highway accident detection processing method and system based on an unmanned aerial vehicle, belongs to the technical field of highway accident detection processing, and is used for solving the problem that the technology of carrying out accident detection processing by using the unmanned aerial vehicle in the related technology is not beneficial to rapid determination of a processing strategy and timely rescue. Acquiring lane occupation data and severity data of the accident based on the unmanned aerial vehicle; determining traffic influence data of the accident based on monitoring video information of the lane where the accident is located within a first preset duration before the accident occurrence moment; and determining accident grade data according to the lane occupation data, the severity data, the traffic influence data and pre-acquired lane type data. By adopting the technical scheme, the grade of the accident on the expressway can be intelligently determined by using the unmanned aerial vehicle, so that a processing strategy can be quickly determined, and rescue can be performed in time.

Description

Technical Field

[0001] The present application relates to the technical field of highway accident detection and processing, and in particular to a method and system for highway intelligent accident detection and processing based on unmanned aerial vehicles. Background Art

[0002] With the development of science and technology, drone technology has become more and more advanced. At present, drones are widely used in various industries. In terms of transportation, drones are used to intelligently detect traffic accidents on highways. Generally speaking, drones will conduct daily patrols along the established inspection routes within the scope of the highway, and use the onboard shooting equipment to record and detect the traffic conditions on the highway, so as to cooperate with the traffic behavior analysis algorithm in the monitoring video to quickly determine the occurrence of the accident and reach the accident point. After the accident occurs, it is generally necessary to report and handle the accident. The accident information reporting based on drone technology is generally difficult to classify the accident information like the accident information reporting by the relevant personnel on the scene. The drone generally uploads the video image of the accident scene, and the relevant personnel manually determine the accident level based on the video image and other information related to the accident, and then determine the handling strategy. However, the efficiency of determining the handling strategy in this way is low and not conducive to timely rescue. Summary of the invention

[0003] The present application provides a method and system for intelligent highway accident detection and processing based on drones, which can intelligently determine the accident level so as to quickly determine the processing strategy and provide timely rescue.

[0004] In a first aspect, the present application provides a method for intelligent highway accident detection and processing based on drones. The method comprises: Acquiring lane occupancy data and severity data of the accident based on the drone, wherein the lane occupancy data is positively correlated with the number and width of lanes affected by the accident; Determine the traffic impact data of the accident based on the surveillance video information of the lane where the accident occurred within a first preset time period before the accident occurred, wherein the traffic impact data is associated with the vehicle volume data and vehicle speed data of each passing vehicle within the first preset time period; Accident level data is determined based on the lane occupancy data, severity data, traffic impact data, and pre-acquired lane type data.

[0005] By adopting the above technical solution, it is possible to use drones to intelligently determine the level of accidents on highways, so as to quickly determine handling strategies and provide timely rescue.

[0006] Furthermore, the lane occupancy data and severity data of the accident obtained by the drone include: Obtain images of the accident scene collected by drones; Analyze the accident scene image to obtain lane condition data and traffic accident condition data, wherein the lane condition data reflects the distribution position and lane width of each lane, and the traffic accident condition data reflects the number, damage degree and distribution position of the accident vehicles; A bird's-eye view of the accident scene is constructed based on the accident scene image, and each lane feature is marked in the bird's-eye view of the accident scene, each accident vehicle is marked with a first rectangular feature, and all accident vehicles are marked with a second rectangular feature, and the number of lanes occupied by the smallest second rectangular feature and the lane width of each occupied lane are determined, and the lane occupancy data is positively correlated to the number of lanes occupied by the smallest second rectangular feature and the lane width of each occupied lane.

[0007] Furthermore, the method for acquiring lane occupancy data includes: Assume that the number of lanes occupied by the smallest second rectangular feature is , the total number of lanes on the road is , the lane width of the i-th lane is , the lane occupancy data is ,but , where is a preset constant greater than zero.

[0008] Further, obtaining images of the accident scene collected by the drone; Analyze the accident scene image to obtain lane condition data and traffic accident condition data, wherein the lane condition data reflects the distribution position and lane width of each lane, and the traffic accident condition data reflects the number, damage degree and distribution position of the accident vehicles; Assume there are n accident vehicles. , the damage degree data of the i-th accident vehicle is , the damage coefficient of the i-th accident vehicle is , the damage coefficient is positively correlated to the volume of the vehicle, determining For the accident vehicles, let the relative distance data between the jth pair of accident vehicles be , the severity data is ,but , where , , All are preset constants greater than zero.

[0009] Furthermore, the traffic impact data of the accident is determined based on the surveillance video information of the lane where the accident occurred within a first preset time period before the accident occurred, including: Obtaining surveillance video information of the lane where the accident occurred within a first preset time period before the time of the accident; Analyze the surveillance video information to determine the vehicle volume data, vehicle speed data and passing time data of each passing vehicle; The first preset duration is divided into a plurality of second preset durations, and a duration flux data is calculated relative to each second preset duration; assuming that the second preset duration has , within the i-th second preset duration, there are The larger i is, the farther the i-th second preset time is from the time of the accident. The vehicle volume data of the j-th passing vehicle in the j-th second preset time is , vehicle speed data is , the duration flux data of the i-th second preset duration is ,but ; Calculate the flux change data of each adjacent second preset time length, the flux change data has The i-th flux change data is ,but ; Determine whether they are all greater than zero or not greater than zero; If yes, then the traffic affects the data ,otherwise, , where The closest to the time of the accident The duration flux data of the second preset duration, , is greater than zero The number of is not greater than zero The number of , All are preset constants greater than zero.

[0010] Further, the determining of accident level data according to the lane occupancy data, severity data, traffic impact data and pre-acquired lane type data includes: Assume the accident level data is z and the lane occupancy data is , the severity data is , the traffic impact data is , the lane type data is ,but .

[0011] In a second aspect, the present application provides a highway intelligent accident detection and processing system based on a drone. The system includes a drone and a server, wherein the drone is used to execute any one of the methods described in the first aspect above, and the server is used to receive the accident level data.

[0012] In summary, this application at least includes the following technical effects: A method and system for intelligent highway accident detection and processing based on drones is provided, which can use drones to intelligently determine the level of accidents on highways, so as to quickly determine processing strategies and provide timely rescue.

[0013] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present application, nor are they intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein: Figure 1 A block diagram of a highway intelligent accident detection and processing system based on a drone in an embodiment of the present application is shown; Figure 2 A flowchart of a method for intelligent accident detection and processing on a highway based on a drone in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0016] In addition, the term "and / or" in this article is only a description of the association relationship between the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0017] The present application provides a method and system for intelligent highway accident detection and processing based on drones, which can use drones to intelligently determine the level of accidents on highways, so as to quickly determine processing strategies and provide timely rescue.

[0018] In a first aspect, the present application provides a highway intelligent accident detection and processing system based on a drone 110.

[0019] Figure 1 A block diagram of a highway intelligent accident detection and processing system based on a drone 110 in an embodiment of the present application is shown.

[0020] Reference Figure 1 The system includes a drone 110 and a server 120. There are generally multiple drones 110, but only one is shown in the figure. The drone 110 is connected to the server 120 for conducting inspections on the highway and collecting inspection image data of the highway. The drone 110 is equipped with an edge computing module for determining whether there is a traffic accident in the inspection image based on image feature analysis technology. The server 120 is connected to the drone 110 and is used to manage the inspection strategy of the drone 110. The server 120 is also connected to the monitoring system of the highway. The monitoring system of the highway includes monitoring cameras installed at each highway node. The monitoring cameras can be used to collect monitoring images of each highway node. By analyzing the monitoring images, the traffic behavior (i.e., traffic flow) of each highway node can be determined. Based on the traffic behavior analysis, it can be determined whether a traffic accident has occurred between the highway nodes. In this way, the monitoring system can be used to determine whether there is a traffic accident and the highway section where the traffic accident occurred. The drone 110 can also be used to inspect traffic accidents and collect inspection image data of the traffic accident scene (i.e., accident scene images). Combining the two is conducive to rapid discovery and using the drone 110 to arrive at the accident scene, collect images of the accident scene, and obtain accident scene images.

[0021] Of course, the server 120 can also be connected to the public service platform, and can also access third-party alarm information to determine the occurrence and location of the traffic accident.

[0022] The edge computing module carried by the drone 110 is also used to execute an accident level determination algorithm, that is, to determine accident level data based on the situation at the accident site, and report the accident level data. The accident level determination algorithm executed by the drone 110 is specifically introduced in conjunction with the disclosure of the second aspect below.

[0023] In a second aspect, the present application provides a method for intelligent highway accident detection and processing based on a drone 110.

[0024] Figure 2 FIG. 1 is a flowchart of a method for intelligent highway accident detection and processing based on drone 110 in an embodiment of the present application. The method can be Figure 1 The drone 110 is executed.

[0025] Reference Figure 2 , the method specifically comprises the following steps: S210: Acquire lane occupancy data and severity data of the accident based on the drone 110.

[0026] In the method of this step, the lane occupancy data is positively correlated with the number and width of lanes affected by the accident.

[0027] The method of this step specifically includes: obtaining the accident scene image collected by the drone 110; analyzing the accident scene image to obtain lane condition data and traffic accident condition data, wherein the lane condition data reflects the distribution position and lane width of each lane, and the traffic accident condition data reflects the number, degree of damage and distribution position of the accident vehicles. The accident scene image collected by the drone 110 reflects the situation of the accident scene in all directions and from multiple angles. By using the accident scene image and the existing image feature analysis technology, the specific lane condition data and vehicle condition data can be determined. The distribution position of the lane and the width of the lane can be specifically determined based on coordinate transformation calculation. The accident vehicle has obvious characteristics such as parking, rollover, deformation and even fire. Based on feature recognition technology, the accident characteristics can be determined, and combined with the accident characteristics that affect the vehicle itself, such as rollover, deformation, fire, etc., the degree of damage to the vehicle can be determined, and then the number and distribution position of the accident vehicles can be determined.

[0028] It should be understood that since most traffic accidents on highways are caused by collisions between two or more vehicles, the present application only addresses traffic accidents in this situation, and does not discuss other situations such as traffic accidents caused by a single vehicle colliding with highway barriers, tunnel walls, etc.

[0029] In the method of the present application, the specific method for obtaining lane occupancy data includes: constructing a bird's-eye view of the accident scene based on the accident scene image, and marking each lane feature in the bird's-eye view of the accident scene, marking each accident vehicle with a first rectangular feature, and marking all accident vehicles with a second rectangular feature, determining the number of lanes occupied by the smallest second rectangular feature and the lane width of each occupied lane, and the lane occupancy data is positively correlated with the number of lanes occupied by the smallest second rectangular feature and the lane width of each occupied lane.

[0030] In one example, let the number of lanes occupied by the smallest second rectangular feature be , the total number of lanes on the road is , the lane width of the i-th lane is , the lane occupancy data is ,but , where is a preset constant greater than zero.

[0031] In the method of the present application, the acquisition of lane occupancy data and severity data of the accident based on the drone 110 includes: assuming that there are n accident vehicles, , the damage degree data of the i-th accident vehicle is , the damage coefficient of the i-th accident vehicle is , the damage coefficient is positively correlated to the volume of the vehicle, determining For the accident vehicles, let the relative distance data between the jth pair of accident vehicles be , the severity data is ,but , where , , All are preset constants greater than zero.

[0032] Specifically, all relative distance data are included in the distribution position of the accident vehicle. Based on feature recognition technology, the distance between any two accident vehicles can be determined. The distance can specifically be the distance between the geometric centers of the first rectangular features corresponding to the two accident vehicles in the overhead view of the accident scene.

[0033] Regarding the determination of the damage coefficient, multiple damage dimension identifiers can be pre-constructed, such as the three damage dimensions of vehicle body deformation degree, whether it rolls over / overturns, and whether it catches fire. One type of damage dimension is an analog quantity, such as the vehicle body deformation degree, which is a deformation degree value determined by combining the deformation area of ​​the vehicle body, the deformation position, and the maximum deformation amount of each deformation position. One type is a digital quantity, such as whether it rolls over / overturns, whether it catches fire, and its status is only yes or no, indicating whether it exists or not. A damage coefficient comparison table can be pre-constructed to determine the damage coefficient result based on the parameter range of the analog quantity in the damage dimension identifier and the presence or absence of the digital quantity. Generally speaking, the greater the deformation degree, the greater the damage coefficient, overturning is greater than rollover is greater than no rollover / overturning, and fire is greater than no fire.

[0034] S220: Determine traffic impact data of the accident based on surveillance video information of the lane where the accident occurred within a first preset time period before the time of the accident.

[0035] In the method of this step, the traffic impact data is associated with the vehicle volume data and vehicle speed data of each passing vehicle within the first preset time period.

[0036] The method of this step specifically includes: obtaining surveillance video information of the lane where the accident occurred within a first preset time period before the time of the accident; Analyze the surveillance video information to determine the vehicle volume data, vehicle speed data and passing time data of each passing vehicle; The first preset duration is divided into a plurality of second preset durations, and a duration flux data is calculated relative to each second preset duration; assuming that the second preset duration has , within the i-th second preset duration, there are The larger i is, the farther the i-th second preset time is from the time of the accident. The vehicle volume data of the j-th passing vehicle in the j-th second preset time is , vehicle speed data is , the duration flux data of the i-th second preset duration is ,but ; Calculate the flux change data of each adjacent second preset time length, the flux change data has The i-th flux change data is ,but ; Determine whether they are all greater than zero or not greater than zero; If yes, then the traffic affects the data ,otherwise, , where The closest to the time of the accident The duration flux data of the second preset duration, , is greater than zero The number of is not greater than zero The number of , All are preset constants greater than zero.

[0037] Vehicle volume data and vehicle speed data can also be determined based on image feature recognition technology and are clear quantitative values. Travel time data can be determined based on a built-in clock.

[0038] It should be understood that since the surveillance video information is obtained by the monitoring system, the method of this step can also be requested by the drone 110, and the surveillance video information can be obtained by the server 120 and the calculation process of the traffic impact coefficient can be executed to save the computing power of the drone 110 and improve the computing efficiency.

[0039] S230: Determine accident level data according to the lane occupancy data, severity data, traffic impact data, and pre-acquired lane type data.

[0040] In the method of this step, the lane type data is positively correlated with the importance of the lane. The lane type data can be preset in the highway network map for specific lanes of specific roads, or a label system reflecting the importance can be adapted for each lane and its affiliated road, and the importance of the lane can be determined based on the importance reflected by the label marked under the label system.

[0041] The accident level data is positively correlated with the lane occupancy data, severity data, traffic impact data, and pre-acquired lane type data.

[0042] In a specific example, let the accident level data be z and the lane occupancy data be , the severity data is , the traffic impact data is , the lane type data is ,but .

[0043] Of course, based on the positive correlation logic, this formula can also be transformed into other forms, which will not be listed here. In short, the accident level data is the total quantitative result of the combination of the severity of the accident and the traffic impact. It can more objectively reflect the urgency and importance of the accident, so that relevant personnel can directly refer to the accident level data to determine the handling strategy of the traffic accident, which is conducive to timely rescue.

[0044] It should be understood that the image feature recognition technology, feature recognition technology, etc. mentioned in this method are all existing common knowledge, which are installed in the edge computing module of the drone 110 and will not be introduced in detail in this article.

[0045] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the embodiments of the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0046] In summary, this application at least includes the following technical effects: A method and system for intelligent highway accident detection and processing based on drone 110 is provided, which can use drone 110 to intelligently determine the level of accidents on the highway, so as to quickly determine the processing strategy and provide timely rescue.

[0047] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

Claims

1. A method for intelligent highway accident detection and processing based on drones, characterized in that: include: Acquiring lane occupancy data and severity data of the accident based on the drone, wherein the lane occupancy data is positively correlated with the number and width of lanes affected by the accident; Determine the traffic impact data of the accident based on the surveillance video information of the lane where the accident occurred within a first preset time period before the accident occurred, wherein the traffic impact data is associated with the vehicle volume data and vehicle speed data of each passing vehicle within the first preset time period; Accident level data is determined based on the lane occupancy data, severity data, traffic impact data, and pre-acquired lane type data.

2. The method for intelligent highway accident detection and processing based on drones according to claim 1 is characterized in that: The lane occupancy data and severity data of the accident obtained by the drone include: Obtain images of the accident scene collected by drones; Analyze the accident scene image to obtain lane condition data and traffic accident condition data, wherein the lane condition data reflects the distribution position and lane width of each lane, and the traffic accident condition data reflects the number, damage degree and distribution position of the accident vehicles; A bird's-eye view of the accident scene is constructed based on the accident scene image, and each lane feature is marked in the bird's-eye view of the accident scene, each accident vehicle is marked with a first rectangular feature, and all accident vehicles are marked with a second rectangular feature, and the number of lanes occupied by the smallest second rectangular feature and the lane width of each occupied lane are determined, and the lane occupancy data is positively correlated to the number of lanes occupied by the smallest second rectangular feature and the lane width of each occupied lane.

3. The method for intelligent highway accident detection and processing based on drones according to claim 2 is characterized in that: The method for acquiring lane occupancy data comprises: Assume that the number of lanes occupied by the smallest second rectangular feature is , the total number of lanes on the road is , the lane width of the i-th lane is , the lane occupancy data is ,but , where is a preset constant greater than zero.

4. The method for intelligent highway accident detection and processing based on drones according to claim 3 is characterized in that: The lane occupancy data and severity data of the accident obtained by the drone include: Obtain images of the accident scene collected by drones; Analyze the accident scene image to obtain lane condition data and traffic accident condition data, wherein the lane condition data reflects the distribution position and lane width of each lane, and the traffic accident condition data reflects the number, damage degree and distribution position of the accident vehicles; Assume there are n accident vehicles. , the damage degree data of the i-th accident vehicle is , the damage coefficient of the i-th accident vehicle is , the damage coefficient is positively correlated to the volume of the vehicle, determining For the accident vehicles, let the relative distance data between the jth pair of accident vehicles be , the severity data is ,but , where , , All are preset constants greater than zero.

5. The method for intelligent highway accident detection and processing based on drones according to claim 1 is characterized in that: Determining the traffic impact data of the accident based on the surveillance video information of the lane where the accident occurred within a first preset time period before the accident occurred includes: Obtaining surveillance video information of the lane where the accident occurred within a first preset time period before the time of the accident; Analyze the surveillance video information to determine the vehicle volume data, vehicle speed data and passing time data of each passing vehicle; The first preset duration is divided into a plurality of second preset durations, and a duration flux data is calculated relative to each second preset duration; assuming that the second preset duration has , within the i-th second preset duration, there are The larger i is, the farther the i-th second preset time is from the time of the accident. The vehicle volume data of the j-th passing vehicle in the j-th second preset time is , vehicle speed data is , the duration flux data of the i-th second preset duration is ,but ; Calculate the flux change data of each adjacent second preset time length, the flux change data has The i-th flux change data is ,but ; Determine whether they are all greater than zero or not greater than zero; If yes, then the traffic affects the data ,otherwise, , where The closest to the time of the accident The duration flux data of the second preset duration, , is greater than zero The number of is not greater than zero The number of , All are preset constants greater than zero.

6. The method for intelligent highway accident detection and processing based on drones according to claim 1 is characterized in that: Determining the accident level data according to the lane occupancy data, severity data, traffic impact data and pre-acquired lane type data includes: Assume the accident level data is z and the lane occupancy data is , the severity data is , the traffic impact data is , the lane type data is ,but .

7. An intelligent highway accident detection and processing system based on drones, characterized in that: The method comprises a drone (110) and a server (120), wherein the drone (110) is used to execute the method according to any one of claims 1 to 6, and the server (120) is used to receive the accident level data.

Images