CN-122022417-A - Unmanned aerial vehicle scheduling method and system for complex environment road detection

CN122022417ACN 122022417 ACN122022417 ACN 122022417ACN-122022417-A

Abstract

An unmanned aerial vehicle dispatching method and system for complex environment road detection belong to the technical field of unmanned aerial vehicle dispatching. The method aims to solve the problems of insufficient judgment basis and resource waste in the scheduling process of the complex environment road detection unmanned aerial vehicle. The method comprises the steps of constructing a terrain complexity scoring function, estimating the perception probability of a ground area, calculating the task entropy of unmanned aerial vehicle information according to the classification probability distribution of images, establishing the task scheduling value of the unmanned aerial vehicle intervention ground area, constructing the comprehensive scheduling cost function of the unmanned aerial vehicle, calculating the task cost performance scoring, normalizing the task cost performance scoring to obtain a scheduling value uniform metric value, calculating the occupation ratio of the scheduling value uniform metric value exceeding an activation threshold value in the vicinity of the statistical area, constructing an area connectivity factor, and constructing an unmanned aerial vehicle scheduling decision function based on the scheduling value uniform metric value and the area connectivity factor to obtain an unmanned aerial vehicle scheduling decision for complex environment road detection. The invention improves the service efficiency of the unmanned aerial vehicle.

Inventors

MENG ANXIN
CHENG GONG
LIU XING
ZHUANG WEIQUN
Ren Bangke
ZHAO HAIYUN
LI JUNYUAN

Assignees

深城交科技集团股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260415

Claims (10)

1. The unmanned aerial vehicle scheduling method for complex environment road detection is characterized by comprising the following steps of: S1, taking the topography gradient, shielding and relief degree of a complex environment road into consideration, and constructing a topography complexity scoring function; S2, estimating the perception probability of a ground area based on the equipment density and layout capacity of the ground sensor; s3, calculating information task entropy of the unmanned aerial vehicle according to the classification probability distribution of the images of the unmanned aerial vehicle based on the images acquired by the classification tasks, and evaluating the recognition difficulty of the images; S4, establishing task scheduling value of the unmanned aerial vehicle intervening ground area based on the terrain complexity score, the perception probability of the ground area and the unmanned aerial vehicle information task entropy; S5, evaluating the flight cost of the unmanned aerial vehicle to the ground area, and constructing a comprehensive scheduling cost function of the unmanned aerial vehicle; S6, carrying out exponential fusion on the task scheduling value obtained in the step S4 and the comprehensive scheduling cost obtained in the step S5, calculating task cost performance scores, and carrying out normalization processing on the task cost performance scores to obtain a scheduling value unified metric value; S7, based on the uniform measurement value of the scheduling value obtained in the step S6, counting the duty ratio of the uniform measurement value of the scheduling value in the neighborhood of the area exceeding an activation threshold value, and constructing an area connectivity factor; And S8, constructing an unmanned aerial vehicle scheduling decision function based on the uniform measurement value of the scheduling value obtained in the step S6 and the regional connectivity factor obtained in the step S7, and obtaining an unmanned aerial vehicle scheduling decision for complex environment road detection.
2. The unmanned aerial vehicle scheduling method for complex environment road detection according to claim 1, wherein the expression of the terrain complexity scoring function constructed in step S1 is: ; Wherein, the Scoring the complexity; The coordinates of the detection points; The regional elevation distribution is obtained by a digital elevation map; 、 Elevation gradients in the x-direction and the y-direction, respectively; for occlusion index, obtained by expert through experience or reference; the geomorphic frequency intensity is obtained by wavelet transformation; the gradient direction entropy is obtained by histogram entropy analysis; The adjustment factors are determined empirically by an expert and are used to adjust the dimensions.
3. The unmanned aerial vehicle scheduling method for complex environment road detection according to claim 2, wherein step S2 fuses various ground sensors according to layout density and confidence weight, and estimates the perception probability of the ground area 。
4. The unmanned aerial vehicle scheduling method for complex environment road detection according to claim 3, wherein the expression for calculating the unmanned aerial vehicle information task entropy in step S3 is: ; Wherein, the The unmanned aerial vehicle information task entropy is used; for the number of classification results, defining according to the target category by expert experience; The pixel duty ratio corresponding to the k class classification result is determined by the image segmentation network model; for a numerical stabilization increment, determined empirically by an expert.
5. The unmanned aerial vehicle scheduling method for complex environment road detection according to claim 4, wherein the expression for obtaining the task scheduling value of the unmanned aerial vehicle intervening in the ground area in step S4 is: ; Wherein, the The value of task scheduling for unmanned aerial vehicles to intervene in ground areas.
6. The unmanned aerial vehicle scheduling method for complex environment road detection according to claim 5, wherein step S5 constructs coupling nonlinear course cost according to flight distance and wind speed, introduces airspace congestion degree, flight time and energy consumption intensity, and establishes comprehensive scheduling cost 。
7. The unmanned aerial vehicle scheduling method for complex environment road detection according to claim 6, wherein the specific implementation method of step S6 comprises the following steps: S6.1, calculating a task cost performance score, wherein the expression is as follows: ; Wherein, the Scoring task cost performance; For the cost penalty rate, determined empirically by an expert; s6.2, normalizing the task cost performance scores, uniformly measuring all areas from 0 to 1 to obtain a uniform measurement value of the scheduling value, wherein the expression is as follows: ; Wherein, the Unifying the measurement values for the scheduling values; is a sharpness regulation parameter, and is determined by expert experience; For the maximum value of the task cost performance scores, expert based on Is determined by the statistical result of the (c).
8. The unmanned aerial vehicle scheduling method for complex environment road detection according to claim 7, wherein the expression of the regional connectivity factor constructed in step S7 is: ; Wherein, the Is a regional connectivity factor; for the neighborhood set, selecting 8 neighborhood or 16 neighborhood according to expert experience; for neighborhood sets Is a dot in (2); To activate the threshold, it is empirically determined by an expert.
9. The unmanned aerial vehicle scheduling method for complex environment road detection according to claim 8, wherein step S8 multiplies the scheduling value unified metric value by the regional connectivity factor, compares the multiplied value with a set threshold value, and outputs a boolean type judgment result, and the expression is: ; Wherein, the Whether to schedule a boolean outcome; to dispatch the decision threshold, it is determined by the expert based on actual demand.
10. A system of unmanned aerial vehicle scheduling methods for complex environment road detection, comprising a processor, a memory and a computer program stored in the memory and executable on the processor, the computer program when run implementing the steps of an unmanned aerial vehicle scheduling method for complex environment road detection as claimed in any one of claims 1 to 9.

Description

Unmanned aerial vehicle scheduling method and system for complex environment road detection Technical Field The invention belongs to the technical field of unmanned aerial vehicle dispatching, and particularly relates to an unmanned aerial vehicle dispatching method and system for complex environment road detection. Background The road detection can ensure traffic operation and urban management safety, the traditional method mainly relies on ground monitoring equipment and manual inspection, but in special environments such as mountain areas, tunnels and overhead, ground equipment is generally difficult to arrange, manual inspection efficiency is low, and coverage range is limited. With the increasing complexity of road networks, in order to improve the coverage rate and timeliness of monitoring, an effective method is formed by using unmanned aerial vehicles to carry out supplementary inspection. The unmanned aerial vehicle has the advantages of flexible course adjustment, rapid development, deployment, control, presentation and the like, and is suitable for high-frequency and high-precision image acquisition. However, in road detection, unmanned aerial vehicle scheduling has the following problems that repeated flight is easy to occur without scientific task priority evaluation standards, unnecessary resource waste is caused, and factors such as terrain conditions, perception deletion degree, image content complexity, flight cost and the like are not considered, so that a dispatching strategy is inaccurate and coverage to an area which needs to be supported most is difficult. Disclosure of Invention The invention aims to solve the problems of insufficient judgment basis and resource waste in the scheduling process of the unmanned aerial vehicle for complex environment road detection, and provides an unmanned aerial vehicle scheduling method and system for complex environment road detection. In order to achieve the above purpose, the present invention is realized by the following technical scheme: the unmanned aerial vehicle scheduling method for complex environment road detection comprises the following steps: S1, taking the topography gradient, shielding and relief degree of a complex environment road into consideration, and constructing a topography complexity scoring function; S2, estimating the perception probability of a ground area based on the equipment density and layout capacity of the ground sensor; s3, calculating information task entropy of the unmanned aerial vehicle according to the classification probability distribution of the images of the unmanned aerial vehicle based on the images acquired by the classification tasks, and evaluating the recognition difficulty of the images; S4, establishing task scheduling value of the unmanned aerial vehicle intervening ground area based on the terrain complexity score, the perception probability of the ground area and the unmanned aerial vehicle information task entropy; S5, evaluating the flight cost of the unmanned aerial vehicle to the ground area, and constructing a comprehensive scheduling cost function of the unmanned aerial vehicle; S6, carrying out exponential fusion on the task scheduling value obtained in the step S4 and the comprehensive scheduling cost obtained in the step S5, calculating task cost performance scores, and carrying out normalization processing on the task cost performance scores to obtain a scheduling value unified metric value; S7, based on the uniform measurement value of the scheduling value obtained in the step S6, counting the duty ratio of the uniform measurement value of the scheduling value in the neighborhood of the area exceeding an activation threshold value, and constructing an area connectivity factor; And S8, constructing an unmanned aerial vehicle scheduling decision function based on the uniform measurement value of the scheduling value obtained in the step S6 and the regional connectivity factor obtained in the step S7, and obtaining an unmanned aerial vehicle scheduling decision for complex environment road detection. Further, the expression of the terrain complexity scoring function constructed in step S1 is: Wherein, the Scoring the complexity; The coordinates of the detection points; The regional elevation distribution is obtained by a digital elevation map; 、 Elevation gradients in the x-direction and the y-direction, respectively; for occlusion index, obtained by expert through experience or reference; the geomorphic frequency intensity is obtained by wavelet transformation; the gradient direction entropy is obtained by histogram entropy analysis; The adjustment factors are determined empirically by an expert and are used to adjust the dimensions. Further, step S2 fuses various ground sensors according to the layout density and the confidence weight, and estimates the perception probability of the ground area。 Further, the expression for calculating the unmanned aerial vehicle information task e