Simplified environment modeling method for autonomous navigation of drone

Abstract

The invention provides a simplified environment modeling method for autonomous navigation of a drone. The method comprises the steps that a depth image is acquired, and the probability density distribution of each column of depth image data is statistically analyzed; the roughness of obstacles is analyzed according to the peak width, and the position of each column of obstacles on the depth imageis calculated; obstacles extracted from the depth image are mapped into a three-dimensional space and called obstacle vertical strips; the obstacle vertical strips are quickly spatially clustered withthe index number of the depth image; gaps between the obstacle vertical strips in a cluster are searched; if the gaps are larger than the reachable size of the drone, the cluster is divided accordingto the position of the gaps, and otherwise the gaps are filled; an obstacle contour plane is fitted according to the clustering result; and a simplified environment model consisting of multiple polygonal prism faces is constructed. According to the invention, the depth sensor data carried by the drone are transformed into the simplified environment model; irrelevant environmental details are eliminated; and services are provided for autonomous navigation of the drone and perceptual information sharing.

Description

technical field [0001] The invention belongs to the field of robot automation, and in particular relates to a simplified environment modeling method serving for autonomous navigation of unmanned aerial vehicles. Background technique [0002] UAV is the abbreviation of unmanned aircraft. In recent years, a wide range of application requirements has brought about the vigorous development of the UAV field. Its applications include search and rescue, intelligent transportation, 3D reconstruction and power line inspection. According to the degree of automation of drones, they can be divided into semi-autonomous drones and fully autonomous drones. The former requires operators to assist in performing tasks, while the latter relies entirely on the sensors carried by drones to sense the environment online and navigate and execute autonomously. decision making. The application of semi-autonomous drones is easily restricted by the environment and tasks. For more dangerous application...

AI Technical Summary

Problems solved by technology

This method depends on the size of the grid size, and the edge of the obstacle is easily affected by quantization, and the modeling accuracy is reduced

Images