Tree contour detection method and device based on unmanned aerial vehicle obstacle avoidance radar

Abstract

The invention discloses a tree contour detection method and device based on an unmanned aerial vehicle obstacle avoidance radar. The method comprises the steps: S1, acquiring an echo signal returned by an unmanned aerial vehicle obstacle avoidance radar, carrying out CFAR detection on the collected echo signal by using a first threshold value, and performing S2 under the condition of detecting a target to be a tree; S2, removing the currently detected target, obtaining the processed echo signal, and performing S3; and S3, performing CFAR detection on the processed echo signal by using a secondthreshold value obtained by reducing the first threshold value and determining the contour of the target tree according to the currently detected target under the condition of target detection. Therefore, the detection of the tree contour boundary can be realized based on the unmanned plane obstacle avoidance radar; and the implementation method is simple, the detection precision is high, and theobstacle avoidance effect is good.

Description

technical field [0001] The invention relates to the technical field of UAV obstacle avoidance radar detection, in particular to a tree outline detection method and device based on UAV obstacle avoidance radar. Background technique [0002] In recent years, with the rapid growth of the industry-level drone market in agriculture, electric power, and industry, the key technologies of drones are also making rapid progress, such as real-time image transmission, target recognition, terrain following and other technologies, making industry-level drones Man-machines are becoming more and more intelligent. Among many technological trends, automatic obstacle avoidance is the key to the safety of drones. The precise UAV autonomous obstacle avoidance system can minimize the occurrence of UAV damage, personal accidents or building accidents caused by human operation errors and line of sight problems. The current mainstream products generally use frequency modulation continuous wave (FM...

AI Technical Summary

Problems solved by technology

[0004] When a drone is flying in the air, it is easy to collide with various buildings, tree crowns, and wires. Once a collision occurs, it is very likely to cause a collision of the drone, or even crash or damage the drone. wounding incident

Timely detection of trees is crucial to the flight safety of UAVs. Compared with the detection of trees, it is not difficult for UAV radar to detect dense trees. However, when using the above-mentioned traditional CFAR processing method, the radar is actually What is detected is often only the dense part of the tree, such as the main part of the trunk, and the relatively sparse parts such as the crown and branches. Due to the weak echo of the reflected signal, there will be a covering effect of the strong reflector of the trunk on the weak target. It is difficult to detect relatively sparse parts of trees such as crowns and branches under the cover of strong trunks, such as figure 2 As shown, even if the millimeter-wave radar detects the trees, it cannot accurately detect the outline of the trees because it cannot detect the crowns and branches of the trees. The UAV may still collide with the branches and cause a collision event.

If it is considered to reduce the detection threshold in the CFAR detection method to improve the detection of relatively sparse parts such as crowns and branches, it will cause a large number of false alarms and cause troubles for practical applications.

Images