无人直升机着陆控制的视觉图像特征处理算法

基于视觉的无人直升机着陆是无人直升机着陆控制中的主要发展方向,视觉图像处理算法将直接影响无人直升机姿态估计以及着陆控制精度.提出了一种应用于无人直升机着陆控制的视觉图像特征处理算法,对目标灰度图像进行阈值处理,边界跟踪,利用凸四边形几何特点提取特征点,最后将所有特征点标记出来.实验结果显示提出的算法即能满足实时处理速度要求,也能满足较高的特征点定位精度要求.这一方法可用于无人直升机着陆控制的相对位置姿态估计.

An Image Feature Processing Algorithm for UAV Helicopter Landing Control

Vision-based landing of the Unmand Aerial Vehicle (UAV) is the main direction of landing control for UAV, in which the image processing algorithm directly influences the precision of pose estimation and landing control. An Image-feature processing algorithm applied to UAV visual-based landing control is presented. The algorithm includes thresholding the grayscale image to a binary image, following feature borders, extracting feature corners based on the convexity of 4-sided polygons and labeling all features at last. It is illustrated by experimental results that the presented algorithm can meet the requirements for real-time processing speed as well as for high accuracy of feature point's location. This algorithm can be used in the estimation of displacement and attitude relative to target platform for UAV helicopter landing control.