异己飞行物检测与跟踪的图像处理方法研究

Research on Image Processing of Alien Flying Object Detection and Tracking

为了实现飞行区域内异己飞行物的快速实时检测处理,提出了一种基于图像处理的异己飞行物检测与路径跟踪算法。通过提取自体飞行器上摄像头视频流的单应矩阵来计算物体的移动参数,运用多假设跟踪方法来确定被测物体是否正在接近自体飞行器。该方法有较强的抵抗图像背景灰度变化的能力,在外界光照明显发生变化时,依然能准确、清晰地检测出目标。通过两架无人飞行器模拟自体飞行器和异己飞行物进行试验,结果表明,该算法性能良好,在检测复杂环境中的表现也非常突出。该方法用于识别异己飞行物是可行的、有效的。

In order to realize a rapid detection and processing of the alien flying objects in the flying area, a kind of alien flying object detection and path tracing algorithm based on image processing is put forward. By extrac- ting the video stream＇ s homography matrix of the camera, the objective＇ s movement parameters is calculated. Mul- tiple hypothesis tracking method is used to make sure whether the measured object is reaching the autologous air- craft or not. The method can resist the change of the image background＇ s gray-level. When the light changes obvi- ously, it can also detect the target accurately and clearly. Through the experiment with two unmanned aircraft simu- lating the autologous aircraft and the alien flying objects, it indicates that the algorithm has a good performance and it＇ s also very prominent when to detect in a complex environment. This method is feasible and effective for identif- ying the alien flying objects.