论文部分内容阅读
针对无人飞艇地面目标检测中细节信息缺失的问题,提出一种静态目标和运动目标的检测方法。利用Lucas-Kanade方法跟踪目标区域内特征点,从而实现静态目标的连续检测。通过图像特征点的跟踪估计相邻帧图像间的全局运动,进而对图像进行运动补偿,利用补偿后的帧差图实现运动目标的检测。采用上海交通大学“致远一号”无人飞艇采集的实际视频数据进行实验与分析,结果验证了该方法的有效性。
Aiming at the lack of detail information in the unmanned airship ground target detection, a static target and a moving object detection method are proposed. The Lucas-Kanade method is used to track the feature points in the target area to achieve continuous detection of static targets. Through the tracking of image feature points, the global motion between adjacent frame images is estimated, then the image is compensated for motion and the detected frame is used to detect the moving object. The actual video data collected by Shanghai Jiao Tong University “Zhiyuan No.1 ” unmanned aerial vehicle was tested and analyzed. The results verify the effectiveness of this method.