论文部分内容阅读
跑道检测识别与跟踪是基于视觉的无人机自主着陆的前提和难点,本文根据固定翼无人机基于视觉自主着陆的特点设计了包括跑道检测、跑道特征提取、跑道识别和跑道跟踪的方案,并在ARM Cortex-A9处理器中基于Linux系统使用OpenCV实现了该方案。最后按照逐步递进的方式分别对检测、检测识别、检测识别与跟踪结果进行了实验验证,并对实时性进行了分析。实验结果表明,通过该方案可以准确地识别图像中的跑道并具有较快的跟踪速度。
The recognition and tracking of runway are the prerequisites and difficulties of autonomous landing of vision-based UAV. Based on the characteristics of autonomous landings of UAVs based on vision, this paper designs programs including runway detection, runway feature extraction, runway identification and runway tracking. This solution was implemented using OpenCV on a Linux-based system in an ARM Cortex-A9 processor. Finally, in accordance with the gradual progress of the way, respectively, testing, testing and identification, testing and identification of the results of the experimental verification, and real-time analysis. Experimental results show that the program can accurately identify the runway in the image and have a faster tracking speed.