论文部分内容阅读
自主障碍检测与回避是无人机低高度飞行时保障其生存性的一项关键技术,有重要的研究意义。通过对机器视觉原理的研究,考虑到支持向量机方法能同时减小匹配难度和计算量,实时性能、泛化性能良好,故采用该方法通过离线监督学习,将无人机前视图像分割为天空与非天空2部分,并将非天空部分作为需要回避的障碍,实现无人机基于视觉的障碍检测系统,为后续的视觉制导提供信息。实验结果表明,支持向量机能有效准确地实现图像的天空分割,并具有良好的泛化性能。
Independent obstacle detection and avoidance is one of the key technologies to ensure the survivability of a UAV during flight at low altitude, which has important research significance. By studying the principle of machine vision, taking into account the support vector machine method can reduce the matching difficulty and the amount of computation at the same time, the real-time performance and the generalization performance are good. Therefore, this method can be used to monitor the unmanned aerial vehicle Sky and non-sky 2 parts, and non-sky part as an obstacle to be avoided to achieve vision-based obstacle detection system for UAV, to provide follow-up visual guidance information. Experimental results show that SVM can effectively and accurately image the sky and has good generalization performance.