论文部分内容阅读
主要研究基于视觉的无人机在自主着陆时,机场跑道在图像中的检测和识别问题。首先,提取一些高亮度斑点,然后采用特殊聚类处理算法对这些含背景噪声、伪目标的点进行聚类,并根据透视原理、矩形约束条件以及目标的前视约束条件共同建立识别模型。在这些聚类中使用识别模型进行识别,这在很大程度上减少计算时间,实验表明,在距离跑道较远、跑道标记不很清晰的情况下,仍然可以有效检测和识别跑道,适合于基于视觉的无人机的自主着陆过程。
Mainly based on visual UAV in autonomous landing, the airport runway in the image detection and identification issues. Firstly, some high-brightness speckles are extracted, then the special clustering algorithm is used to cluster these points with background noises and false targets. The recognition model is established based on the perspective principle, the rectangular constraint and the target forward-looking constraints. Recognition using these recognition models in these clusters greatly reduces computational time and experiments show that runways can still be effectively detected and identified at runway distances and runway markings that are not well defined Visual drone autonomy landing process.