论文部分内容阅读
针对由于CamShift算法跟踪特征单一引起的对颜色相似目标或背景的干扰和对目标遮挡情况较敏感的问题,提出了一种CamShift融合局部特征匹配的无人机目标跟踪算法。实验表明,局部特征匹配算法中BRISK匹配算法在特征检测和特征描述阶段都表现出了较好的性能,融合CamShift算法和BRISK算法的目标跟踪算法在能保证目标跟踪的实时性要求的前提下,改善了CamShift对颜色相似目标或背景的干扰的敏感性,同时增强了对目标遮挡鲁棒性。该方法通过颜色特征和局部特征的共同定位目标,实现了目标的准确跟踪。
Aiming at the problem that the CamShift algorithm can only track the characteristics of a single object and interfere with the similar objects or backgrounds and is more sensitive to the target occlusion, a CamShift fusion algorithm based on local feature matching is proposed. Experiments show that the BRISK matching algorithm in local feature matching algorithm shows good performance in the feature detection and feature description stages. The target tracking algorithm based on CamShift algorithm and BRISK algorithm can guarantee the real-time target tracking requirements, Improved CamShift sensitivity to interference with color-alike targets or backgrounds while enhancing robustness to occlusion of targets. The method achieves the goal of accurate tracking through the co-location of color features and local features.