激光测距雷达与视觉传感器的配准是视觉图像与激光距离信息融合的前提。激光测距雷达与视觉传感器的配准可分为点匹配与平面匹配。由于混合像素和非结构空间特征激光点丢失现象,使得基于点匹配的视觉传感器与激光测距雷达配准算法精度难以提高。通过对混合像素和非结构空间特征激光点丢失问题的分析,提出了一种结合平面建模思想,利用假想激光光束与平面模型相交提取匹配特征点的方法。验证了实验结果并进行了算法精度比较。实验结果表明,这种改进的特征匹配点提取算法解决了特征点缺失,并且提高了点匹配的精度,使匹配性能大大地改善。

Space alignment is the first step in data fusion for vision sensor and LMS. Point matching and plane matching are commonly used in space alignment of vision sensor and LMS. But the precision of point matching method is hardly improved, since mixed pixel and matching point loss in non-structure space. Mixed pixel and matching point loss in non-structure space are analyzed.　A method for matching point extracting by modeling of pattern and LMS laser is proposed. The result is verified and space alignment precision is compared. The simulated results show that this algorithm can improve the precision of point matching method, and finally the performance of point matching method is improved.