基于双焦成像的单眼立体视觉算法

A Monocular Stereo Vision Algorithm Based on Bifocal Imaging

对单眼立体视觉的双焦成像算法进行了分析.在1 m以内的深度范围里,双焦成像算法可以极好地恢复空间物点的深度.在此范围内,物点越远离中心区域,其深度值的精度越高.为了提高算法精度,需要校正非理想变焦图像对的中心点.分析了中心点校正方法,并给出了校正后空间物点深度恢复的实验结果.

This paper analyzes the monocular stereo vision algorithm based on bifocal imaging. In the depth range less than 1 meter, this algorithm demonstrates excellent recovery of objects＇ depth, and the farther the object point is away from the center area of the frame, the more precise the depth value will be. In order to improve the algorithm precision, the cen- ter point of a pair of zoom images needs adjusting if it is not an ideal pair. This paper introduces the method of center point adjustment and illustrates the experimental result of depth recovery of a spatial object after adjustment.