关键词: 盲环境, 移动定位, 序列编码图形, 单目视觉, POSIT算法

Abstract: This paper proposes a mobile positioning algorithm for moving targets in a blind environment without satellite navigation signals and based on coded sequence pattern signposts and a monocular vision system. The coded sequence pattern signposts were fixed along the blind environment. The 3D coordinates of the center of each coded pattern identified a signpost. A pinhole imaging model was then established using a camera equipped with through-the-lens distortion analysis. A calibrated monocular vision system was mounted on the target object to capture images with coded patterns. Image processing methods were used to interpret the coded sequence pattern signposts, and to inquire about the coordinates of the center of each signpost according to its code pattern. The translation and rotation parameters of the monocular vision system were computed by using a POSIT algorithm so as to compute the pose of a moving target. Experiment results indicated that the proposed method delivers fast and accurate pose estimation for moving targets in a blind environment without GNSS signals.