Plane Depth Estimation Based on Single Image and Projection of Four Laser Points

摘要

Single image depth estimation is a challenging problem and provides many application scenarios like object detection, pose calculation and path planning. To avoid the large data restriction in specific applications with deep learning based methods, this paper proposes a method of depth estimation for an arbitrary spatial plane using four laser points projected in the plane. Single plane is addressed at this work and the experimental results based on construction robot show that the proposed depth estimation method is feasible and valid.