Projected image correction technology research on autonomous-perception anisotropic surfaces

摘要

In this paper, a projection correction method is proposed for autonomous-perception depth anisotropic surfaces, in order to improve the adaptive perceptual projection of the projection equipment under different circumstances. During the parameter calibration process of a projector-camera system with low recognition precision, as well as low noise resistance at angular points, an angular-point subpixel detection algorithm based on color information guidance was proposed to effectively improve the identification precision of the angular-point detection. Meanwhile, the projection geometry correction algorithm was proposed, which was based on the topology analysis, in order to analyze the spatial topology distribution of the depth anisotropic surface and also to solve the homography matrix in the different regions of the anisotropic surface. Eventually, the homography matrix was used for the geometric distortion correction of the projection distortion image. The experimental analysis showed that the identification accuracy of the proposed method achieved 0.25 pixels, with simultaneous high planar parallelism and liner perpendicularity. Through utilizing the method proposed by us for the projection correction of the depth anisotropic surface, it was confirmed that the geometric distortion correction precision of the proposed method had reached the subpixel level, and the consistency level of the imaging picture was also confirmed. (c) 2018 Optical Society of America