Autonomous Perceptual Projection Correction Technique of Deep Heterogeneous Surface

This paper proposes a projection correction method which to improve the adaptive perception projection of the projection equipment in different environments. Firstly, in the process of photon signal transmission, projector-camera can cause the loss of photon signal due to the coupling of system channel. Therefore, this paper proposes a system coupling correction scheme, which effectively reduces the system coupling crosstalk. Secondly, in order to establish the feature mapping relationship between the projection image and the deep heterogeneous surface quickly, a projection feature image of color structured light mesh fringe is designed. Finally, due to the feature point of the heterogeneous surface is quite different in topological structure, it will lead to the problem of inconsistent geometric mapping relation. For this reason, a projective geometric correction algorithm for topological analysis is proposed, analyzing the spatial topological distribution of the depth heterogeneous surface and solving the homography matrix of each region in the heterogeneous surface, then the geometric correction of the projected distortion image is solved by using the homography matrix set. From the experimental analysis we can see that, in the deep heterogeneous surface environment, the average error, the maximum error and the root-mean-square error of the correction image respectively are 0.424 pixels, 0.862 pixels and 0.216 pixels. At the same time, the parallelism of the distortion correction image is kept 90° substantially. It can be seen that the geometric distortion correction accuracy of this method has reached the sub-pixel level and the imaging screen consistency level.