Research On Structured Light Imaging Based On Light Field Camera

Three-dimensional shape measurement technology is becoming more and more important for today's technology and life.Structured light measurement technology is one of the widely used technologies,with no contact and high precision.However,due to its triangulation system principle,this technology can cause self-occlusion problems for measuring some complex topography objects,affecting the integrity of the final result.In this paper,an emerging image recording device—light field camera is introduced into the structured light system,which becomes light field structured light system.A fusion algorithm integrating light field multi-view data is proposed to solve the self-occlusion problem and improve the final result.Furthermore,a self-adapting view selection algorithm is designed to reduce information redundancy.Finally,the experimental system work-flow of this paper is introduced.Firstly,in order to integrate the multi-view data of the light field,we design a subaperture fusion algorithm based on the filter kernel.Inspired by the angle superresolution of the light field,I regard the sub-apertures as "pixels".The central subaperture is regarded as base,and the other sub-apertures are used to optimize the central sub-aperture through designed filter kernel.By applying the algorithm to actual models,the results of our algorithm was proved to improve the integrity of the results of central sub-aperture,which validate the availability of our algorithm.Then,in order to reduce the time complexity of our fusion algorithm,we propose a sub-aperture selection algorithm based on contribution criterion.The sub-aperture contribution is defined as the valid value provided by sub-aperture at the area where contain invalid pixel of central sub-aperture.The calculation method of the contribution degree is determined through experimental analysis,which is then applied to the actual model.The selected sub-aperture fusion results are compared with the fusion results of all sub-aperture,which verified the availability of the method.Finally,based on the sub-aperture fusion algorithm and selection algorithm proposed in this paper,the work-flow of the light field structured light system is briefly analyzed,including the system work-flow framework,light field camera data decoding,and system calibration.Through the system calibration parameters,the absolute phase map of all previous fusion results are transformed into the final depth maps.By compared these fusion depth maps with the central sub-aperture depth map,we verified the validity of the work.The light field structured light system constructed in this paper and its related algorithms has achieved good results in solving the self-occlusion problem,this work would be useful for future study.