Computational holography is an important branch of three-dimensional display technology.Compared with common two-dimensional image,computer-generated hologram(CGH)can record phase information represented depth perception,and the display effect of reconstructed images meet the visual requirement of human eyes.However,there are still some problems in the computational holography,such as slow calculation and dealing with occlusion problems difficultly.In this paper,it is based on the ray tracing algorithm and look-up table method,Graphics Processing Unit(GPU)is used to handle data in parallel during the calculation process,which can greatly improve the computing efficiency and show the occlusion effect.The specific work is listed as follows:1.A real-time CGH method based on extended look-up table technics is proposed.Different layers is carried out according to the depth values of object points,and object points with the same depth value are located in the same layer.In the first frame,a table is calculated for each layer and stored in the memory.The complex amplitude information of wavefront plane is stored in each table.The calculation domain of Fresnel zone plate corresponding to each object point is gradually extended as the number of frames increases.The CGH is completed until the maximum resolution of each table is reached.The corresponding computation of each object point is processed by GPU in parallel.The CGHs are generated by the proposed method and other methods respectively,and the calculation speed is compared.The experimental results show that the method can calculate CGH in real-time.The calculated frame rate can reach 81.95 fps to generate CGH with resolution of 1024×1024 based on 150×150 rays.2.A real-time CGH method based on ray tracing occlusion processing technics is proposed.The camera located at holographic plane can launch multiple rays to the three-dimensional scene in parallel.The information obtained from the intersections can be used to calculate complex amplitude of the object light wave,as well as position coordinate and coordinate system formulas of the cameras placed at the intersection points.By placing cameras on the partially blocked model and launching rays to the holographic plane in reverse,it is used to determine whether to call the information of tables.The information about the table corresponding to the ray that can reach the wavefront plane can be called during each frame calculation.The experimental results show that the holograms calculated by the proposed method can be loaded into a spatial light modulator(SLM),and the reconstructed images can reflect the occlusion relationship between the models. |