Research on three-dimensional animation character hair rendering and virtual scene fusion technology based on ray tracing algorithm

Abstract

The increasingly powerful performance of computational graphics chips, the rise of augmented reality and virtual reality, real-time graphics rendering has a pivotal role in today's social life, and realistic hair rendering has been one of the difficult problems in computer graphics. In this paper, based on the changes of viewpoints, objects, and light sources in the 3D scene changes, we propose ray tracing model and calculate its pixel unbiasedness to ensure the correctness of the calculation results. A simplified Fresnel formula is used to calculate the contour light of hairs, and at the same time, an improvement of the basic Fresnel formula is proposed to enhance the calculation of the table, contour light intensity and width. Based on the principle of hierarchical shadow mapping, realize the simulation of hair sub-shadow effect. Construct the virtual scene, obtain the body skeleton and other data of the 3D animated character, and improve the fusion between the virtual scene and the 3D animated character. After analyzing the rendering performance and the fusion effect of the virtual scene, after turning on the simulation module, the GPU time increases by 16.98%, while the CPU time increases by 25 μs, and the system's overhead of the Compute Shader reaches 18.65%. The algorithm seeks the mean image in 40-60 images, and the 2×2 chunk fitting method can get better experimental results, at this time, the PSNR is 38.7651dB and 38.8879dB, respectively.