Texture Mapping Methods For Multi-View Three Dimensional Reconstruction

Three-dimensional reconstruction is an important subject in the field of computer vision and computer graphics.In particular,3D colored human model is of great significance in costume customization,virtual reality and game entertainment,movie production and so on.As a crucial technology,texture mapping can greatly enhance the realism of the model and is an indispensable part of 3D reconstruction.In the process of reconstruction,how to effectively deal with the errors caused by camera illumination,calibration and geometric model reconstruction,and so on,to obtain the high-quality texture mapping result on the surface of human body model is an important step in the reconstruction of high-precision static human body and also a challenging issue.In order to solve these problems,this paper proposes a hierarchical optimization texture mapping method based on multi-view 3D reconstruction,which handles the image color,camera pose,local image details and texture boundaries separately from the global and the local.As for global consistency,it minimizes color difference between images using color transfer.Then it corrects global camera pose by the Guess-Newton method.As for local detail,based on the idea of bidirectional similarity,the local detail of the image is synthesized to realize the accurate correspondence between the image and the geometric model.Finally,we choose best view for every face on the model and adjust the color on texture boundaries to smooth the seam.Taking the 3D color human model reconstruction as an example,the effectiveness of the algorithm is verified by comparing several groups of experiments.The results show that our algorithm can effectively correct the geometric and color errors,and realize high-quality texture mapping.