A Robust Method for High Quality Texture Mapping of 3D Models

Abstract

The invention discloses a high-quality texture mapping method for a three-dimensional robust model. The high-quality texture mapping method comprises the following steps of firstly, dispersing an input mesh model into point cloud with pixel granularity; secondly, building color consistency target energy, optimizing an input external reference of a camera according to the target energy, and deforming and adjusting a color image; and finally, building the optimal spliced target energy to acquire a texture image mapping relation of which each patch is optimal. Compared with a traditional method, the high-texture mapping method has obvious advantages on input data with a large aligning error and high light.

Description

technical field [0001] The invention relates to a three-dimensional model high-quality texture mapping method, in particular to a robust three-dimensional model high-quality texture mapping method. Background technique [0002] Texture mapping is the process of establishing the correspondence between the surface of a three-dimensional object and the pixel coordinates of a two-dimensional image space. In 3D graphics, texture mapping allows users to collect real natural scene photos as 2D texture data, which is very important for expressing realistic 3D object models. A key issue in the realistic texture mapping problem is how to deal with the registration error of the input data. The registration error mainly comes from the geometric error between the model and the real object, the error of the camera parameters, and the error introduced by the simplification of the camera pinhole model. The registration error of the input data has a direct impact on the final result. [00...

Images