A Fast 3D Reconstruction Method for Non-Lambertian Surfaces

Abstract

The invention belongs to the technical fields of image understanding and computer vision, and particularly relates to a fast three-dimensional reconstruction method for a non-Lambert surface. The technical scheme is that the fast three-dimensional reconstruction method for the non-Lambert surface sequentially comprises the following steps of: 1, constructing an imaging model, wherein the imaging model is suitable for the non-Lambert surface, and is nearly consistent with an actual imaging condition; 2, establishing an image irradiance equation for the shape from shading (SFS) problem of the non-Lambert surface by using the imaging model and a grayscale image acquired by a video camera; 3, calculating the viscosity solution of the image irradiance equation by utilizing a viscosity solution theory; and 4, reconstructing the three-dimensional shape information of the non-Lambert surface according to the solution of the image irradiance equation. The method has the advantages that the image irradiance equation established under the non-Lambert surface is directly solved, and meanwhile, the viscosity solution of the image irradiance equation is calculated by utilizing the viscosity solution theory which is a weak solution theory, so that the SFS problem of the non-Lambert surface is improved; and the method has the advantages of high speed and accuracy.

Description

technical field [0001] The invention belongs to the technical field of image understanding and computer vision, and mainly relates to a technology for realizing three-dimensional reconstruction by light and shade changes of a single image, in particular to a fast three-dimensional reconstruction method for a non-Lambertian surface. Background technique [0002] In order to meet the needs of 3D modeling in the fields of industrial and agricultural product detection, document image restoration, medical image analysis, biometric recognition, and topographic and geomorphic measurement, and to effectively reconstruct the 3D shape information of the object surface, image understanding and computer It is a very important research direction in the field of visual technology. However, how to quickly and accurately reconstruct the 3D shape of non-Lambertian surfaces has not yet been well resolved. [0003] At present, there are two main types of methods for obtaining 3D shape informa...

AI Technical Summary

Problems solved by technology

Although the contact measurement method is relatively mature and can obtain high-precision data, it also has many shortcomings: ①The contact measurement makes the probe easy to wear and scratch the measured surface, and the measurement process requires manual intervention. Therefore, the measurement cost is high and the degree of automation is low; ②The contact measurement adopts the point-by-point scanning method, and the scanning speed is limited by the mechanical movement, so the measurement time is long and the measurement efficiency is low, which is not suitable for rapid measurement. Occasions; ③The measurement effect of large objects, especially soft materials, is not good, and the surface that cannot be touched by the probe cannot be measured, and the requirements for the measurement environment are high, so its application range is often limited

Unfortunately, for the research on the image irradiance equation established by non-Lambertian surfaces, no relevant literature reports have been found at home and abroad.

Images