Three-dimensional reconstruction method and system

Abstract

The invention provides a three-dimensional reconstruction method and a system. The three-dimensional reconstruction method comprises steps of adopting dynamic averaging techniques to captured images under any resolution to perform three-dimensional matching based on one-dimensional phase correlation; and obtaining three-dimensional coordinates of each point of a three-dimensional model according to three-dimensional matching results and prestored internal and external parameters of a binocular video camera, and performing a coloring process for output. Three-dimensional matching is performed by using one-dimensional phase related algorithm under different resolution in combination with the dynamic averaging techniques, matching precision is provided through a small amount of calculation, and matching degree is further utilized for matching selectively before images of native resolution are matched. The dynamic averaging techniques are subjected to weighting processing during the matching process, thereby greatly reducing calculation time of outputting high-precision three-dimensional models, improving matching efficiency, and enabling rapid achievement of three-dimensional model reconstruction.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a three-dimensional reconstruction method and system. Background technique [0002] A 3D model is a 3D polygonal representation of an object. With the development of 3D technology, 3D models have been gradually applied to various fields. As an effective means to obtain 3D information, 3D reconstruction has become a research hotspot in the fields of computer graphics and computer vision. The passive acquisition of 3D information technology based on stereo vision only requires simple equipment, such as two cameras, which is more acceptable to the public than active equipment such as expensive laser scanners and CT. Although passive technology is suitable for many applications, the low accuracy of 3D reconstruction is still the main problem of passive 3D reconstruction technology. [0003] Due to its anti-noise, robustness and high-precision characteristics, the stereo matching tec...

AI Technical Summary

Problems solved by technology

In the literature S.Nagashima, K.Ito, T.Aoki, H.Ishii and K.Kobayashi, "High-accuracy estimation of image rotation using 1D phase-only correlation," IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, vo1.E92-A, no.1, pp.235-243, 2009 proposed to use the epipolar limit to adopt one-dimensional phase correlation technology, and use the mean value technology to improve the matching accuracy, although this method improves the matching speed, in The author's experimental conditions also have good results, but the averaging technique cannot adapt to the multi-resolution structure of the image, and it is easy to cause edge mismatches; and processing each sequence equally will affect the matching accuracy and speed to a certain extent

Images