Three-dimensional reconstruction method and system in intelligent transportation

Abstract

The invention discloses a three-dimensional reconstruction method and system in intelligent transportation. The three-dimensional reconstruction method comprises the steps of obtaining information of a detection scene, achieving extraction and dynamic tracking of target feature points in the detection scene, achieving feature point automatic completion and conducting three-dimensional reconstruction on targets through a distribution reconstruction algorithm. The three-dimensional reconstruction method comprises the following steps of S101, carrying out camera calibration, S102, synchronously obtaining original video images with scene depth information, S103, tracking feature points of the targets in a video scene, S104, selecting video key frames, S105, reconstructing the selected frames, S106, packaging subsets, and S107, obtaining a final three-dimensional model. According to the three-dimensional reconstruction method and system in intelligent transportation, an improved KLT tracking algorithm is adopted, a feature point model is optimized, and tracking efficiency is improved; the key frames are extracted to serve as the subsets, and the subsets are reconstructed and finally are merged, so that the problems that the obtained video frames are relatively short, and depth obtaining of image information is influenced by tracking losing of feature points in the scene, faults and the like are solved.

Description

technical field [0001] The invention relates to the technical field of digital image processing, in particular to a three-dimensional reconstruction method and system in intelligent transportation. Background technique [0002] In recent years, 3D reconstruction has been an important research direction in the field of computer vision, and it has a wide range of applications in virtual reality, object recognition and visualization. How to quickly and effectively reconstruct the scene in the computer has always been a hot and difficult point in the field of computer vision research. [0003] 3D reconstruction (also known as "multi-view geometry") is a series of images (such as buildings, streets, vehicles, etc.) in a static scene to estimate the 3D structure of the scene. Simultaneously estimating multi-view geometry may not be as complex as object recognition, which requires many tasks; some examples include obtaining dense reconstructions, creating detailed 3D models or cri...

AI Technical Summary

Problems solved by technology

However, for a long time, due to the lack of necessary technology, intelligent transportation applications such as vehicle detection and road condition analysis have been limited to simple qualitative analysis

Images