Three-dimensional global visualization monitoring system construction method based on skeleton-point local dynamic updating

Abstract

A three-dimensional global visualization monitoring system construction method based on skeleton-point local dynamic updating is disclosed. In the method, a mobile-robot visual-angle-scene three-dimensional reconstruction system, a large-range multiple visual-angle-scene three-dimensional reconstruction system and a skeleton-point three-dimensional positioning and matching system are mainly included. The method mainly comprises the following steps of mobile-robot visual-angle-scene three-dimensional reconstruction, large-range multiple visual-angle-scene three-dimensional reconstruction, skeleton-point three-dimensional positioning under a global coordinate system, and accumulative error control and large-range three-dimensional scene real-time updating. In the invention, a multiple-visual-angle camera (camera) which is fixedly placed and an array camera (camera) carried by a mobile robot are combined and a skeleton-point three-dimensional positioning and matching method is used so as to complete real-time three-dimensional scene reconstruction under a large range. And a local differentiation updating method is combined, a changing portion in a three-dimensional scene is updated so that updating efficiency is greatly increased; and a good global three-dimensional map of a large range scene is acquired and the map can be updated in real time.

Description

technical field [0001] The invention relates to a method for constructing a three-dimensional global visualization monitoring system based on local dynamic updating of skeleton points, which can realize three-dimensional global visual monitoring of a large-scale scene through the local dynamic updating of skeleton points, and belongs to the field of computer vision. Background technique [0002] The 3D global visual monitoring of large-scale 3D scenes is a hot research direction in the field of computer vision, and it has broad application prospects in many fields such as industrial monitoring, archaeology, and virtual reality. As basic supporting facilities, at present, major factories, substations and some public facilities have deployed cameras, ranging in number from a few to hundreds, to monitor whether each part of the site is operating normally, but the current monitoring images are all completed manually , and the two-dimensional images acquired by the camera have a ...

AI Technical Summary

Problems solved by technology

However, since the establishment of the entire 3D scene needs to patrol the entire scene area, all the image data can be collected. For a large-scale 3D scene, it takes a long time, and the reconstruction of the entire 3D scene cannot be realized in a short period of time, let alone It meets the requirements of real-time update of 3D scenes, and does not make full use of the video information of multiple cameras already on site

However, using the video information of multiple cameras already on site for 3D reconstruction can only achieve rapid reconstruction of local 3D scenes on site, but cannot obtain a global 3D map. If the local information is spliced ​​with 3D point clouds to obtain a global dynamic 3D map, and Can not meet the needs of real-time updates

Images