Scene 3D reconstruction method, device and terminal equipment

Abstract

A scene three dimensional reconstruction method, a device and terminal equipment are provided; the method comprises the following steps: building a first characteristic point set and a first characteristic plane set, and using the first characteristic points and first characteristic planes to form map initial markers; building a second characteristic point set from a present frame color picture, and building a second characteristic plane set from corresponding point clouds in a point cloud graph; determining a matching relation between the second characteristic point set and the first characteristic point set and the matching relation between the second characteristic plane set and the first characteristic plane set; determining whether the present frame is a key frame or not according to the determined matching relations and the gap frames between the present frame and the previous key frame; if yes, determining whether the present frame initial camera pose is valid or not; projecting the second characteristic point and the second characteristic plane of the key frame of the valid initial camera pose on a coordinate system, thus obtaining the added first characteristic point and first characteristic plane; creating a map according to all markers of the map, and realizing three dimensional reconstruction of the scene. The method can more accurately reconstruct the scene.

Description

technical field [0001] The invention belongs to the technical field of three-dimensional scene reconstruction, and in particular relates to a three-dimensional scene reconstruction method, device and terminal equipment. Background technique [0002] Since the 3D reconstruction method of the scene is beneficial to many practical applications, such as virtual reality and augmented reality, localization and path planning of robots, and indoor work of autonomous guided vehicles, etc., it has received increasing attention from researchers. Using RGB-D cameras for 3D reconstruction has unique advantages over binocular cameras, because RGB-D cameras can provide depth values ​​corresponding to image pixel coordinates without spending a lot of computing resources for pixel parallax calculations. Traditional 3D reconstruction methods based on RGB-D cameras mainly use feature points for feature detection and matching. [0003] Generally speaking, the 3D reconstruction based on RGB-D c...

AI Technical Summary

Problems solved by technology

[0005] In view of this, the embodiment of the present invention provides a three-dimensional scene reconstruction method, device and terminal equipment to solve the problem that it is difficult to detect a sufficiently accurate feature matching relationship in a scene with low texture and high texture repeatability in the prior art. It is difficult to get an accurate camera pose, which leads to the problem that the scene is difficult to reconstruct accurately

Images