Real augmented reality method based on ORB-SLAM and depth camera

Abstract

A real augmented reality method based on ORB-SLAM and a depth camera comprises the following steps: (1) determining an initial coordinate plane by use of a Marker-less technology, and initializing a slam coordinate system with real scales; (2) importing a map, and combining a directly expressed map with a feature map according to extracted feature points; (3) using monocular SLAM o build a real scale map, and saving the map; (4) placing a virtual object in a scene, and calculating the distance between the virtual object and the camera through a 3D engine; and (5) using the depth camera to get scene depth data, and performing deep fusion according to the distance data of the virtual object obtained in the 3D engine to achieve a blocking effect and make the real object and a virtual model have strong interaction directly. The method is of very strong robustness, and realizes blocking interaction between a real object and virtual data.

Description

technical field [0001] The invention belongs to the technical field of AR (Augmented Reality, Augmented Reality), which combines SLAM perception environment information technology and depth camera acquisition scene three-dimensional information technology, and relates to a realistic augmented reality method. Background technique [0002] SLAM (simultaneous localization and mapping, real-time positioning and map construction) refers to the creation of maps by robots in a completely unknown environment under the condition of uncertain positions, and at the same time use maps for autonomous positioning and navigation. ORB-SLAM (a real-time accurate monocular SLAM system based on ORB features, SLAM system based on ORB feature points), ORB refers to a rotation invariant feature, the entire algorithm is based on ORB features, the latest ORB- The progress of SLAM is based on the key frame of ORB-SLAM for semi-dense scene reconstruction. Augmented reality is to apply the informatio...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the shortcomings of the existing AR technology that cannot accurately calculate the three-dimensional information of objects, resulting in occluded objects that cannot be correctly understood and constructed, and poor robustness, the present invention provides a robust ORB-SLAM-based and depth camera The realistic augmented reality method relies on SLAM to perceive the three-dimensional environment of the object, the depth camera obtains the scene distance information, and deeply integrates the real object with the virtual model to achieve the "parallax" effect like the human eye, enabling gesture recognition, perspective tracking, Scene reconstruction is more accurate, so that real objects can block the virtual model