Abstract:With the development of computer vision technologies, 3D reconstruction has become a hotspot. At present, 3D reconstruction relies heavily on expensive equipment and has poor real-time performance. In this paper, we aim at solving the problem of 3D reconstruction of an indoor scene with large vertical span. In this paper, we propose a novel approach for 3D reconstruction of indoor scenes with only a Kinect. Firstly, this method uses a Kinect sensor to get color images and depth images of an indoor scene. Secondly, the combination of scale-invariant feature transform and random sample consensus algorithm is used to determine the transformation matrix of adjacent frames, which can be seen as the initial value of iterative closest point (ICP). Thirdly, we establish the relative coordinate relation between pair-wise frames which are the initial point cloud data by using ICP. Finally, we achieve the 3D visual reconstruction model of indoor scene by the top-down image registration of point cloud data. This approach not only mitigates the sensor perspective restriction and achieves the indoor scene reconstruction of large vertical span, but also develops the fast algorithm of indoor scene reconstruction with large amount of cloud data. The experimental results show that the proposed algorithm has better accuracy, better reconstruction effect, and less running time for point cloud registration. In addition, the proposed method has great potential applied to 3D simultaneous location and mapping.

Indoor Scene Reconstruction: from Panorama Images to CAD Models.

A Simple 3D Reconstruction Method Based on Single Camera

A Flexible Method for 3D Reconstruction of Urban Building

Automatic 3D Indoor Scene Modeling from Single Panorama

Outdoor Scene 3D Reconstruction from Multiple Point Cloud

Indoor Structure Understanding from Single 360 Cylindrical Panoramic Image

Modeling room structure from indoor panorama.

Hybrid 3D Reconstruction of Indoor Scenes Integrating Object Recognition

Model-driven Indoor Scenes Modeling from a Single Image.

Understanding of Indoor Scenes Based on Projection of Spatial Rectangles.

3D Reconstruction of Indoor Scenes Via Image Registration

Real-time indoor scene reconstruction with Manhattan assumption

3D Reconstruction Approach for Outdoor Scene Based on Multiple Point Cloud Fusion

Total3DUnderstanding: Joint Layout, Object Pose and Mesh Reconstruction for Indoor Scenes from a Single Image

Indoor Panorama Planar 3D Reconstruction via Divide and Conquer

3D Residential Reconstruction from a Single Indoor Floor Plan.

Online Structure Analysis for Real-Time Indoor Scene Reconstruction

DeepPanoContext: Panoramic 3D Scene Understanding with Holistic Scene Context Graph and Relation-based Optimization

Singe Image-Based Data-Driven Indoor Scenes Modeling

Indoor scene scanning reconstruction method and apparatus

Planar Reconstruction of Indoor Scenes from Sparse Views and Relative Camera Poses