Paper Info
Title
Hand-Held 3d Dense Map Construction System Based On Improved Icp Algorithm
Abstract
In view of the shortcomings of low-dimensional modeling, difficult to move and expensive, the 3D reconstruction system based on the improved ICP (Iterative Closest Point) algorithm is proposed. Firstly, the Kinect sensor is employed to collect color images and depth images, and the Kinect pose is calculated by tracking and acquiring image feature points, which effectively improves the camera's error in obtaining depth images and image noise, and realizes local map construction. Based on the ORB-SLAM (Simultaneous localization and mapping based on ORiented brief feature recognition) system, the key frames are extracted to obtain more image pixels. The improved ICP algorithm is presented to construct a three-dimensional dense map, and a good three-dimensional model of indoor scene is obtained. The experimental results show that the proposed high-resolution three-dimensional reconstruction system with large scenes has good stability, accuracy and robustness, and has great application value.
Year
Venue
Field
2019
2019 12TH ASIAN CONTROL CONFERENCE (ASCC)
Local map,Computer science,Feature recognition,Algorithm,Robustness (computer science),Image noise,Pixel,Simultaneous localization and mapping,3D reconstruction,Iterative closest point
DocType
ISSN
Citations 
Conference
2072-5639
0
PageRank 
References 
Authors
0.34
0
4
Name
Order
Citations
PageRank
Lei Yu1777.87
Xiaofan Fu200.34
Haonan Xu311.36
Shu-Min Fei4115096.93