Interested sample point pre-selection based dense terrain reconstruction for autonomous navigation

Authors:
Lili Lin;Wenhui Zhou
Affiliations:
College of Information and Electronic Engineering, Zhejiang Gongshang University and Zhejiang Provincial Key Laboratory of Information, Network Technology, Zhejiang University, Hangzhou, China;College of Computer and Software, Hangzhou Dianzi University, Hangzhou, China
Venue:
IITA'09 Proceedings of the 3rd international conference on Intelligent information technology application
Year:
2009

Citing 3
Cited 0

Modeling rugged terrain by mobile robots with multiple sensors

Modeling rugged terrain by mobile robots with multiple sensors
High-Resolution Terrain Map from Multiple Sensor Data

IEEE Transactions on Pattern Analysis and Machine Intelligence - Special issue on interpretation of 3-D scenes—part II
Comparison of energy minimization algorithms for highly connected graphs

ECCV'06 Proceedings of the 9th European conference on Computer Vision - Volume Part II

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper addresses the problem of terrain reconstruction for autonomous navigation. Dense stereo matching based terrain reconstruction methods are sensitive to mismatch pixels in disparity map, and consume lots of computation on pixels in uninterested regions. Traditional sample point pre-selection methods (SPPS) are effective, but they can only obtain sparse terrain map for grid-based representation. We extend the idea of SPPS methods and propose an interested sample point pre-selection (ISPPS) based dense reconstruction method. In this method, we choose appropriate interested sample points and set their reconstruction results as local control points, which can reduce the ambiguity and computation complexity in successive dense reconstruction procedures. The most prominent advantage of this method is it directly recovers the 3D terrain model without dense disparity map. Experiments show the proposed method is robust, and can achieve precise dense reconstruction with low computation complexity.