Kernel based approach for accurate surface estimation. (November 2016)

• Record Type: Journal Article
• Authors: Singh, Mahesh K.; Venkatesh, K.S.; Dutta, Ashish
• Is Part Of: Computers & electrical engineering. Volume 56(2016)
• Page Start: 763
• Page End: 772
• Publication Date: 2016-11
• Subjects: Laser range scanner -- Max-margin formulation -- Surface estimation -- Kernel -- Sub-gradient method; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2016.03.001
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: The problem of detecting the surface of an unknown arbitrarily-shaped scene from a set of points is proposed. Range data is obtained from in-house developed Laser range scanner. The surface estimation problem is described as a max-margin based formulation of a kernel function and solving the objective function using sub-gradient method. Additional geometric ray based information is used to eliminate the unnecessary bumps on the surface and increase the precision. Graphical abstract: Abstract: Accurate surface estimation is a critical step for autonomous robot navigation on a rough terrain. In this paper, we present a new method for estimating the surface of an unknown arbitrarily shaped terrain from the range data. The terrain modeling problem is generally formulated as the estimation of a function whose zero-set corresponds to the surface to be reconstructed. A Laser range scanner has been built for acquisition of range data. The range data from the scanner samples the terrain unevenly, and is more sparse for distant regions from the sensor. The paper describes the surface estimation problem as a max-margin based formulation of a non-stationary kernel function and minimizes the objective function using sub-gradient method. Unlike other methods, additional geometric ray based information is used to eliminate the unnecessary bumps on the surface and increase the precision. The experimental results validate the robustness of the proposed approach.