Home > Published Issues > 2021 > Volume 9, No. 4, December 2021 >

Monocular-Based Drivable Area Segmentation by Fusing 3-D and Texture Information

Takehito Ogata
Center for Technology Innovation-Controls, R&D Group, Hitachi, Ltd., Hitachi, Japan

Abstract—In this paper, a monocular camera based drivable area segmentation algorithm is described. The feature-point based motion stereo algorithm is a well-known method to measure 3D environment by using Monocular-camera. However, one of the disadvantages of this algorithm is that it is not suitable to measure 3D information of the area around the traveling direction and non-texture area (e.g. road surface). It is important to know, 3D measurement of the traveling direction and road surface are critical for the driving assistance system. In this paper, we propose the unique drivable area segmentation algorithm. One of its uniqueness is that it combines 3D information of feature points calculated from motion stereo, and segmentation based on similarity of grid-based texture feature. We implement this algorithm in automotive embedded SoC and evaluate various situations.

Index Terms—motion stereo, texture-based feature vector, k-nearest neighbor, occupancy grid map, wide-angle camera

Cite: Takehito Ogata, "Monocular-Based Drivable Area Segmentation by Fusing 3-D and Texture Information," Journal of Image and Graphics, Vol. 9, No. 4, pp. 140-145, December 2021. doi: 10.18178/joig.9.4.140-145

Copyright © 2021 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.