Stereovision-based road boundary detection for intelligent vehicles in challenging scenarios

Abstract

Road detection is a crucial problem for intelligent vehicles and mobile robots. Most of the methods proposed nowadays only achieve reliable results in relatively well-arranged environments. In this paper, we proposed a stereovision-based road boundary detection method by combining homography estimation and MRF-based belief propagation to cope with challenging scenarios such as unstructured roads with unhomogeneous surfaces. In the method, each pixel in the reference image is firstly labeled as "road" or "non-road" by minimizing a well defined energy function that accounts for the planar road region. Subsequently, both of the road boundaries are generated using Catmull-Rom splines based on RANdom SAmple Consensus (RANSAC) algorithm with varying road structure models to help the intelligent vehicle understand the structure as well as safe range of current road. In the suggested framework, both intensity and geometry information of road scenarios are used to contain all the regions belonging to the planar road plane, and the left and right road boundaries are generated separately using a robust fitting algorithm to handle different road structures. Therefore, more accurate as well as robust detection of the road can be expected. Experimental results on a wide variety of typical but challenging scenarios have demonstrated the effectiveness of the proposed method.978-1-4244-3804-4/09/$25.00 ©2009 IEEE

Perception, Planning, Control, and Coordination for Autonomous Vehicles

“…This is the main reason why stereo is not widely adopted in this research field. In [108], the author used dense mapping to obtain disparity while in [109], Maximum A Posteriori -Markov Random Field (MAP-MRF) approach was applied. However, both methods are not very effective and subject to smoothing noise.…”

A single 3-D feature-based visual servoing

“…However, due to noise influence, it seems tricky in practice to refer to a single 3-D feature for performing an accurate positioning or motion control. In fact, a 3-D feature is usually the result of several robust estimations [5] issued from a high volume of visual data. Thanks to a subpixel interpolation and fine edges detection in highresolution images, a 2-D straight line may be accurately fitted (or located if it is a projection of a real 3-D line, like object borders) and is very stable through repetitive segmentation [8], since this kind of feature is unsensitive to a phase distorsion during the detection ( [19], [20]) and can rely on many input subpixel edges.…”

Self-supervised learning method for unstructured road detection using Fuzzy Support Vector Machines

“…It is noted that smaller i s reduces the effect of the parameter i  in problem (2) such that the corresponding point i x has less contribution in the minimization in (2). Like SVMs, searching the optimal hyperplane in (2) is a QP problem, which can be solved by constructing a Lagrangian and transformed into the dual maximize 1 1 1 In proposed algorithm, the training set can be automatically found out from online.…”