Concurrent visual multiple lane detection for autonomous vehicles

Abstract

This paper proposes a monocular vision solution to simultaneous detection of multiple lanes in navigable regions / urban roads using accumulator voting. Unlike other approaches in literature, this paper first examines the extent of lane parameters required for continuous control of any vehicle manually or autonomously. The accumulator-based algorithm is designed using this fundamental control knowledge to vote for the required lane parameters (position of lanes and steering angle required) in the image plane. The novel accumulator voting scheme is called "Parametric Transform for Multi-lane Detection." This paper not only adapts predictive control in the image plane, but also detects multiple lanes in the scene concurrently in the form of multiple peaks in the accumulator. This method is robust to shadows and invariant to color, texture, and width of the road. Finally, the method is designed for dashed/continuous lines.

Lane Level Vehicle Positioning Based on Integrated Multi-sensor Fusion

“…However, multiple lanes should be detected in our system to locate the vehicle within a particular lane. A statistical voting approach is proposed by Gupta [8], however, the computational complexity of the algorithm restricts its implementation in embedded device. An easy but robust approach, named patch-based Hough detection with lane verification, is proposed to detect multiple lanes.…”

Wrong Roadway Detection for Multi-lane Roads

“…Triggered by the digital map information, straight roads and curved roads are handled with different methods. Accumulator voting (in parameter space) for paired lines is used in [9] for generating lane parameters for autonomous vehicle navigation tasks. This algorithm mainly focuses on detecting the driven lane for calculating control parameters for the ego-vehicle (e.g.…”

“…This method detect in some experiments up to four lanes, due to the used expansion of the search space. These methods can potentially be used to detect multiple lanes, but none of them considers the robustness of detection results for positioning with lane-level accuracy [7,9,21]. This paper proposes (1) a low-cost system for reducing wronglane related accidents, (2) a solution for wrong-lane driving detection, and (3) a new multiple-lane detection method.…”