Obstacle detection and classification fusing radar and vision

Abstract

Abstract-This paper presents a system whose aim is to detect and classify road obstacles, like pedestrians and vehicles, by fusing data coming from different sensors: a camera, a radar, and an inertial sensor. The camera is mainly used to refine the vehicles' boundaries detected by the radar and to discard those who might be false positives; at the same time, a symmetry based pedestrian detection algorithm is executed, and its results are merged with a set of regions of interest, provided by a Motion Stereo technique.The tests have been performed in several environments and traffic situations, their results showed how the vision based filtering provides an effective reduction of radar's false positives; furthermore, the regions of interest detected by the Motion Stereo algorithm, truly improves the pedestrian detector's performance again by keeping low the number of detection errors.The system has been shown during the APALACI-PReVENT European IP final demonstration 1 in September 2007 in Versailles (France).

Looking at Vehicles on the Road: A Survey of Vision-Based Vehicle Detection, Tracking, and Behavior Analysis

“…In [166], obstacles are detected using vision operations on the inverse perspective mapped image and ranged using radar. In [167], vehicles are detected with a boosted classifier using Haar and Gabor features and ranged using radar.…”

Overview of Environment Perception for Intelligent Vehicles

“…ROI extraction was performed by using color and intensity information [19]. In [20], a set of Regions of Interests (ROIs) was detected by a Motion Stereo technique to improve the pedestrian detector's performance. Using dense stereo for both ROIs generation and pedestrian classification, a novel pedestrian detection system for intelligent vehicles was presented in [21].…”

Robust monocular lane detection in urban environments

“…• vehicle pitch can be corrected at runtime by means of an IMU, or using a pitch detector algorithm, such as the one presented in [16]; • the image areas corresponding to obstacles can be masked, thus reducing the chance of incurring in false detections. The obstacle detection step can exploit various techniques, such as stereovision or LIDAR-based solutions, as in [17] and [1].…”