Laser Intensity-Based Obstacle Detection and Tracking

Abstract

One of the major challenges in designing intelligent vehicles capable of autonomous travel on highways is reliable obstacle detection. Highways present an unknown and dynamic environment with real-time constraints. In addition, the high speeds of travel force a system to detect objects at long ranges. Because of its necessity for mobile robot platforms and intelligent vehicles, there has been a great amount of research devoted to the obstacle detection problem. Although there are a number of methods that can successfully detect moving vehicles, the more difficult problem of detecting animals or small, static road debris such as tires, boxes, or crates remains unsolved. Laser range scanners, or ladars, have been used for many years for obstacle detection. Laser scanners operate by sweeping a laser across a scene and at each angle, measuring the range and returned intensity. Past researchers have ignored the intensity signal while focusing on the range returned from the laser, since the range provides direct 3-D informa-tion useful for mapping. In this thesis, I demonstrate how laser intensity alone can be used to detect and track obstacles. While laser ranging demands fast, complicated electronics, intensity can be measured cheaply. Minimizing cost will be extremely important for any consumer system, and is a strong motivating factor for this thesis. Laser intensity provides different information from ordinary video data since lighting and viewing directions are coincident. At long ranges and grazing angles, vertical obsta-cles reflect significantly more laser energy than the horizontal road. The obstacle detection system uses a high-performance laser scanner which provides fast single-line laser scans. Histogram analysis on the returned intensity signal is used to select obstacle candidates. After candidates are matched and merged with candidates from previous scans, the range to each obstacle is estimated by a novel intensity and position tracking method. Finally, the positions of all obstacles are updated based on vehicle motion before the next laser scan is acquired. Since laser intensity has been ignored in the obstacle detection literature, I have devoted a significant portion of the thesis to examining the intensity measurements pro-vided by the scanner. I present a laser reflectance model which provides good results for a wide variety of object surfaces. The reflectance model is based on experimental results and a combination of two popular reflectance theories from the computer graphics and com-puter vision literature. I also discuss road and system geometry in detail, since geometry affects the obstacle detection problem significantly.