Real-Time Fusion of Visual Images and Laser Data Images for Safe Navigation in Outdoor Environments

Abstract

[EN]In recent years, two dimensional laser range finders mounted on vehicles is becoming a
\nfruitful solution to achieve safety and environment recognition requirements (Keicher &
\nSeufert, 2000), (Stentz et al., 2002), (DARPA, 2007). They provide real-time accurate range
\nmeasurements in large angular fields at a fixed height above the ground plane, and enable
\nrobots and vehicles to perform more confidently a variety of tasks by fusing images from
\nvisual cameras with range data (Baltzakis et al., 2003). Lasers have normally been used in
\nindustrial surveillance applications to detect unexpected objects and persons in indoor
\nenvironments. In the last decade, laser range finder are moving from indoor to outdoor rural
\nand urban applications for 3D imaging (Yokota et al., 2004), vehicle guidance (Barawid et
\nal., 2007), autonomous navigation (Garcia-Pérez et al., 2008), and objects recognition and
\nclassification (Lee & Ehsani, 2008), (Edan & Kondo, 2009), (Katz et al., 2010). Unlike
\nindustrial applications, which deal with simple, repetitive and well-defined objects, cameralaser
\nsystems on board off-road vehicles require advanced real-time techniques and
\nalgorithms to deal with dynamic unexpected objects. Natural environments are complex
\nand loosely structured with great differences among consecutive scenes and scenarios.
\nVision systems still present severe drawbacks, caused by lighting variability that depends
\non unpredictable weather conditions. Camera-laser objects feature fusion and classification
\nis still a challenge within the paradigm of artificial perception and mobile robotics in
\noutdoor environments with the presence of dust, dirty, rain, and extreme temperature and
\nhumidity. Real time relevant objects perception, task driven, is a main issue for subsequent
\nactions decision in safe unmanned navigation. In comparison with industrial automation
\nsystems, the precision required in objects location is usually low, as it is the speed of most
\nrural vehicles that operate in bounded and low structured outdoor environments.
\nTo this aim, current work is focused on the development of algorithms and strategies for
\nfusing 2D laser data and visual images, to accomplish real-time detection and classification
\nof unexpected objects close to the vehicle, to guarantee safe navigation. Next, class
\ninformation can be integrated within the global navigation architecture, in control modules,
\nsuch as, stop, obstacle avoidance, tracking or mapping.Section 2 includes a description of the commercial vehicle, robot-tractor DEDALO and the
\nvision systems on board. Section 3 addresses some drawbacks in outdoor perception.
\nSection 4 analyses the proposed laser data and visual images fusion method, focused in the
\nreduction of the visual image area to the region of interest wherein objects are detected by
\nthe laser. Two methods of segmentation are described in Section 5, to extract the shorter area
\nof the visual image (ROI) resulting from the fusion process. Section 6 displays the colour
\nbased classification results of the largest segmented object in the region of interest. Some
\nconclusions are outlined in Section 7, and acknowledgements and references are displayed
\nin Section 8 and Section 9.