Christopher Paul Urmson

Traffic light mapping and detection

The outdoor perception problem is a major challenge for driver-assistance and autonomous vehicle systems. While these systems can often employ active sensors such as sonar, radar, and lidar to perceive their surroundings, the state of standard traffic lights can only be perceived visually. By using a prior map, a perception system can anticipate and predict the locations of traffic lights and improve detection of the light state. The prior map also encodes the control semantics of the individual lights. This paper presents methods for automatically mapping the three dimensional positions of traffic lights and robustly detecting traffic light state onboard cars with cameras. We have used these methods to map more than four thousand traffic lights, and to perform onboard traffic light detection for thousands of drives through intersections.

Exploiting publicly available cartographic resources for aerial image analysis

Cartographic databases can be kept up to date through aerial image analysis. Such analysis is optimized when one knows what parts of an aerial image are roads and when one knows locations of complex road structures, such as overpasses and intersections. This paper proposes self-supervised computer vision algorithms that analyze a publicly available cartographic resource (i.e., screenshots of road vectors) to, without human intervention, identify road image-regions and detects overpasses. Our algorithm segments a given input image into two parts: road- and non-road image regions. It does so not by learning a global appearance model of roads from hand-labeled data, but rather by approximating a locally consistent model of the roads appearance from self-obtained data. In particular, the learned local model is used to execute a binary classification. We then apply an MRF to smooth potentially inconsistent binary classification outputs. To detect overpasses, our method scrutinizes screenshots of road vector images to approximate the geometry of the underlying road vector and use the estimated geometry to localize overpasses. Our methods, based on experiments using inter-city highway ortho-images, show promising results. Segmentation results showed on average over 90% recall; overpass detection results showed 94% accuracy.

Ortho-image analysis for producing lane-level highway maps

This paper presents new aerial image analysis algorithms that, from highway ortho-images, produce lane-level detailed maps. We analyze screenshots of road vectors to obtain the relevant spatial and photometric cues of road image-regions. We then refine the obtained patterns to generate hypotheses about the true road-lanes. A road-lane hypothesis, since it explains only a part of the true road-lane, is then linked to other hypotheses to completely delineate boundaries of the true road-lanes. Finally, some of the refined image cues about the underlying road network are used to guide a linking process of road-lane hypotheses. We tested the accuracy and robustness of our algorithms with high-resolution, inter-city highway ortho-images. Experimental results show promise in producing lane-level detailed highway maps from ortho-image analysis -- 89% of the true road-lane boundary pixels were successfully detected and 337 out of 417 true road-lanes were correctly recovered.