I. INTRODUCTION( 1 INCE traffic scenes are complex and include lots of information, keeping constant attention on the traffic signs is not an easy task for drivers. Therefore some trafficdata (signs) can be missed for several causes such as the complexity of the road scene, the high number of visual information, or even the driver' s stress or visual fatigue [1]. In order to assist this task, several driver assistant systems have been suggested in the past years using either database information (i.e. learned Geographic Information Systems) or on-vehicle sensors (i.e. laser, camera, etc.) to provide various environment information such as traffic signs, speed limits, traffic lights, crosswalks, .... -,or any other information like pedestrian or obstacles. The specific functionality of traffic lights detection shall be very useful since traffic lights position and state (80, stop or caution) provide good knowledge of the traffic environment such as high probability of crossroads/crosswalk, dangerous area, etc. Furthermore detecting traffic lights with an on-vehicle camera could also be used to improve fusion of GPS and camera visual data in order to make visual projection of roadinformation on the windshield. Because all (or the upmost of) the previous works were only applied for suspended trafic lights recognition, we propose in this paper a generic modular real-time algorithm
I. INTRODUCTION<br>( 1 INCE traffic scenes are complex and include lots of information, keeping constant attention on the traffic signs is not an easy task for drivers. Therefore some traffic<br>data (signs) can be missed for several causes such as the complexity of the road scene, the high number of visual information, or even the driver' s stress or visual fatigue [1]. In order to assist this task, several driver assistant systems have been suggested in the past years using either database information (i.e. learned Geographic Information Systems) or on-vehicle sensors (i.e. laser, camera, etc.) to provide various environment information such as traffic signs, speed limits, traffic lights, crosswalks, .... -,or any other information like pedestrian or obstacles. The specific functionality of traffic lights detection shall be very useful since traffic lights position and state (80, stop or caution) provide good knowledge of the traffic environment such as high probability of crossroads/crosswalk, dangerous area, etc. Furthermore detecting traffic lights with an on-vehicle camera could also be used to improve fusion of GPS and camera visual data in order to make visual projection of road<br>information on the windshield. Because all (or the upmost of) the previous works were only applied for suspended trafic lights recognition, we propose in this paper a generic modular real-time algorithm
正在翻译中..