Environmental Perception Using Fish-Eye Cameras for Autonomous Driving

The fish-eye camera has become an indispensable sensor in autonomous driving. Due to the unique projection principle of the fish-eye camera, it provides a large field of view (FoV). Because of this special feature, the fish-eye camera has abundant applications in environmental perception and autonomous driving. However, many challenges still exist in the practical application of fish-eye cameras. In this chapter, typical fish-eye datasets, including real data and generated virtual data, are introduced. Then, the projection principle of the fish-eye camera and four classical fish-eye image representation models are given. By sorting and summarizing relevant research, we show various applications of fish-eye cameras in environmental perception, including semantic understanding and target detection. These works have designed various strategies to take advantage of fish-eye cameras and prevent image distortions, showing the broad application prospects of fish-eye cameras. The development trend of fish-eye cameras in autonomous driving is discussed.