Surround-View Vision-based 3D Detection for Autonomous Driving: A Survey

Abstract

Vision-based 3D Detection task is a fundamental task for the perception of an autonomous driving system, which has piqued interest amongst many researchers and autonomous driving engineers. However, achieving a rather good 3D BEV (Bird’s Eye View) performance is not an easy task using 2D sensor input data of monocular cameras. This paper provides a literature survey of the existing Vision-Based 3D detection methods focused on autonomous driving. We have made detailed analyses of over 60 papers leveraging Vision BEV detection approaches and binned them into different sub-groups for an easier understanding of the common trends. Moreover, we have highlighted how the literature and industry trends have moved towards surround-view image-based methods and noted thoughts on what special cases these surround-view methods address. In conclusion, we provoke thoughts of 3D Vision techniques for future research based on the shortcomings of the current methods, including the direction of collaborative perception.