An Augmented Around View Monitor System Fusing Depth And Image Information During the Reversing Process 2020-01-0095
The Around view monitor (AVM) system for vehicles usually suffers from the distortion of surrounding objects caused by incomplete rectification and stitching, which seriously affects the driver's judgment of the surrounding environment during the reversing process. In response to solve this problem, an augmented around view monitor (AAVM) system is proposed. First, an around view image is generated from four fisheye cameras, and a single depth camera calibration is conducted to improve its accuracy of depth estimation. After that, a DCNN detector is used to detect objects and localize the corresponding point cloud, like pedestrians or vehicles, in the reflectance image of depth camera. Then, the reconstruction of objects in the depth camera coordinate is performed using the detection results and information of the corresponding point cloud. Next, a joint calibration method of multiple depth cameras is adopted to establish a vehicle-centric coordinate. Finally, the around view image, the object model, and the car model are unified in the same vehicle-centric coordinate, which will be projected at a suitable angle of view so as to easy to observe. The field experiments are conducted using a passenger car with 4 fisheye cameras and 3 depth cameras. The results show that the AAVM can help the driver to effectively judge the real environment behind the vehicle during the reversing process.
WEI LI, Wenbo Wang, Lu Jiang, Dongjie Xiao, Libo Cao