Coupled Global-Local object detection for large VHR aerial images

Xi Chen, Chaojie Wang, Zhihong Li, Min Liu, Qingli Li, Honggang Qi, Dongliang Ma, Zhiqiang Li, Yong Wang

Abstract: Highlights•Global & local branches fuse information in a layer and merge detection results.•Kernels are adaptively and easily chosen to find better receptive fields.•Proposal filters screen qualified global proposals to extract better features.•The unqualified objects are efficiently suppressed on global branch.•Our method captures more contextual information and yields superior results.