Inspired by 2D GrabCut for still images, this paper proposes automated abandoned object segmentation by introducing 3D GrabCut in surveillance scenario. This allows the method to produce precise object extraction without needing user intervention. Both RGB and depth input are utilized to build abandoned object detector which can resist to shadow and brightness changes. We performed the indoor experiments to show that our system obtains an accurate detection and segmentation. Both quantitative and qualitative measurements are provided to analyze the result.