Depth image application in analysis of automatic 3D reconstruction

In this paper, we present an algorithm which can reconstruct arbitrary scene in real-time automatically by using depth image. ICP algorithm[3] will be used to estimate the pose of the depth sensor, after image was taken from depth sensor. Depth data will be fused into global model using signed distance function[4]. Finally, according to the Marching Cubes[2] algorithm, the model will be presented in real-time automatically. It is worth mentioning that practicing this algorithm, we just need to fix a commodity depth sensor on the robotic hand, and provide a trajectory for the robot to scan the scene, and the model will be generated in real-time. In the last part, experimental result will be shown. Overall, we can make a conclusion that using depth sensor to reconstruct scene in realtime is feasible. We believe that this method will be an excellence solution of simultaneous localization and mapping.