Image-Based ICP Algorithm for Visual Odometry Using a RGB-D Sensor in a Dynamic Environment

摘要: This paper proposes a novel approach to calculate visual odometry using Microsoft Kinect incorporating depth information into RGB color generate 3D feature points based on speed up robust features (SURF) descriptor. In particular, the generated are used for calculating iterative closest point (ICP) algorithm between successive images from sensor. The ICP works image of differently previous approaches. suggests one modified versions state-of-the-art implementation algorithm. Such an makes accurate calculation rigid body transformation matrix in dynamic environment. From this step, dynamically moving can be separated outliers. Then, outliers filtered with random sample consensus (RANSAC) matrix. experiments demonstrate that is successfully obtained proposed