Unsupervised Learning of Visual Odometry with Depth Warp Constraints

2019 International Conference on Advanced Electronic Materials, Computers and Materials Engineering

无线电电子学;计算机科学;材料科学

Shi, Haibin^1 ; Guo, Menghao^1 ; Xu, Zhi^1 ; Zou, Yuanbin^1

College of Information Science and Engineering, Northeastern University, Shenyang

110819, China^1

关键词: Cascade networks;    End to end;    Geometric method;    Network learning;    Supervised methods;    Unsupervised method;    Visual odometry;    Visual SLAM;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/563/4/042024/pdf
DOI  :  10.1088/1757-899X/563/4/042024

来源: IOP

【 摘 要 】

Visual Odometry (VO) is one of the important components of Visual SLAM system. Some impressive work on the end-to-end deep neural networks for 6-DoF VO has appeared. We propose two-part cascade network structure to learn depth from binocular image and to infer ego-motion from consecutive frames. We propose depth warp constraints to make the Network learning more geometrically information. A lot of experiments on KITTI data set show that our model is superior to previous unsupervised methods and has comparable results with the supervised method, verifying that such a depth warp constraints perform successfully in the unsupervised deep method which is an important complement to the geometric method.