2019/03/15 by Yilun Zhang, Zhang, Yilun, Ty Nguyen +12 · 15 citations
Computer Science · Engineering · #Advanced Vision and Imaging #Artificial intelligence #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Computer vision #Depth map #FOS: Computer and information sciences #Geology #Image (mathematics) #Measured depth #Optical measurement and interference techniques #RGB color model #Robotics (cs.RO) #Robotics and Sensor-Based Localization #cs.CV #cs.RO
paper · pdf · doi:10.48550/arxiv.1903.06397
published in arXiv (Cornell University) (Cornell University)
openalex publication_date 2019/03/15 · openalex created_date 2019/04/11 · arxiv created 2019/08/14 · arxiv updated 2019/08/15 · openalex updated_date 2026/07/28
Depth estimation is an important capability for autonomous vehicles to understand and reconstruct 3D environments as well as avoid obstacles during the execution. Accurate depth sensors such as LiDARs are often heavy, expensive and can only provide sparse depth while lighter depth sensors such as stereo cameras are noiser in comparison. We propose an end-to-end learning algorithm that is capable of using sparse, noisy input depth for refinement and depth completion. Our model also produces the camera pose as a byproduct, making it a great solution for autonomous systems. We evaluate our approach on both indoor and outdoor datasets. Empirical results show that our method performs well on the KITTI~\citekittigeiger2012we dataset when compared to other competing methods, while having superior performance in dealing with sparse, noisy input depth on the TUM~\citesturm12iros dataset.