Abstract
Depth estimation can provide tremendous help for object detection, localization, path planning, etc. However, the existing methods based on deep learning have high requirements on computing power and often cannot be directly applied to autonomous moving platforms (AMP). Fifth-generation (5G) mobile and wireless communication systems have attracted the attention of researchers because it provides the network foundation for cloud computing and edge computing, which makes it possible to utilize deep learning method on AMP. This paper proposes a depth prediction method for AMP based on unsupervised learning, which can learn from video sequences and simultaneously estimate the depth structure of the scene and the ego-motion. Compared with the existing unsupervised learning methods, our method makes the spatial correspondence among pixel points consistent with the image area by smoothing the 3D corresponding vector field based on 2D image, which effectively improves the depth prediction ability of the neural network. Our experiments on the KITTI driving dataset demonstrated that our method outperformed other previous learning-based methods. The results on the Apolloscape and Cityscapes datasets show that our proposed method has a strong universality.
Highlights
We propose a depth prediction method for autonomous moving platforms (AMP) based on unsupervised learning, which can learn from video sequences and simultaneously estimate the depth structure of the scene and the ego-motion
Since the depth of the object in the image cannot be directly calculated, the simultaneous localization and mapping (SLAM) methods based on monocular camera need to first analyze the relationship of the position of the same points in different views, and estimate the ego-motion and the depth of the points in the scene simultaneously for example [6,7]
To be able to evaluate in the KITTI dataset, we needed to map the laser measurements into the graph space to generate the ground-truth depth corresponding to the original image
Summary
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. We propose a deep prediction method based on unsupervised learning This method trains the neural network by analyzing the geometric constraint relationship of the three-dimensional (3D) scene among sequence of pictures and constraining the correspondence of pixels among images according to image gradient. It can simultaneously predict the depth structure of the scene and the ego-motion. Our method makes the spatial correspondence between pixel points consistent with the image area by smoothing the 3D corresponding vector field by 2D image This effectively improves the depth prediction ability of the neural network. The results of the assessment indicate that our unsupervised method is superior to existing methods of the same type and has better quality than other self-supervised and supervised methods in recent years
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
