Depth Completion with RGB Prior

TL;DR

This paper introduces a deep learning model for depth correction in RGBD images tailored for industrial environments, addressing reflection issues with a new dataset and improving depth accuracy in challenging settings.

Contribution

The paper presents a novel deep model for depth correction in industrial RGBD images and provides a new publicly available dataset for training and evaluation.

Findings

Improved depth accuracy in reflective industrial environments

Effective correction of depth distortions caused by reflections

Public dataset enables further research in industrial depth perception

Abstract

Depth cameras are a prominent perception system for robotics, especially when operating in natural unstructured environments. Industrial applications, however, typically involve reflective objects under harsh lighting conditions, a challenging scenario for depth cameras, as it induces numerous reflections and deflections, leading to loss of robustness and deteriorated accuracy. Here, we developed a deep model to correct the depth channel in RGBD images, aiming to restore the depth information to the required accuracy. To train the model, we created a novel industrial dataset that we now present to the public. The data was collected with low-end depth cameras and the ground truth depth was generated by multi-view fusion.