Self-Supervised Monocular Depth Estimation in the Dark: Towards Data Distribution Compensation
This addresses the challenge of nighttime depth estimation for autonomous driving systems, offering a novel approach that avoids the need for night training data.
The paper tackles the problem of unreliable self-supervised depth estimation at night by proposing a method that trains only on day images and uses physical priors to compensate for day-night differences, achieving state-of-the-art results on nuScenes-Night and RobotCar-Night datasets.
Nighttime self-supervised monocular depth estimation has received increasing attention in recent years. However, using night images for self-supervision is unreliable because the photometric consistency assumption is usually violated in the videos taken under complex lighting conditions. Even with domain adaptation or photometric loss repair, performance is still limited by the poor supervision of night images on trainable networks. In this paper, we propose a self-supervised nighttime monocular depth estimation method that does not use any night images during training. Our framework utilizes day images as a stable source for self-supervision and applies physical priors (e.g., wave optics, reflection model and read-shot noise model) to compensate for some key day-night differences. With day-to-night data distribution compensation, our framework can be trained in an efficient one-stage self-supervised manner. Though no nighttime images are considered during training, qualitative and quantitative results demonstrate that our method achieves SoTA depth estimating results on the challenging nuScenes-Night and RobotCar-Night compared with existing methods.