Self-supervised monocular depth estimation method based on deep learning

Abstract

The invention discloses a self-supervision monocular depth estimation method based on deep learning, and the method comprises the steps: respectively extracting the pyramid features of an original right view Ir and a synthesized left view, carrying out the horizontal correlation operation of the pyramid features to obtain a multi-scale correlation feature Fc, and obtaining a completed multi-scalecorrelation feature Fm; sending the Fm to a visual clue prediction network in a binocular clue prediction module, generating an auxiliary visual clue Dr, reconstructing a right view from the synthesized left view, and optimizing the binocular clue prediction module through image reconstruction loss between the reconstructed right view and a real right view Ir; using a visual clue Dr generated by the binocular clue prediction module for constraining a disparity map D1 predicted by the monocular depth estimation network, and enhancing consistency between the monocular depth estimation network and the disparity map D1 by using consistency loss; and constructing constraints of occlusion guidance to allocate different weights for reconstruction errors of occluded area pixels and non-occluded area pixels.

Description

technical field [0001] The invention relates to the fields of computer vision and depth estimation, in particular to a self-supervised monocular depth estimation method based on deep learning. Background technique [0002] As one of the basic tasks of computer vision, depth perception can be widely used in areas such as autonomous driving, augmented reality, robot navigation, and 3D reconstruction. Although active sensors (e.g., lidar, structured light, and time-of-flight) have been widely exploited to directly acquire scene depth, active sensor devices are usually bulky, expensive, and have high energy consumption. In contrast, the method of predicting depth based on RGB (color) images has the advantages of low price and easy implementation. Among existing image-based depth estimation methods, monocular depth estimation does not rely on multiple acquisitions of the perceived environment, and has received extensive attention from researchers. [0003] In recent years, deep...

AI Technical Summary

Problems solved by technology

However, existing methods usually only focus on utilizing synthesized target views to construct supervisory signals, and do not fully explore and exploit the geometric correlation between source views and synthesized target views.

Moreover, due to the occlusion between the source and target views, existing methods directly minimize the appearance difference between the synthesized target view and the real target view in disparity learning, which will lead to inaccurate predicted disparities near occluded regions.

Images