Self-Supervised Monocular Depth Estimation Using Global and Local Mixed Multi-Scale Feature Enhancement Network for Low-Altitude UAV Remote Sensing

Article Computer Science, Artificial Intelligence

Unsupervised Scale-Consistent Depth Learning from Video

Jia-Wang Bian et al.

Summary: This study introduces a monocular depth estimation method SC-Depth, using unlabelled videos for training and achieving scale-consistent prediction at inference. Contributions include a novel geometry consistency loss and self-discovered mask, demonstrating efficacy with detailed ablation studies and high-quality depth estimation results in various datasets.

INTERNATIONAL JOURNAL OF COMPUTER VISION (2021)

Add to Collection

Article Geography, Physical

Self-supervised monocular depth estimation from oblique UAV videos

Logambal Madhuanand et al.

Summary: This paper introduces the technique of depth estimation using UAV aerial images, training deep learning models on video frames, and proposing a self-supervised learning method to optimize the accuracy of depth estimation. Experimental results show that the proposed model outperforms existing methods in estimating depths.

ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Channel-Wise Attention-Based Network for Self-Supervised Monocular Depth Estimation

Jiaxing Yan et al.

Summary: In this paper, the Channel-wise Attention-based Depth Estimation Network (CADepth-Net) is proposed to address the lack of modeling of scene structure and handling of details information in self-supervised learning for monocular depth estimation. The two effective modules in CADepth-Net enhance the perception of scene structure and emphasize local details information, leading to state-of-the-art results on benchmarks such as KITTI and Make3D datasets.

2021 INTERNATIONAL CONFERENCE ON 3D VISION (3DV 2021) (2021)

Add to Collection

Article Geography, Physical