| Abstract:Monocular depth estimation is an important research direction in the field of computer vision, which plays an important role in 3D reconstruction, target tracking, scene understanding and many other appliactions. Due to the characteristics of monocular camera, such as low cost, more popular equipment, and easy image acquisition, the topic of depth information acquisition from monocular images has become a popular research topic. Firstly, the common deep learning models for monocular depth estimation are summarized, mainly including CNN, RNN and GAN. Then, the deep learning methods for monocular depth estimation are summarized from the perspective of training methods, which are mainly divided into supervised learning, unsupervised learning and semi-supervised learning. Finally, the development trend of monocular depth estimation is summarized in this paper. |