MCC-DEM: Multi-Scale Context Constraint-Based Depth Estimation Method for Monocular Camera

Monocular depth estimation is the task of estimating the depth value of each pixel given a monocular RGB image. For the obstacle avoidance and map building of mobile robots, the requirement of accuracy of monocular depth estimation is engaged. In this work, we propose a monocular depth estimation network based on an encoder-decoder structure, to obtain accurate depth estimation. An improved atrous spatial pyramid pooling module is used to extract the multi-scale context information, in which structural similarity error is used to optimize the loss function and better reflect the structural cues of an object. Experimental results show that the proposed method outperforms state-of-the-art methods on both a public dataset and a collected dataset in terms of overall performance.