MONOCULAR DEPTH ESTIMATION IN FOREST ENVIRONMENTS

Abstract. Depth estimation from a single image is challenging task, especially inside the highly structured forest environment. In this paper, we propose supervised deep learning model for monocular depth based on imagery. We train our new data set of RGB-D images that collected using terrestrial laser scanner. Alongside input RGB image, uses sparse channel as to recover dense information. The prediction accuracy significantly higher than state-of-the-art methods when applied in context estimation. Our brings RMSE down 2.1 m, compared 4 m and above reference methods.