Visual Depth Mapping from Monocular Images using Recurrent Convolutional Neural Networks