Learn to have Color and Detail: An End-to-End Panchromatic Image Enhancement
Minjian Zhou, Yuxuan Wang, Guangming Wu, Ryosuke Shibasaki · 2021
Due to the limited resolution and chrominance information, panchromatic images can not be widely used in accurate earth observation applications, such as road extraction, vehicle detection, and building segmentation. In this research, we propose a cascaded fully convolutional network (CFCN) to achieve panchromatic image super-resolution and image colorization in an end-to-end manner. Experiments on a multispectral image dataset demonstrate that panchromatic images enhanced by the proposed CFCN can achieve high learned extraction similarity as compared to aerial images.