An integrated framework for aerial image restoration
N. M. Kwok, Haiyan Shi · 2015
The restoration of aerial images is a crucial process for exposing ground features of interest to users. This is particularly important in applications concerned with the surveillance or remote sensing of the land. In this work, an efficient framework was developed that aims at providing a higher quality image of balanced tonal rendering, vivid colouring, and object boundary sharpening. These goals were achieved by processing the aerial image through an integration of colour channel mean value exponent alignment, pixel based min-max shifting and scaling, and object edge emphasizing. Real-world aerial images were used in experiments. Satisfactory results, evaluated both qualitatively and quantitatively, were obtained and verified the effectiveness of the proposed method.