Investigating the Distinctive Character of Intensity Values in Spatially Enhancement of Historical Aerial Photographs Using Adaptive Gamma Correction

Due to the lack of color information in historical aerial photographs, each pixel is expressed in grayscale depending on its intensity value. This representation is one of the obstacles to super-resolution enhancement. In this study, the effect of pixel intensity value range on the spatially enhanced historical aerial photographs was investigated, and adaptive gamma correction was proposed as a data augmentation technique. Three datasets were produced as the dataset in which the pixel intensity value range is kept limited, the dataset in which the pixel intensity value range is kept wide, and the dataset in which the pixel intensity value range is kept wide assisted by adaptive gamma correction. Test images enhanced with super-resolution were comparatively analyzed via image quality metrics. The AGC-assisted approach gave better results in the lower-intensity versions of the test images. This situation was much more evident at higher scale factors. The trends of pixel-based and structural similarity-based metrics were similar to each other. However, the high darkness found in lower-intensity images did not allow the progression in metric values to be supported by visual interpretation. In general, the findings revealed that intensity values in historical aerial photographs have a distinctive characteristic in terms of enhancement.