Aerial and Satellite Image Processing Applications

Aerial and satellite images are used extensively in many trip and mapping software packages, such as Google Earth etc. Aerial images differ significantly due to the variations in atmospheric clarity, atmospheric layer density, humidity, temperature, and season under which the aerial images are captured. Among the issues in effective utilization of aerial images, color tone discrepancy is the inconsistency in brightness, saturation, or color balance between images representing adjacent areas, causing adjacent areas appear significantly different that would otherwise be similar. Aenal images are also routinely used as ground textures to simulate Earth surface and 3D features on Earth in terrain visualization, and flight simulations. In these applications satellite images of different regions are not captured in the same season and hence they exhibit substantial variations due to seasonal color changes of vegetation on the ground, causing the trainee to witness widely differing and constantly changing level of fidelity. Therefore m many applications such as flight simulations it is desired that, satellite images captured in the same season be used instead of the images based on different seasons. This work presents two techniques; one to minimize the color tone discrepancies between adjacent areas in aerial and satellite images, and second to simulate seasonal color adaptation of vegetation in aerial and satellite imagery. First part of this research work presents two algorithms that alter the color tone of aerial and satellite images, resulting adjacent pairs look similar. Proposed algorithm I utilizes a cost function approach and is suitable for images that have overlapped areas. First, a single parameter is β utilized to model the attenuated incident light on the captured image. The second model assumes maximum incident light, but models of the atmospheric effects using two parameters c and IA. The atmosphere color caused by scattering is represented by IA. The attenuation caused by atmospheric absorption is modeled by a parameter c. The third model utilizes the three parameters β, c and IA to enhance the target image. Proposed algorithm II is based on statistical estimation and is suitable for arbitrary images representing adjacent areas. The second part proposes a novel method to model the seasonal color variations in vegetation using image analogy technique. The proposed method first generates a vegetation map for pixels corresponding to vegetative areas, using ISODATA clustering and vegetation classification. It then generates seasonal color adaptation of a target input image based on a pair of training images, which depict the same area but were captured in different seasons, using image analogies technique. The vegetation map ensures that only the colors of vegetative areas in the target image are altered and also improves the performance of the original image analogies technique. An interesting and worth mentioning application of image analogies technique, the color night vision is also presented.