Fast Colour Balance Adjustment of IKONOS Imagery Using CUDA

With the advancement of sensor technology, the image resolution of optical satellites improves, increasing image size in the process. For satellite imagery mosaic composition, the task of performing colour balancing becomes more intensive due the sheer size of data needed to be processed. In order to reduce the processing time for color balancing, the proposed procedure takes advantage of the graphics processor unit's data-parallel capabilities using compute unified device architecture (CUDA). By repartitioning the image data and modifying the conventional colour balancing algorithm to suit CUDA's requirements, we are able to cut down the processing time to a fraction of the original time taken by conventional methods, showing CUDA's capabilities in remote sensing image processing applications.