A novel method of destriping for airborne hyperspectral image

Airborne hyperspectral image often suffers from stripe noises, which greatly affect the visual interpretation and radiometric integrity of the image. To reduce the stripe noises, a new method of destriping was developed, by combining outdoor calibration with image statistic method. In this method, outdoor calibration is conducted to get the relative radiometric calibration coefficients, which are used for the first correction of raw image. Then image statistic information is used to refine the calibration coefficients and do a second correction. The experimental results reveal that the proposed method can effectively remove stripe noises, and has better effect than either calibration method or image statistic method individually. It provides a simple, effective, practical way for the destriping of airborne hyperspectral image and can also be applied in the destriping of other types of remote sensing images.