An Operational Improvement of Haze/Clear Line Identification from Satellite Imagery Based on Multi-Resolution Segmentation

摘要

Haze removal and atmospheric correction are the important steps prior to the application of imagery. ATCOR is a more effective haze removal and atmospheric correction modeling which combined several improved methods. However, boundary effect along haze/clear lines in the image processed by ATCOR also be needed to overcome. In this paper, an operational improvement of haze/clear line identification based on multi-resolution segmentation was introduced. Firstly, the haze/clear line was generated by ATCOR, and then the image was segmented using multi-resolution segmentation algorithm. Secondly, the haze/clear line generated by ATCOR was intersected with the multi-resolution segmentation, and the final haze/clear line was generated by merging intersection vector. At last, the image was atmospherically corrected by ATCOR. This method combined the haze/clear line result processed by ATCOR and multi-resolution segmentation, and the boundary effect paralleled the haze/clear lines were disappeared obviously.