An improved algorithm for shadow restoration of high spatial resolution imagery

Abstract

Based on the principle of "static gains and offsets during the process of image data acquirement," an algorithm named "shadow restoration by matching the statistical feature of neighboring pixels" was proposed here. The histogram method used to calculate the threshold and to efficiently separate shadow from water was improved by incorporating into shape factors. By matching the statistical feature of shadow with the pixels of none shadow segments on the same line, this improved method was able to restore weak spectral information in shadow. The results from the restoration experiment using QuickBird imagery of Nanjing city indicated its good performance for shadow restoration which is characterized by a fast speed, high degree of automation and wide application scope.