Abstract: The shadow on the high resolution remote sensing image influences the image quality badly.An automatic property-based approach is presented for shadow removal of color image.The RGB image is first transformed into HSI color space,according as shadow regions has lower luminance and higher saturation than non-shadowed regions,threshold detection and small regions removal and morphological algorithm are performed to get accurate shadow mask.Then the shadow regions are matched to their adjacent non-shadowed regions on I component,S component and H component respectively,and by inverse HSI transform to finish shadow removal.The experiment result shows that this method can remove the influence of shadow effectively and does not change the pixel value of non-shadowed regions.