利用数学形态学方法,研究与探讨了IKONOS多光谱图像的分割技术.提出一种结合图像边缘特征和纹理特征的混合分割新算法.在高分辨率多光谱遥感图像K-L变换的基础上,采用多尺度多方向形态学梯度算子提取边缘特征.应用数学形态学滤波及局部方差统计特征对图像对象进行标记,最后采用强制最小过程,进行标记控制的分水岭分割.研究结果表明,提出的分割算法优于仅利用边缘特征的分水岭分割算法,同时,该算法能较好地解决分割过程中存在的过分割与欠分割问题,是一种适合高分辨率多光谱遥感图像的分割算法.

Image segmentation has been an mi portant research area in miage analysis and interpretation. An ideal segmentation strategy of remotely sensed data should considerproblems of over-segmentation and under-segmentation smi ultaneously and find a good tradeoffbetween them.In thispaper,miage segmentation for IKONOS multispectral data is investigated by using techniques of mathematical morphology, and a novelhybrid segmentation algorithm is proposed by combining both edge and texture featuresof mi ages. Based on theK-L transform of multispectral data, edge features are detected by morphological multiscale andmultidirection gradient algorithms, and mi age objects aremarked throughmorphological filtering and localvariance features extracting. Finally, themarkercontrolledwatershed algorithm is mi plemented. The results indicate that the performance of the proposed algorithm is superior to the gradientbased watershed segmentation. Moreover, this approach ismore suitable forhigh resolution remotely sensed data to overcome over-segmentation and underseg mentation problems effectively.