论文部分内容阅读
高分辨率遥感影像中地物对象的最优分割是遥感图像理解、地物识别的关键问题,也是有效利用地物空间信息的基础.针对遥感影像中简单地物具有内部光谱均一和外界光谱差异大的特点,提出图斑显著性的概念,并在多尺度分析中的图像分割过程中建立图斑演变的显著性变化曲线,应用显著性变化曲线的极大值和分割过程中图斑之间的包含关系建立尺度序,并以此解决图像分割中的简单地物的最优分割问题,确定可能为地物对象的图斑.并将此原理和方法应用于高分辨率海岸带遥感影像的图像分割中,效果明显.
The optimal segmentation of objects in high-resolution remote sensing images is the key issue in understanding remote sensing images and object recognition, and also the basis for the effective use of the spatial information of the objects.For the simple objects in remote sensing images have the internal spectral uniformity and the external spectral differences Large features, the concept of plaque significance was proposed, and in the process of image segmentation in multiscale analysis, a significant curve of the evolution of the plaque was established. Between the maxima of the saliency curve and the patches in the segmentation process , And then to solve the problem of the optimal segmentation of simple objects in image segmentation and to determine the speckle that may be the object of the object.And this principle and method are applied to the high resolution remote sensing images of coastal zone Image segmentation, the effect is obvious.