论文部分内容阅读
针对分形网络演化多尺度分割方法对高分辨率遥感影像存在的欠分割问题,文章提出以KL散度为核心的区域合并标准,研究出一种基于KL散度原理的改进分形网络演化多尺度分割方法。该方法相对于分形网络演化多尺度分割方法,在合并标准上更充分利用了遥感影像的光谱特征,能够很好地消除因光谱差异造成的欠分割,对同质性较强的地物效果尤为明显。通过WorldView-2高分辨率影像分割实验,比较欠分割率,结果证明该方法更适合于高分辨率遥感影像分割,能够为地物信息提取提供与自然更吻合的实体对象。
In order to solve the problem of under-segmentation of high-resolution remote sensing images based on the multi-scale segmentation method of fractal network evolution, this paper proposes a regional consolidation criterion with KL divergence as the core, and develops an improved multi-scale segmentation based on KL divergence principle method. Compared with the fractal network evolution multi-scale segmentation method, this method makes full use of the spectral features of the remote sensing images in the merger standard, which can well eliminate the under-segmentation due to spectral differences, especially for the strong homogeneous objects obvious. Through WorldView-2 high-resolution image segmentation experiment, the under-segmentation ratio is compared. The results show that this method is more suitable for high-resolution remote sensing image segmentation, and can provide physical objects more consistent with nature.