基于概率距离的高分辨率遥感影像分割结果的优化方法及其应用研究

Research on Optimization Method of High Resolution Remote Sensing Image Segmentation Results Based on Probability Distance and Its Application

不同于传统的通过选择合适的尺度参数来优化分割结果的方法,本文从统计学的角度,提出了一种基于概率距离的高分辨率遥感影像分割结果优化方法,并以Jeffries Matusita距离为例,介绍了利用该方法优化分割结果的流程。该方法通过为影像中的地物构建最优影像对象达到优化分割结果的目的。使用无人机影像对本文提出的优化方法进行了测试,测试结果表明,本文所提出的分割结果优化方法具有较好的效果,优化后的分类精度得到明显提高。

Different from the traditional method of optimizing segmentation results by selecting appropriate scale parameters,this paper proposes a method of optimizing segmentation results of high-resolution remote sensing images based on probability distance from the perspective of statistics.The process of optimizing segmentation results using this method is illustrated by taking Jeffries Matusita distance as an example.This method optimizes the segmentation results by constructing the optimal image object for the ground object in the image.The optimization method proposed in this paper is tested using UAV images.The test results show that the segmentation result optimization method proposed in this article has good results,and the optimized classification accuracy has been significantly improved.