遥感影像变化检测方法的研究  

Research on remote sensing imagery change detection method

作　　者：陕唐剑 钟宏伟 李林辉[1] SHAN Tangjian;ZHONG Hongwei;Li Linhui(College of Information and Computer Engineering,Northeast Forestry University,Harbin 150040,China)

机构地区：[1]东北林业大学信息与计算机工程学院

出　　处：《智能计算机与应用》2019年第5期198-199,204,共3页Intelligent Computer and Applications

基　　金：2018年大学生创新训练计划项目(201810225186)

摘　　要：图像变化检测一直是数字图像处理研究的热点之一,遥感图像由于具有空间信息,其变化检测更具有重要的意义。变化检测被广泛应用到森林、地质、环保、军事等领域。本文根据哈尔滨市道里区2006年和2016年两个不同时期的遥感影像,根据数据的实际特点进行了预处理,再将影像结合监督学习的方法分解成多个像素组的集合,形成对象,最后进行面向对象的变化检测方法,实验结果表明,此方法较传统的基于像素变化检测精度有所提高。Imagery change detection has always been one of the hotspots in digital image processing.Due to its spatial information,remote sensing imagery change detection has more important significance.Change detection is widely used in forestry,geology,environmental protection,military and other fields.According to the remote sensing image of Daoli District in Harbin in 2006 and 2016,this paper preprocesses the image according to the actual characteristics of the data,then decomposes the image into a set of multiple pixel groups,forms objects,finally carries out object-oriented change detection method.The experimental results show that this method has better accuracy than the traditional method based on pixel change detection.

关 键 词：遥感影像 变化检测 图像变化检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]