摘要
针对土地利用变化中局部像斑变化较难提取和自动化监测较难实现等问题,该文提出了一种基于类别的矢量图与遥感影像变化检测方法。在矢量图约束下,对遥感影像进行影像分割获取像斑;提取像斑在遥感影像上的直方图特征,采用G统计量度量像斑之间的特征距离;利用像斑与其他相同类别像斑之间的特征距离,构建单波段上像斑的类别异质度,自适应加权组合各波段上像斑的类别异质度构建像斑的类别异质度;依据最大熵方法获取各地物类别对应的异质度阈值,以类别为单位对各像斑进行变化判别,获取变化检测结果。在GF-2号卫星影像上的实验验证了该方法的有效性,实现了矢量图与遥感影像的自动变化检测。
For the problems of the difficult extraction of local image spot changes and the difficult implementation of automatic monitoring in land use changes,this paper proposes a classification-based change detection method for vector images and remote sensing images.Under the constraint of vector images,this paper performs the image seg‐mentation of remote sensing images to obtain image spots,extracts the histogram features of image spots on remote sensing images,uses G statistics to measure the feature distance among image spots,uses the characteristic distance between image spots and other similar types of image spots to construct the category heterogeneity of image spots on a single wavelength band,constructs the category heterogeneity of image spots by the adaptive weighted combina‐tion of the category heterogeneity of image spots on each wavelength band,uses the maximum entropy method to obtain the heterogeneity threshold corresponding to each ground feature category,and performs change discrimina‐tion on each image spot by the category to obtain of change detection results.Experiments on GF-2 satellite images have verified the effectiveness of the method and achieved the automatic change detection of vector images and re‐mote sensing images.
作者
曾灵芝
ZENG Lingzhi(Hezhou Bureau of Natural Resources,Hezhou,Guangxi Zhuang Autonomous Region,542899 China)
出处
《科技资讯》
2023年第20期176-179,共4页
Science & Technology Information
关键词
矢量图
变化检测
像斑
类别异质度
Vector graph
Change detection
Image spot
Category heterogeneity
