摘要
为实现滑坡遥感影像的自动提取,以贵州省水城县为研究区,运用GF-1遥感影像数据,剔除明显非滑坡地物后,应用云变换计算样本区影像像元亮度值,统计得到聚类中心值和聚类个数,结合快速广义模糊C-均值聚类算法进行像元聚类,提取出128处准确滑坡,kappa系数达0.752,总体精度达0.880。该方法可自动提取出滑坡范围,减少工作量和主观影响,提高了基于遥感影像进行滑坡识别的效率和精度。
In order to realize the automatic extraction of landslide from remote sensing images,Shuicheng County,Guizhou Province,was taken as the research area.Using the GF-1 remote sensing image data and removing the obvious non-landslide features,the image digital number in the sample area was calculated by cloud trasforation,and the cluster center value and cluster number were obtained statistically.And combined with the fast generalized fuzzy C-means clustering algorithm for pixel clustering,128 accurate landslides were extracted.The kappa coefficient was 0.752,and the overall accuracy was 0.880.This method can automatically extract the landslide range,reduce the workload and subjective influence,and improve the efficiency and accuracy of landslide recognition based on remote sensing images.
作者
谭秋焰
吴彩燕
贾菊桃
朱新婷
廖军
TAN Qiuyan;WU Caiyan;JIA Jutao;ZHU Xinting;LIAO Jun(School of Environment and Resource,Southwest University of Science and Technology,Mianyang 621010,Sichuan,China;Tianfu Institute of Research and Innovation,Southwest University of Science and Technology,Chengdu 610299,Sichuan,China;Mianyang S&T City Division,the National Remote Sensing Center of China,Mianyang 621010,Sichuan,China;Sichuan Academy of Safety Science and Technology,Chengdu 610000,Sichuan,China)
出处
《西南科技大学学报》
CAS
2023年第1期54-60,共7页
Journal of Southwest University of Science and Technology
基金
国家自然科学基金项目(41301587)
四川省科技厅项目(2020YFS0389)
第三次新疆综合科学考察项目(2021xjkk1404)。
关键词
滑坡
遥感影像
快速广义模糊C-均值聚类算法
云变换
Landslide
Remote sensing images
Fast generalized fuzzy C-means clustering algorithm
Cloud transformation
