摘要
[目的]研究基于无人机低空遥感影像的面向对象分类技术在开发建设项目水土保持监测中的应用,为水土保持监测工作的信息化能力提升提供技术支撑。[方法]利用旋翼无人机获取水土保持监测目标区域的低空遥感影像,通过倾斜摄影技术构建数字表面模型,结合ESP分割尺度评价工具获取最优分割尺度参数,采用多元特征空间指标参与最邻近分类法的监督分类,并依据位置信息的评价方法和误差矩阵对分类解译精度进行验证。[结果]本研究的水土保持监测目标区域的地物分类总体精度达到了86.10%,Kappa系数为0.841,有较好的一致性,能够满足精度需求。[结论]利用无人机低空遥感影像的面向对象分类技术实现了开发建设项目水土保持监测区域地物的快速、精确识别和分类。
[Objectives]With the object-oriented classification techniques for low-altitude unmanned aerial vehicle（UAV）remote sensing image data,this study monitored soil and water conservation for development and construction projects,to provide technical supports for soil and water conservation monitoring.[Methods]Unmanned multi-rotor aircraft was used to obtain the low altitude remote sensing images in soil and water conservation monitoring target areas.Digital surface models were constructed by oblique photography.The optimal segmentation scale parameters were selected by the estimating the scale parameter（ESP）segmentation scale evaluation tool,and supervised classification by the nearest neighbor classification of multivariate feature space metrics was used.The classification accuracy was verified through the location information validation method and error matrix.[Results]The total accuracy and the Kappa coefficients of target area in monitoring the soil and water conservation were 86.10%and 0.841,respectively.This result could meet the precision requirements.[Conclusion]The object-oriented classification techniques used for low-altitude UAV remote sensing image data achieved fast,accurate identification and classification in the monitoring of soil and water conservation for development and construction projects.
作者
周湘山
秦甦
魏凡
戴松晨
张磊
周杰
詹晓敏
ZHOU Xiangshan;QIN Su;WEI Fan;DAI Songchen;ZHANG Lei;ZHOU Jie;ZHAN Xiaomin(Chengdu Engineering Corporation Limited, Chengdu, Sichuan 610041, China)
出处
《水土保持通报》
CSCD
北大核心
2018年第3期130-135,共6页
Bulletin of Soil and Water Conservation
基金
中国电建集团成都勘测设计研究院有限公司青年基金计划项目"水电水利项目水土保持监测近景摄影测量技术及应用研究"(P298-2015)
关键词
水土保持监测
无人机低空遥感
面向对象分类
最优分割尺度
soil and water conservation monitoring
low-altitude UAV remote sensing
object-oriented classification
optimal segmentation scale
