摘要
为实现高原山区现代烟草定量监测,针对贵州高原山区天气多云雨,导致光学数据难以获取,提出一种利用高分辨率合成孔径雷达(SAR)应用于烟草快速监测的方法。结果表明,利用不同地物后向散射特征差异,采用支持向量机分类方法提取烟草种植分布情况,分类总精度达到了85.08%,烟草识别精度达到了78.36%;运用回归分析,将烟草生长参数与后向散射系数建立监测模型,所建立的叶长、叶宽都与HH极化后向散射系数,叶片鲜重与HH、VV极化后向散射系数的模型都能反映其线性回归的耦合关系。表明高分辨率SAR数据为数据源的新型遥感监测手段能提供丰富的高原山区典型地物的后向散射信息,能为现代烟草农业大范围实时遥感监测提供技术支持。
In order to carry out quantitative monitoring of modern tobacco in plateau mountain area, a new method was proposed inwhich high resolution synthetic aperture radar （SAR） method was applied to rapidly monitor tobacco, since weather in Guizhouplateau is too rainy to obtain optical data. The results showed that the after scattering feature of different ground objects wasextracted by difference, and used classification method of SVM to classify ground objects, the accuracy of total classification was85.08%, and tobacco identification accuracy was 78.36%. Using of regression analysis, the monitoring model between the tobaccogrowth parameters and the backscattering coefficient was established. The model which was between the leaf length/leaf width andHH polarization backscatter coefficient or leaf fresh weight and HH, VV polarization backscattering coefficients could reflect thelinear regression equation. The high resolution SAR data as the data source of the new remote sensing techniques can provide richbackscatter information of typical objects in plateau mountainous area, and provide technical support for modern tobacco agriculturein large-scale and real-time monitoring by remote sensing.
出处
《中国烟草科学》
CSCD
2014年第6期74-79,共6页
Chinese Tobacco Science
基金
国家重点基础研究发展计划(973计划)课题"人为干预下喀斯特山地石漠化的演变机制与调控"(2012CB723202)
贵州省科技计划"喀斯特山区SAR遥感平台监测与识别关键技术研究与应用"{黔科合GY字〔2013〕3062}
贵州省国际科技合作计划"不同生态恢复措施干预下喀斯特石漠化演变及调控研究"{黔科合外G字[2012]7022号}
关键词
SAR
后向散射系数
遥感监测
烟草
SAR
backscattering coefficient
remote sensing monitoring
tobacco
