遥感监测适用性分析

Research on the Applicability of Remote Sensing Monitoring of Inedible Grass Pedicularis sp. by GF-1 WFV Satellite in Bayanbulak Grassland

【目的】验证GF-1 WFV(16 m分辨率)在不食草马先蒿遥感监测的适用性,了解巴音布鲁克草原不食草马先蒿的分布。【方法】利用校正后的GF-1 WFV遥感卫星影像,结合实测光谱数据,采用目视解译结合监督分类的方法,对研究区的不食草马先蒿进行识别,验证其分类精度,并对研究区马先蒿的分类精度和分布情况进行分析。【结果】影像解译整体精度为80.91%,Kappa系数为0.71,分类结果较好;得到了巴音布鲁克草原不食草马先蒿分布范围和影响分类精度的主要原因。【结论】利用16 m分辨率GF-1 WFV多光谱遥感数据对巴音布鲁克草原不食草马先蒿进行遥感监测是可行的。通过对研究区遥感解译的分类结果和实地踏查发现,不食草马先蒿危害面积较大,且多为连片分布、顺河流分布;危害较为严重的区域为巴音布鲁克总场巴音布鲁克分场、巴音郭楞乡哈尔萨拉村、巴音布鲁克镇巴西里克村。

[ Objective] To validate the applicability of remote sensing monitoring of inedible grass ularis sp. by GF -1 WFV （16m ） satellite and understand the inedible grass Pedicularis s distribution of the study area in Bayanbulak grassland.[Method ]The corrected GF - 1 WFV remote sensing satellite images were combined with the measured spectral data. The method of visual interpretation was applied combined with supervised classification of the inedible grass Pedicularis sp. identification to verify its classification accuracy. And then the accuracy of classification and distribution of Pedicularis sp. were discussed.[ Result] The result of accuracy evaluation indicated that the overall accuracy was 80. 9 1 % , and the Kappa Coefficient was 0. 71. The classification result was better. Thus the distribution range of the poisonous grass Pedicularis sp. in Bay-anbulak grassland and the main reasons affecting the classification accuracy were obtained. [ Conclusion] The result showed that the applicability of remote sensing monitoring of inedible grass Pedicularis sp. by GF - 1 WFV （16 m） satellite in Bayanbulak grassland was feasible. The classification results of the study area of re-mote sensing interpretation and field survey found that the damaged area of grass Pedicularis sp. area is rather large,and most of them are contiguous and distributed along the river. The seriously damaged areas were Bay-anbulak parvial field of the Bayanbulak total f ie ld , Hal Sara village of Bayingolin Township, the Gabriele Ba- silico village of Bayanbulak Town.