摘要
库尔勒香梨是一个地域性极强的名优特优品种,是库尔勒地区乃至全国最优异的地方梨品种之一,利用遥感技术快速、准确获取库尔勒市香梨的种植面积、长势等信息,可为政府管理部门制定各种决策提供数据支持。本研究利用高分一号(GF-1,分辨率2 m和16 m)、Landsat 8(分辨率30 m)2015年4—10月的库尔勒市遥感影像,依据地面调查建立的感兴趣区域(ROI),采用5种不同方法进行监督分类,通过识别精度比较筛选出最佳识别方法,并对库尔勒市的香梨种植面积进行识别、提取。结果表明,GF-1影像数据的最佳识别月份为8月,最佳识别方法是最大似然法,识别精度为95.94%;Landsat 8影像数据的最佳识别月份为10月,最佳识别方法也是最大似然法,识别精度为94.43%。因此,选择利用8月份的GF-1影像数据提取库尔勒市的香梨栽植面积,经最大似然法进行监督分类,最终解译面积为2.827万公顷,与实际统计资料相比,识别精度96.45%,说明利用该方法监测库尔勒市的香梨栽植情况可行。
Korla fragrant pear is a well-known and excellent variety with strong regionalism,and it is one of the most excellent local pear cultivars in Korla and even in China. With the remote sensing technology,we can quickly and exactly gain informations of Korla fragrant pear including planting area and growth vigor,which could provide data support for government administration section making decision. Using the Korla's remote sensing images of GF-1( resolving power as 2 m and 16 m) and Landsat 8( resolving power as 30 m)from April to October of 2015,according to the region of interest( ROI) set by ground survey,adopting 5 different supervised classification methods,the best recognition method was screened out with recognition accuracy,and the Korla fragrant pear planting area was recognized and extracted. The results showed that the best recognition month of GF-1 image was August,and the best recognition method was maximum likelihood method,and its recognition accuracy reached 95. 94%; the best recognition month of Landsat 8 image was October,the best recognition method was also maximum likelihood method,and the recognition accuracy reached 94. 43%. With the GF-1 image in August,supervised classification by the maximum likelihoodmethod,the final planting area of Korla fragrant pear was extracted as 28. 27 thousand hectare,and the recognition accuracy could reach 96. 45% compared with the practical statistics,which showed the method was suitable for supervising the planting situation of fragrant pear in Korla.
出处
《山东农业科学》
2018年第1期153-157,共5页
Shandong Agricultural Sciences
基金
"十二五"国家科技支撑计划课题"西部特色作物科技信息服务互动平台构建与示范"(2014BAD10B01)
新疆自治区科技支疆项目计划(指令性)项目(2016E02083)
