面向农业区划的作物种植结构遥感提取

发布时间：2018-05-11 04:41

本文选题：作物种植结构 + 面向对象分割　；参考：《中国农业资源与区划》2017年08期

【摘要】：[目的]尝试将作物种植结构提取结果由行政单元发展为基于相对均质的地理网格单元,解决当前农作物种植结构信息提取的空间局限性问题,文章在大尺度的土地利用/覆被分类与地块尺度的作物分类之间提出了作物种植结构单元概念,并构建了一种快速、低成本、准确的区域尺度作物种植结构提取方法。[方法]利用黑龙江省2014年250m分辨率的植被指数产品构建时间序列曲线提取物候信息,在耕地物候分区基础上对各物候区进行面向对象的多尺度分割,提取作物种植结构单元,利用光谱特征和NDVI指数构建特征空间,最终采用最邻近分类方法提取作物种植结构。[结果](1)利用MODIS时间序列数据提取物候特征进行多尺度分割的方法,能够有效的提取区域尺度农作物种植结构单元;(2)作物种植结构提取总体精度为95.70%;(3)黑龙江省2014年作物种植类型共有12种。其中,三江平原主要是水稻单一种植区、水稻混作区;松嫩平原以玉米单一种植区以及玉米-大豆混作区种植为主;西北部种植结构较复杂;东南部因地势等影响多种植玉米、大豆。[结论]利用物候数据进行种植结构提取可以有效划分农业区划,研究成果不仅为作物种植结构调整和农业发展布局提供科学依据,也是不同区域产量预测的基础,为合理布局农业生产、改进耕作制度以及引入和推广新产品等提供依据。
[Abstract]:[objective] to develop the result of crop planting structure extraction from administrative unit to geographical grid unit based on relative homogeneity, so as to solve the problem of spatial limitation of current crop planting structure information extraction. In this paper, the concept of crop planting structure unit is proposed between large scale land use / cover classification and land scale crop classification, and a rapid, low cost and accurate extraction method of crop planting structure on regional scale is constructed. [methods] the time series curve extract phenological information was constructed by using the 250m resolution vegetation index product of Heilongjiang province in 2014. Based on the phenological partition of cultivated land, the multi-scale segmentation of each phenology area was carried out, and the crop planting structure unit was extracted. The spectral feature and NDVI index were used to construct the feature space and the nearest neighbor classification method was used to extract crop planting structure. [results] A multi-scale segmentation method based on the extraction characteristics of MODIS time series data was used. The total precision of crop planting structure extraction is 95.70) there are 12 crop planting types in Heilongjiang Province in 2014. Among them, the Sanjiang Plain is mainly a rice monoculture area, where rice is mixed; the Songnen Plain is mainly planted in a maize monoculture area and a corn-soybean mixed cropping area; the planting structure in the northwest is more complicated; and the southeast area has more maize planting due to the influence of topography and so on. Soybeans. [conclusion] using phenological data to extract the planting structure can effectively divide the agricultural regionalization. The research results not only provide scientific basis for crop planting structure adjustment and agricultural development layout, but also serve as the basis for yield prediction in different regions. It provides basis for rational distribution of agricultural production, improvement of farming system and introduction and promotion of new products.
【作者单位】：东北农业大学资源与环境学院;
【基金】：国家自然科学基金项目“基于土壤不同理化参数光谱特征差异的黑土有机质高光谱遥感反演研究”(40801167) 黑龙江省普通高等学校新世纪优秀人才培养计划“松嫩平原北部主要土壤有机质光谱速测研究”(1254-NCET-002) 黑龙江省自然科学基金“区域土壤有机质野外高光谱遥感预测模型研究”(D201404)
【分类号】：S127

【相似文献】