摘要
农作物种植结构信息是作物长势监测和农业结构调整的重要参考依据,及时准确地通过土地分类和农作物适宜性评价获取农作物空间分布与适宜性种植信息对农业可持续发展意义重大.该文以河北张家口沽源县为研究对象,利用Sentinel-2多光谱数据提取的NDV,NDBI,NDWI,NDRE1,SR_(re)和CI_(red-edge)数据为特征,分别采用SVM支持向量机、决策树法、随机森林法对研究区内典型农作物进行精细化提取,探究主要作物空间分布情况,并通过对比kappa系数探讨不同方法对农作物分类的精度,选择最优分类方法.选取土壤性质、土壤侵蚀度、高程、坡度、坡向5个指标建立农作物适宜性评价体系,采用GIS层次分析法与土地适宜性分级指标对沽源县农作物适宜性进行评价.结果表明:基于随机森林分类法对研究区内8种主要农作物进行分类的精度最高,其总体准确率为65.10%,Kappa系数为0.5871;研究区内主要农作物在空间上整体呈现出镶嵌结构;研究区内中度适宜种植当地主要农作物的用地面积最大,其次是适宜种植面积居中,不适宜种植地区的面积最小.
Crop planting structure information is an important reference basis for crop growth monitoring and agricultural structure adjustment.Timely and accurate acquisition of regional crop spatial distribution and planting information is of great significance for sustainable agricultural development.This study took Guyuan County,Zhangjiakou,Hebei Province as the research object,used NDV,NDBI,NDWI,NDRE1,SRre and CIred edge data extracted from Sentinel-2 multispectral data as the characteristics,respectively used SVM support vector machine,decision tree method and random forest method to fine extract typical crops in the study area,explored the spatial distribution of main crops,and discussed the accuracy of different classification methods for crop classification by comparing kappa coefficients to choose the optimal classification method.Five indicators:soil properties,soil erosion degree,elevation,slope,and aspect were selected to establish a crop suitability evaluation system.GIS Analytic Hierarchy Process and land suitability grading indicators were used to evaluate the crop suitability of Guyuan County.The results show that the classification accuracy of 8 main crops in the study area based on the random forest classification method is the highest,with an overall accuracy of 65.10%and a Kappa coefficient of 0.5871.The main crops in the research area exhibit a mosaic structure in space as a whole.The area of moderately suitable for planting local main crops in the research area is the largest,followed by the suitable planting area,and the area unsuitable for planting is the smallest.
作者
赵孟辰
阿里木江·卡斯木
ZHAO Mengchen;ALIMUJIANG Kasmu(College of Geography and Tourism,Xinjiang Normal University,Urumqi 830054,China)
出处
《西南大学学报（自然科学版）》
CAS
CSCD
北大核心
2023年第11期176-185,共10页
Journal of Southwest University(Natural Science Edition)
基金
新疆维吾尔自治区创新环境(人才、基地)建设专项(2022D04007)
新疆维吾尔自治区社会科学基金项目(22BJY020).
作者简介
赵孟辰,博士,副教授,主要从事地理信息研究.
