论文部分内容阅读
作物种植结构监测和估产是精准农业遥感的重点领域,其研究对于指导作物种植结构和制定农业政策具有重要意义。该文以黑龙江省北安市为研究区,以2015年的Landsat8 OLI和多时相GF-1为遥感数据源,基于物候信息和光谱特征确定的农作物识别关键时期和特征参数,构建面向对象的决策树分类模型,开展作物种植结构监测研究;综合植被光谱指数和地面采样数据,采用逐步回归方法建立产量遥感估算模型。结果表明:多源与多时相的遥感数据可以反映不同农作物的季相特征,应用本文所构建的决策树分类模型,作物分类效果较好,总体精度达87.54%,Kappa系数为0.8115;2015年,北安市的主要作物类型为大豆、玉米、水稻和小麦,面积分别为2204、1955、122和19 km~2,其中大豆的种植面积最大,占作物种植面积的51.24%。基于NDVI、EVI和GNDVI构建的多元回归模型为北安市大豆和玉米产量估算最优模型(R~2=0.823 7,均方根误差135.45 g/m~2,精度80.55%);北安市玉米高产区集中分布在西部,大豆的高产区主要分布在东部;2015年北安市玉米和大豆的单产分别为8 659、2 846 kg/hm~2,总产量分别为16.93×10~8、6.27×10~8 kg。利用作物关键物候期的多源多时相遥感数据能够精确高效地提取作物种植结构,构建的产量估算多元回归模型,为精准农业科学发展提供参考。
The monitoring and estimation of crops planting structure are the key fields of precision agriculture remote sensing. Their research is of great significance for guiding crop planting structure and formulating agricultural policies. In this paper, Bei’an City, Heilongjiang Province as the research area, 2015 Landsat8 OLI and multi-phase GF-1 remote sensing data sources, based on the phenological information and spectral characteristics of the identified key period and characteristic parameters of the crop, to build object-oriented decision-making Tree classification model to carry out crop cultivation structure monitoring research; vegetation spectral index and ground sampling data, using stepwise regression method to establish yield estimation model of remote sensing. The results show that the multi-source and multi-phase remote sensing data can reflect the seasonal characteristics of different crops. Using the decision tree classification model constructed in this paper, the crop classification effect is better, the overall accuracy is 87.54%, Kappa coefficient is 0.8115; in 2015, The main crop types in Beian are soybean, corn, rice and wheat with the areas of 2204, 1955, 122 and 19 km ~ 2, respectively, of which soybeans have the largest acreage, accounting for 51.24% of the crop area. The multivariate regression model based on NDVI, EVI and GNDVI was the best model for estimation of soybean and maize yield in Beian City (R ~ 2 = 0.823 7, root mean square error 135.45 g / m ~ 2 and accuracy 80.55%); The high yielding areas of corn are mainly distributed in the west and the high yielding areas of soybean are mainly distributed in the east. The yields of maize and soybean in Beian in 2015 were 8 659,2 846 kg / hm 2, respectively, and the total yields were 16.93 × 10 8, 6.27 × 10 ~ 8 kg. The multi-source and multi-temporal remote sensing data of crop phenology can accurately and efficiently extract the crop planting structure and build a multiple regression model of yield estimation, which can provide a reference for the development of precision agriculture science.