基于GF1-NDVI时序影像对春小麦进行提取研究
|
摘要
利用传统方法对农作物种类、分布和种植面积等调查,需要耗费大量的人力、物力和财力。该研究以西宁市为研究区域,采用高分一号影像,对西宁市春小麦进行分类和提取模型设计。在全生育期波谱特征曲线分析基础上,提取春小麦的NDVI(归一化植被指数)曲线特征。采用基于NDVI阈值的决策分类技术,进行作物识别与提取。最后设计精度自检方案,通过混淆矩阵得出其总体精度达到93.8%,kappa系数为0.875。其用户精度和制图精度分别为93.7%和94.9%。从分类精度可以看出,利用中高分辨率遥感卫星影像,在作物NDVI时间序列变换规律分析的基础上,可以准确的进行大面积农作物的分类与提取。在全国农作物面积与农作物种类等资源调查中具有非常大的应用潜能。
Using traditional methods to investigate crop species,distribution,planting area and so on,takes a lot of manpower and material resources.This study takes Xining as the study area.The classification and extraction model of spring wheat of Xining city is carried out by using high GF-1 image.The NDVI(normalized differential vegetation index)curve characteristics of spring wheat is extracted on the basis of spectral curve analysis of the whole growth period.The recognition and extraction of the crop are carried out using the decision classification technique based on the NDVI threshold.Finally,the design accuracy self checking scheme and obtained the confusion matrix,calculated the overall accuracy is 93.8% and the kappa coefficient is 0.875.Its user precision and mapping accuracy are 93.7% and 94.9% respectively.It can be seen from the classification accuracy that the middle and high resolution remote sensing images can be used for the classification and extraction of large area crops on the basis of the analysis of the crop NDVI time series change rule.It has great potential in investigation of crop areas and types in China.
Using traditional methods to investigate crop species,distribution,planting area and so on,takes a lot of manpower and material resources.This study takes Xining as the study area.The classification and extraction model of spring wheat of Xining city is carried out by using high GF-1 image.The NDVI(normalized differential vegetation index)curve characteristics of spring wheat is extracted on the basis of spectral curve analysis of the whole growth period.The recognition and extraction of the crop are carried out using the decision classification technique based on the NDVI threshold.Finally,the design accuracy self checking scheme and obtained the confusion matrix,calculated the overall accuracy is 93.8% and the kappa coefficient is 0.875.Its user precision and mapping accuracy are 93.7% and 94.9% respectively.It can be seen from the classification accuracy that the middle and high resolution remote sensing images can be used for the classification and extraction of large area crops on the basis of the analysis of the crop NDVI time series change rule.It has great potential in investigation of crop areas and types in China.
引文
[1]刘佳,王利民,杨福刚,等.基于HJ时间序列数据的农作物种植面积估算[J].农业工程学报,2015(3):199-206.
[2]王利民,刘佳,杨玲波,等.基于NDVI加权指数的冬小麦种植面积遥感监测[J].农业工程学报,2016,32(17):127-135.
[3]张煦,马驿,郑雯,等.基于时序MODIS-NDVI的油菜种植面积变化趋势分析——以江汉平原为例[J].长江流域资源与环境,2016(3):412-419.
[4]TURNER M D,CONGALTON R G.CIassification of Multi-temporal SPOT-XS Satellite Data for Mapping Rice fields on a West African Floodplain[J].Int J Remote Sensing,1998,19(1):21-41.
[5]马扶林,周秉荣,许正福.基于GIS的青海省主要农作物种植区划[J].现代农业科技,2016(12):207-208,212.
[6]柳雄,王秀娟,韩建琪.青海省小麦种植现状以及推广前景[J].中国种业,2017(2):14-17.
[7]刘寒,祝晓坤,左琛,等.基于时序SPOT6遥感影像的审批土地监测应用[J].北京测绘,2017(6):150-152.
[8]宋富强,邢开雄,刘阳,等.基于MODIS/NDVI的陕北地区植被动态监测与评价[J].生态学报,2011,31(2):354-363.
[9]葛广秀,李卫国,景元书.基于NDVI密度分割的冬小麦种植面积提取[J].麦类作物学报,2014,34(7):997-1002.
[10]贺振,贺俊平.基于NOAA-NDVI的河南省冬小麦遥感估产[J].干旱区资源与环境,2013,27(5):46-52.
[11]许礼林,雷莹,王明.基于高分辨率遥感影像提取地理变化信息的辅助生产方法研究[J].北京测绘,2016(2):119-122.
[12]方杰诗,周晓霖.基于遥感影像和NDVI阈值法的银川市植被覆盖度反演与监测[J].节水灌溉,2014(11):68-72,76.
[2]王利民,刘佳,杨玲波,等.基于NDVI加权指数的冬小麦种植面积遥感监测[J].农业工程学报,2016,32(17):127-135.
[3]张煦,马驿,郑雯,等.基于时序MODIS-NDVI的油菜种植面积变化趋势分析——以江汉平原为例[J].长江流域资源与环境,2016(3):412-419.
[4]TURNER M D,CONGALTON R G.CIassification of Multi-temporal SPOT-XS Satellite Data for Mapping Rice fields on a West African Floodplain[J].Int J Remote Sensing,1998,19(1):21-41.
[5]马扶林,周秉荣,许正福.基于GIS的青海省主要农作物种植区划[J].现代农业科技,2016(12):207-208,212.
[6]柳雄,王秀娟,韩建琪.青海省小麦种植现状以及推广前景[J].中国种业,2017(2):14-17.
[7]刘寒,祝晓坤,左琛,等.基于时序SPOT6遥感影像的审批土地监测应用[J].北京测绘,2017(6):150-152.
[8]宋富强,邢开雄,刘阳,等.基于MODIS/NDVI的陕北地区植被动态监测与评价[J].生态学报,2011,31(2):354-363.
[9]葛广秀,李卫国,景元书.基于NDVI密度分割的冬小麦种植面积提取[J].麦类作物学报,2014,34(7):997-1002.
[10]贺振,贺俊平.基于NOAA-NDVI的河南省冬小麦遥感估产[J].干旱区资源与环境,2013,27(5):46-52.
[11]许礼林,雷莹,王明.基于高分辨率遥感影像提取地理变化信息的辅助生产方法研究[J].北京测绘,2016(2):119-122.
[12]方杰诗,周晓霖.基于遥感影像和NDVI阈值法的银川市植被覆盖度反演与监测[J].节水灌溉,2014(11):68-72,76.