基于多源卫星数据的黑河中游绿洲区土地利用分类与作物类型提取及其时空变化分析

研究土地利用与作物种植结构变化是分析农业生产活动和生态环境变化的重要手段。黑河中游绿洲是我国西北地区最重要的绿洲之一,也是甘肃省主要的商品粮基地。为准确研究黑河中游绿洲的时空变化情况,本文利用2001—2015年MODIS卫星数据,提取分析了近15年黑河中游绿洲范围的时空变化特征;基于2001—2016年的Landsat/TM,利用支持向量机分类法获得该绿洲区土地利用分类;基于2011—2016年的HJ1A/CCD数据,利用HANTS滤波后的NDVI时间序列曲线,获得作物种植面积的精细分类。经验证,该土地利用分类精度高于88.46%,kappa系数为0.81;小麦和玉米与实地验证点对比的分类精度高于90.8%。结果表明:1)黑河中游绿洲面积总体为增加趋势,仅2014年、2015年略有减少,绿洲面积从2001年的2 701 km2增加到2015年的2 936 km2。2)2001—2016年间,裸地面积减少436.7 km2,耕地面积增加91.3 km2,草地、林地面积增加289.6 km2。3)2011—2016年,小麦、玉米种植面积均有减少,小麦共计减少195.77 km2,玉米减少144.37 km2。研究结果可为当地农业种植结构的调整和作物估产及绿洲生态保护提供一定依据和参考。

Land classification and spatio-temporal changes of wheat and corn in the middle reaches oasis of Heihe using multi-satellite data

Study on land use/cover and crop planting structure in agricultural regions is critical for analyzing changes in agricultural production activities and the ecological environment. The middle reaches oasis of Heihe is one of the most important oases in Northwest China and it is also the main commodity grain production base in Gansu Province. The use of surface water in Heihe is huge and has resulted in excessive utilization of surface water in the middle reaches of Heihe. The shortage of water resources has led to the deterioration of ecological environment in the oasis area in the middle reaches of Heihe, which has in turn caused significant damage to the development of Heihe oasis and agriculture. In order to accurately study the temporal and spatial changes of the oases in the middle reaches of Heihe, this study used MODIS satellite data to analyze the time-space variation characteristics of oases in the middle reaches of Heihe for the period from 2001 to 2015. Based on the Landsat/TM of 2001-2016, land use classification for the oasis area was obtained using the support vector machine classification. Then based on the HJ1A/CCD data for 2011-2016, a fine classification of crop planting structure was obtained from the HANTS filtered NDVI time series curve. The reason for the difference in satellite data selection was due to the failure in data quality and the late launch of the satellites. The land use classification method had high accuracy, with classification result accuracy higher than 88.46% and kappa coefficient of 0.81. The classification accuracy of wheat and maize compared with field verification points was higher than 90.8%. The results showed that 1) the area of the oasis in the middle reaches of Heihe generally increased. With the exception of 2014 and 2015 when the area of oasis slightly reduced, it increased from 2 701 km² in 2001 to 2 936 km² in 2015. 2) For 2001-2016, the area of bare land reduced by 436.7 km², that of cultivated land increased by 91.3 km², then of grassland and woodland increased by 289.6 km² and water body remained unchanged. 3) For 2011-2016, the planted area of wheat and corn decreased. Wheat total area decreased by 195.77 km² and that of corn decreased by 144.37 km². The results provided the basis and reference for adjustment of the local agricultural structure and crop yield estimation, and water resources and oasis ecology protection. The innovation point of the paper was that the area of oasis calculated based on area calculation formula for vegetation coverage. This correctly reflected the trend of change in the middle reaches of Heihe in the over 15-year period. It provided a new method for the calculation of oasis area, which greatly improved the accuracy of calculation when using low resolution (over 250 m) satellite data.