【摘要】 精准掌握耕地撂荒分布情况对于耕地保护和粮食安全尤为重要。为此，该文以南方丘陵区地带粮食主产区湖南省宁远县作为研究区域，依据该地水稻种植物候特征，以2020年和2021年获取的6期卫星遥感影像和航空影像作为数据源，以第三次全国国土调查水田和1∶2 000 DLG田埂范围作为最小提取单元，基于时序归一化植被指数(normalized difference vegetation index, NDVI)分析和改进支持向量机(support vector machine, SVM)方法，提取水稻种植图斑；通过连续2 a差值计算，筛选出疑似撂荒区域，结合无人机抽样航空调查，对疑似区域进行精细化筛查，分析得到耕地撂荒区域，经过现场核查监测精度可达85%以上。结果表明，融合空、天、地一体化的遥感监测方法可以为农业管理部门的耕地撂荒现象治理工作提供科学有效的数据支持。更多还原

【Abstract】 With continuously accelerated urbanization, urban expansion-induced farmland occupation and rural hollowing have gradually aggravated arable land abandonment, posing challenges to China’s food security. Hence, accurately determining the distribution of abandoned arable land is critical to arable land protection and food security. This study investigated a county in the major crop production area in a hilly region in southern China. Based on the phenological characteristics of local rice planting, this study adopted six phases of satellite remote sensing images and aerial images acquired in 2020 and 2021 were used as the data source and selected paddy fields determined based on the Third National Land Survey and field ridges in 1∶2 000 digital line graphics(DLGs) as the minimum information extraction unit. Then, the rice planting patches were extracted through time-series normalized difference vegetation index(NDVI) analysis and the improved SVM method. Suspected abandoned areas were screened out through the difference calculation for two consecutive years and then further identified in the unmanned aerial vehicle(UAV)-based sampling aerial survey. Consequently, the abandoned arable land areas were determined, with monitoring accuracy exceeding 85%, as revealed by on-site verification. The results of this study show that the space-air-ground integrated remote sensing monitoring method can provide scientific and effective data support for agricultural management departments to control and manage cultivated land abandonment.更多还原