基于多时相SAR对农作物种植区信息的提取研究

Extraction of Crop Planting Area by Multi-temporal SAR

针对我国农作物种类种植区域难以精细化统计,区域性农作物产量难以精准估计的问题,提出了基于多时相SAR,利用雷达影像,根据不同植被的后相散射机制不同的原理,并利用监督分类的方法对农作物种植区域进行高效精准划分,以此来估算农作物种植面积。以黑龙江省东部区域的某一水稻种植区域为例,根据获取的该区域的ENVISAT-1雷达影像进行时序性分析,并根据水稻生长周期的反射机制进行监督分类,获取了该区域的水稻种植面积,并通过与此地的实地调查面积比对分析,验证了此估算方法的可靠性,为此区域的水稻估产及管理提供了精准依据。

Considering the problem that the crop area is difficult to be accurately counted and the regional crop yield is diffi- cult to be accurately estimated, this paper puts forward the different principles of using the radar image according to the dif- ferent scatter mechanism of different vegetation, and uses supervised classification method to classify the crop planting area ef- ficiently. Taking a rice planting area of Heilongjiang Province as an example, the temporal analysis of ENVISAT-1 images is carried out and the rice planting area is obtained by supervising and sorting according to the reflection mechanism of rice growth cycle. By comparing with the field survey in this area, the reliability of the estimation method is verified, which pro- vides a precise basis tbr rice yield estimation and management of the area.