基于多源数据的渍害田识别

Identification of waterlogged field based on multi-source data

渍害田是耕作层土壤含水率长期过高不利于作物生长的中低产田.渍害田的大范围识别是渍害低产田治理工作要解决的一个重要课题.文中分析了渍害田多种形成因素,认为耕作层土壤含水率过高是造成渍害的根本原因,选择土壤相对湿度作为渍害田识别指标.同时利用安徽省内土壤类型分布资料及土壤特征参数,土壤墒情资料和多期MODIS遥感影像数据及地面产品数据,建立土壤含水率反演模型,并转换为土壤相对湿度分布图.根据土地利用类型和土壤相对湿度的指标,对安徽省内渍害田进行识别,获取渍害田分布图,经过资料对比分析,确定最优的识别方法和成果.

The waterlogged fields are medium-low yielding farmland, where too high soil moisture content is kept for a long time in the plough layer. In order to improve the soil output competencies, the waterlogged fields should be identified and upgraded. Identification of waterlogging field is an important problem which must be solved. This paper will probe into how to identification waterlogging field according to multisource data. The main contents of the article include the followings. By collecting relevant literatures at home and abroad, we identified three soil moisture inversion models, which are based on remote sensing data. We analyze the coefficient between apparent thermal inertia, the relative temperature, the seventh band of MODIS data with soil moisture of 10cm and 20cm depth in study area according to MODIS data and soil moisture information from Anhui soil moisture station. Then we build three models between soil moisture and remote sensing data and check those models by correlation analysis. According to correlation in- dex and the soil attribute distribution map we choose two suitable models for further studies. We combine the soil moisture distribution information and arable land distribution information of the study area. Then based on ＂relative moisture of the soil greater than 80% should be waterlogged field＂, we extract water- logged field distribution information of the study area. After checked the result by comparing with statistical data, we can determine the optimal recognition methods and results