一种基于时序遥感影像混合像元分解的耕地种植强度估算方法——以湖北省为例

在耕地总量减少、质量下降与人口持续增长、对粮食需求增加的背景下，耕地集约化利用成为保障国家粮食安全的重要手段.现有耕地集约化利用指标定义主要有种植频率和复种指数等.种植频率是对种植模式的硬划分(如单季或双季)，会损失部分信息.复种指数是以统计数据为基础计算得到，忽略了行政区内作物分布状况的空间异质性.该文引入“种植强度”这一概念，实现对耕地集约化利用状况的精细表达.通过分析不同土地覆盖及不同复种状况的耕地在时间序列植被指数上的差异，构建特征空间，利用线性光谱分解的方法在时间维度上进行混合像元分解.以湖北省为研究区估算耕地种植强度，基于湖北省县级尺度和样区像素尺度验证，相关系数达到0.9.该研究为复杂地形、复杂种植结构下的种植强度估算提供了新的方法，可用于区域尺度耕地集约利用状况的精细制图.

A method for estimating cropping intensity based on mixed pixel unmixing using time series remote sensing images——A case study of Hubei Province

Under the background of decreasing quantity and quality of croplands， and increasing demand for food， intensive utilization of croplands palys an important role in national food security. The existing indicators of intensive utilization of cropland mainly include cropping frequency and multiple cropping index. The crop patterns include single cropping and double cropping， which were norminal measurement and will lose some information. The multiple cropping index is calculated based on the statistical data， ignoring the spatial heterogeneity whinin the administrative region. In this paper， the concept of “cropping intensity” is introduced to express the intensive utilization of cropland delicately. By analyzing the difference of vegetation index in time series under different cropping conditions， the feature space was constructed， then the linear spectral unmixing was carried out in time dimension. Taking Hubei Province as the research area， the cropping intensity was estimated， and the correlation coefficient was 0.9 based on the validation in both county and pixel level. This study provides a new method for estimating cropping intensity under heterogeneous topography and crop patterns. It can be used for accurate mapping of cropping intensity at regional scale.