基于多时相GF-1WFV 数据的南方丘陵地区甘蔗种植面积提取

南方丘陵地区甘蔗种植具有分散、形状多样等特点,利用中低分辨率遥感数据提取甘蔗种植面积时,数据的有效性和精度都难以满足要求.利用高分辨率GF-1WFV遥感数据,分析比较甘蔗与其他地物类的光谱特征、纹理特征及植被指数时间变化差异,采用多时相迭代方法构建甘蔗提取特征向量决策树模型.利用该模型提取了广西江州区的甘蔗种植面积,经野外实地调查验证,甘蔗面积提取的用户精度为90.13％,生产精度为88.78％,表明GF-1WFV数据是复杂地形下提取甘蔗等农作物的潜力数据源,该提取思路可以为在全国范围内不同地区了解甘蔗种植情况,进行甘蔗长势监测和产量估测提供技术参考.

Sugarcane extraction in the southern hills using multi-temporal GF-1 WFV data

As sugarcane planting in southern hills exhibits many characteristics, such as scattered, diverse shape and so on, the extraction of sugarcane information using low and medium resolution remote sensing data is hard to satisfy the requirements of accuracy and availability. GF -1 WFV images are used as data source in this paper, spectral characteristics, texture features and vegetation index changes are comparatively analyzed, and then the sugarcane extraction model are established based on decision tree method using multi-temporal images. With this model, sugarcane planting information is extracted in Guangxi Jiangzhou area and fields survey data are used to check the extraction accuracy. The results show that the user accuracy is 90.13% and production accuracy is 88.78%. GF-1 WFV satellite data with its high spatial resolution and convenient obatain means can provide a potential data source for crop information identification, monitoring, extraction in complex area.