基于多时相卫星影像的复杂山区烤烟种植区域提取

为提高卫星影像对复杂山区烤烟种植区的提取精度。以高分六号2m分辨率卫星影像作为遥感数据源，基于多时相卫星遥感影像序列，根据烤烟不同生长期的卫星影像特征，采取面向对象分类方法，分别完成3个不同时相卫星影像的烤烟种植区初步提取，再利用多时相植被指数规则初步提取烤烟种植区，最后通过交叉验证获取烤烟种植区域，并完成精度验证。结果表明：该研究方法能较为准确的提取复杂山区烤烟种植区，能够保持较低的错分率（5.3%），提取精度达94.67%，较基于单一时相卫星影像提取烤烟种植区域精度（90%）明显提升。

Extraction of Flue-cured Tobacco in Complex Mountainous Area Based on Multi-temporal Satellite Images

In order to improve the extraction accuracy of flue-cured tobacco planting area in complex mountainous areas, this paper used GF6 multi-temporal satellite remote sensing image series, adopted object-oriented classification method to preliminary extract the flue-cured tobacco planting areas from three different time-phase satellite images according to the satellite image characteristics of different growth stages of flue-cured tobacco,then used the multi-temporal vegetation index rules to extract the planting area of flue-cured tobacco. Through cross validation, the flue-cured tobacco planting area was finally obtained,then completed the accuracy verification. The experiment result showed that this research method can accurately extract the flue-cured tobacco planting area in complex mountainous areas. The higher recognition accuracy (94.67%) was obtained under the lower misjudgment ratio (5.3%), compared with the extraction of flue-cured tobacco planting area based on single temporal satellite image, the accuracy is significantly improved.