【摘要】 为了建立一个用于环青海湖地区草地蝗虫发生空间分布的遥感监测模型,分别在1997年6月下旬到7月上旬、1999年8月中下旬及2000年8月中下旬进行了野外考察,获得了198个蝗虫密度样本资料,收集了同期的TM5遥感图像。通过TM5遥感影像计算了10种植被指数,如归一化植被指数(NDVI)和褐色植被指数(BVI)等。采用统计方法,研究了环青海湖地区草地蝗虫密度(GD)与各种植被指数之间的相关关系。结果表明:尽管GD与10个VI的线性相关并不显著,但GD对NDVI和BVI比较敏感。86%的GD>5头/m2样本出现在满足条件0.2<NDVI<0.5和75<BVI<110的地方,而低GD样本中只有29%满足该条件。表明这两个指数对草地蝗虫发生区域有较强的指示性,可用于草地蝗虫的遥感监测。更多还原

【Abstract】 To develop a remote sensing model for monitoring the spatial distribution of rangeland grasshopper infestation in the regions around Qinghai Lake, 198 site grasshopper density (GD) data, and three TM5 images were collected during the last ten days of June and the first ten days of July in 1997, the middle and last ten days of August in 1999, the middle and last ten days of 2000. Ten vegetable indexes (VI), for example the normalized deviation vegetable index (NDVI) and brown vegetable index (BVI), were calculated from TM5 images. The correlations between the GD and VI were investigated with statistics methods. It shows that though the linear correlation between GD and the ten VI are not significant, GD is sensitive to NDVI and BVI, 86% of high GD samples appear in the span 0.2<NDVI<0.5 and 75<BVI<110, and just 29% low GD samples appear in this bound. This implicate that NDVI and BVI are good indicators of rangeland grasshopper location, so that can be used as predictors for rangeland grasshopper remote monitoring.更多还原