论文标题
优化SAR数据处理和用于森林变化检测的阈值：东部亚马逊东部的森林砍伐警告的申请
Optimizing SAR data processing and thresholding for forest change detection: an application for early deforestation warnings on eastern Amazonia
论文作者
论文摘要
目前的工作提出了一种用于早期森林砍伐检测的操作方法的原型，可对多云的热带雨林进行检测。所提出的方法利用Sentinel-1 SAR数据处理在Google Earth Engine Platform中，以标记近期森林砍伐可能性很高的区域。对东部亚马逊盆地东部地区的结果的评估表明，在生产商的准确性方面，共具有共聚数据（VV频段）提供了最佳的结果（95,4％的5％意义，88.9％的意义为88,9％，1％的意义），而交叉极化数据（VH频段）在用户的准确性方面具有出色的效果（86％的5％显着性，1％的显着性1％。
The present work proposes a prototype for an operational method for early deforestation detection of cloudy tropical rainforests. The proposed methodology makes use of Sentinel-1 SAR data processed into the Google Earth Engine platform for flag the areas where the probability of recent deforestation is high. The evaluation of the results over a region on the Eastern Amazon basin showed that copolarized data (VV band) offers the best results in terms of producer's accuracy (95,4% for a 5% significance, 88,9% for 1% significance), while crosspolarized data (VH band) offered excellent results in terms of user's accuracy (86% for a 5% significance, 100% for 1% significance).
