基于Himawari-8卫星的中国逐小时火点产品数据集

考虑到中国JMA-Himawari-8火灾产品的不确定性，我们在原始Himawari-8源的基础上，通过采用火灾提取的动态阈值和地面热源数据库，提出了一种自适应的每小时NSMC-Himawari-8消防产品。基于视觉提取的参考文献和一致性检查，我们发现NSMC-Himawari-8消防产品有效地消除了JMA-Himawari-8火灾产品中包含的大多数虚假火灾警报。基于罕见的现场收集的地面参考数据集，我们评估了JMA-Himawari-8和NSMC-Himawari-8消防产品在中国各地的可靠性。JMA-Himawari-8消防产品的总体精度为54 % 和59 % （不考虑遗漏错误）。相比之下，通过识别更多的真实火灾像素并避免大多数虚假火灾警报，NSMC-Himawari-8火灾产品的总体准确度为80 % 和84 % （不考虑遗漏错误）。NSMC-Himawari-8消防产品可以成为改善中国实时火灾监测的一个很有前途的来源。这项研究还为使用下伏表面和热源的局部数据集来提高特定地区全球火灾产品的准确性提供了有用的参考。

Given the uncertainties associated with the China-focused JMA-Himawari-8 fire product, we propose an adaptive hourly NSMC-Himawari-8 fire product based on the original Himawari-8 data source, utilizing dynamic fire extraction thresholds and a ground heat source database. Based on visually interpreted reference datasets and consistency checks, we determined that the NSMC-Himawari-8 fire product effectively eliminates most false fire alarms present in the JMA-Himawari-8 fire product. Using a rare ground reference dataset collected in situ, we assessed the reliability of both the JMA-Himawari-8 and NSMC-Himawari-8 fire products across mainland China. The JMA-Himawari-8 fire product yields overall accuracies of 54% and 59% (excluding omission errors). In comparison, the NSMC-Himawari-8 fire product achieves overall accuracies of 80% and 84% (excluding omission errors) by identifying more true fire pixels and avoiding the majority of false fire alarms. The NSMC-Himawari-8 fire product serves as a promising resource for improving real-time fire monitoring in China. This study also provides a valuable reference for enhancing the accuracy of global fire products in specific regions by leveraging local datasets of underlying surfaces and heat sources.

搜集汇总

背景与挑战

背景概述

该数据集是中国地区基于Himawari-8卫星的逐小时火点监测产品，通过优化算法显著提高了火灾检测的准确性，总体精度从原始产品的54-59%提升至80-84%，有效减少了虚假警报。