A Wildfire Detection Algorithm Based on the Dynamic Brightness Temperature Threshold

Satellite remote sensing plays an important role in wildfire detection. Methods using the brightness and temperature difference of images to determine if a has occurred are one main research directions forest fire monitoring. However, common detection algorithms mainly based on fixed threshold distinguish pixels non-wildfire pixels, which reduces applicability algorithm different space–time regions. This paper presents adaptive algorithm, DBTDW, dynamic threshold. First, regression dataset, MODIS_DT_Fire, was constructed moderate resolution imaging spectroradiometry (MODIS) Then, meteorological information, normalized vegetation index (NDVI) elevation information provided by DBTDW used calculate obtain minimum burning area Planck Otsu algorithm. Finally, six models were trained establish correlation between factors wildfire. The root-mean-square error (RMSE) mean absolute (MAE) evaluate performance. results show that under XGBoost model, best prediction effect (leave-one-out method: RMSE/MAE = 0.0730). Compared with method threshold, proposed this adaptively higher universality, will help improve effectiveness satellite