Determination of Forest Fire Intensity Level Using Multi-Temporal Satellite Remote Sensing Imagery

Sudden changes in forest fire behavior are the primary cause of casualties, and sudden increases in forest fire intensity are the primary driver of these behavioral changes. Accurately quantifying forest fire intensity can significantly reduce suppression risks for firefighters. The method of quantifying forest fire intensity by calculating the fireline's energy release rate per unit length requires precise acquisition of key parameters such as fuel load, fire spread rate, and calorific value of combustion. However, the extreme conditions of forest fires make it very difficult to obtain key parameters directly. The existing indirect methods, empirical interpretation, flame height sensing, and satellite FRP inversion, still struggle to overcome the limitations of high ground deployment costs, strong human subjectivity, and insufficient remote sensing resolution. This study used the Himawari 8/9 multi-temporal satellite remote sensing images to derive image differential expressions based on the Structural Similarity Index Measure (SSIM). Forest fires are identified using image format conversion and the Single-band Image Fusion Method (SIFM). Independent variables, including duration, burned forest area, and maximum FRP, are selected, with forest fire intensity level as the dependent variable to construct a forest fire intensity level model. Experimental results demonstrate that the SIFM outperforms the Image Format Conversion Method (IFCM) in terms of accuracy for identifying forest fires, and the constructed forest fire intensity level model effectively determines forest fire intensity levels.