Forest Fire Information Detection and Vegetation Restoration Dynamic Monitoring Based on Multi-source Remote Sensing Images

The novel fractional vegetation coverage model was developed from normalized difference vegetation index and the dimidiate pixel model based on Landsat 5 TM and GF-1 WFV images, the aerial CCD data of 0.2 m resolution and the global land cover map at 30 m. The fire information of Daxing'anling area from 2001 to 2020 was calculated using the difference of NDVI before and after fire. The change of forest coverage and its influencing factors were quantitatively analyzed by FVC, and the vegetation coverage of hot spot area was analyzed through the change of FVC. Vegetation resilience after different forest fires was analyzed in this study, and the results demonstrated: the vegetation coverage estimation method based on dimidiate pixel model is simple yet accurate, which is suitable for remote sensing data of different resolutions; dNDVI could differentiate the information of fire area effectively. FVC of the whole study area is high, frequent fire is the main factor affecting the distribution and dynamic change of FVC in the study area. FVC restoration effect varies different degrees of fire area. The FVC in small fire area can be restored to the original state, the FVC in medium fire area still has a certain gap with the original FVC after a long time of natural restoration; while FVC will increase in large fire area, especially in large fire area after a long time of restoration, but it is far from the original FVC. The research shows that it is difficult to restore the forest after the fire disturbance, which has destructive effects on forest ecosystem.