Using Images Generated by Sentinel-2 Satellite Optical Sensor for Burned Area Mapping

Remote Sensing is identified as an effective and efficient tool for monitoring fire events, and for quantifying fire effects on environment. Satellite images are used both to identify active fires and to analyse their effects, as well as to define burnt areas and map the severity of fires. Fires modify the structure and the reflectance of vegetation as well as the soil properties within the burned area; the produced changes are detectable in the visible, infrared and microwave parts of the electromagnetic spectrum. One of the most useful approach is based on classification of images using the spectral properties of burnt residues. This paper aims to use data obtained from optical sensor mounted on Sentinel-2 platform for mapping areas damaged by fire in a precise and rapid way. Sentinel-2 offers multispectral medium and high spatial resolution images with 13 spectral bands and about 5-day temporal resolution. Two images concerning the same scene in Campania Region but acquired on different dates are considered: pre-fire and post-fire. For each image, the Normalized Burned Ratio Index (NBR) is calculated, which allows to identify the areas affected by the fire and the relative degree of severity. Using change detection techniques, burned map can be identified. The evaluation of the accuracy is carried out using some indexes widespread in remote sensing literature, such as User’s Accuracy, Producer’s Accuracy, Overall Accuracy and Kappa coefficient. The values obtained in the confusion matrix showed the high quality of the developed method based on the use of the NBR index.