Maximum Likelihood Classifier Analysis of UAV Image Classification Post-Fire Area Estimation

Forest fire incidents are a prevalent hazard in Malaysia and globally. The repercussions of this situation are significant for all stakeholders, and they will have a profoundly adverse impact on nature, the environment, and the nation in particular. Numerous prior studies have employed space-borne or airborne multi-spectral imagery to assess the burned area; however, this approach incurs significant costs. This study employs supervised classification for image classification to estimate the burned area. The maximum likelihood classifier was selected for its superior accuracy in image classification and was chosen for use in this study. Two distinct software applications were used in estimating the burned area results. An analysis of the two software programs reveals that they yield nearly identical outcomes when classifying the same burned area. The findings indicate that both programs utilizing the same classification method are capable of estimating the burned area. Therefore, we can assess the damage incurred by the fire in the area, implement measures to mitigate this loss in targeted locations, and strive to prevent future incidents from happening.