Remote Sensing-Based Ecosystem Monitoring and Disaster Management in Urban Environments Using Machine Learnings

A key component of managing natural resources is the use and cover of the land. Maps of environmental changes are created using it in order to monitor ecosystems. For forestry, urban planning and agriculture, automatic mapping has many benefits. The science of remote sensing has benefited greatly from the development of deep learning techniques, which have yielded impressive results in image classification. This research proposes novel technique in ecosystem monitoring with disaster management in urban environments based on hyperspectral image analysis using a machine learning model. Here the input has been collected as hyperspectral image as well as processed for noise removal, normalisation and smoothening. Ecosystem monitoring is carried out utilizing Gaussian attention linear discriminant logistic regression. Then, the disaster management has been carried out based on the ecosystem monitoring model using cloud-based genetic spatio algorithm. Experimental analysis is carried out in terms of prediction accuracy, precision, F-measure, recall and RMSE for various hyperspectral images: prediction accuracy of 96%, precision of 97%, F-measure of 88%, recall of 95% and RMSE of 60% for proposed technique.