LiDAR-Based Climate Change Imaging in Geoscience Using Spatio Extreme Fuzzy Gradient Model

In order to estimate high-resolution air temperatures, computationally demanding procedures are required because climate models do not provide the resolution required for urban climate studies. Data-driven methods, on the other hand, provide air temperature downscaling that is quicker and more precise. Lidar technology has become a vital tool for researching these atmospheric components because of its capacity to measure backscattered signals and investigate the atmosphere at extremely high spatial and temporal resolution. However, machine learning has become a potent technique for sophisticated data processing and analysis because of the complexity and diversity of lidar technology. This research proposes a novel technique in urban–rural interface analysis in remote sensing by machine learning (ML) model. Here the input is collected as urban–rural landscape data and processed for noise removal as well as normalisation. Then these data features have been extracted and classified using a spatio extreme fuzzy gradient model with reinforcement Markov vector perceptron neural networks. Experimental analysis has been carried out for various geographical regions in terms of training accuracy, average precision, recall, RMSE, and F-1 score. The proposed model attained an F-1 score of 92%, training accuracy of 98%, average precision of 95%, recall of 93%, and RMSE of 55%.