Analysis of urban environmental comfort using Landsat-8 multitemporal data and Artificial Neural Network

The presence of greenery in urban residential and office areas can improve the comfort of residents who live in these environments. In an urban setting, vegetation serves an ecological purpose by absorbing carbon dioxide, supplying oxygen, lowering the temperature to produce a tolerable microclimate, acting as a water catchment area, and reducing noise. Urbanization and anthropogenic activity-driven growth of urban and sub-urban regions put stress on the local vegetation and have the potential to lower environmental comfort. To promote the creation of a sustainable urban environment, a thorough analysis of the urban environment is required. Applications for remote sensing in all spectral, geographic, and temporal dimensions have increasingly adopted the usage of deep learning methods with artificial neural networks. This study attempted to predict the application of remote sensing data for analyzing environmental comfort in metropolitan areas based on multitemporal Landsat-8 data. The study area is Greater Jakarta. The approach was based on supervised classification with neural network techniques and land parameters like surface temperature, brightness index, greenness index, and wetness index. According to the study's findings, the proposed method could accurately predict that very uncomfortable classes predominated in Jakarta, Bogor, Depok, Tangerang, Bekasi, and surrounding areas by more than 92%. In addition to being densely populated with communities, urban environments are uncomfortable due to a lack of vegetation cover, which increases surface temperatures. In the future, this research can provide input for similar research, especially in the use of deep learning Artificial Neural Network methods for environmental analysis.