Urban Expansion Prediction using Supervised Classification and ConvLSTM

Thailand has a strategy to shift its national positioning to Thailand 4.0 through a key strategic initiative called EEC. Chachoengsao Province is part of EEC, which will be supported in industrial development. Mueang Chachoengsao district still has a career in agriculture, which is a major source of food. However, rapid urbanization will have an impact on environmental issues. This study aims to predict urban growth in Mueang Chachoengsao district using supervised classification with minimum distance algorithm and ConvLSTM. The satellite images were acquired from 2013 to 2023. The land uses were classified into five categories and extracted using the minimum distance algorithm. The urban areas were extracted to construct the dataset. The ConvLSTM model was trained with different number of stacked layers and kernel sizes. The performance was evaluated and contrasted using three image measurements. The results show that the ConvLSTM can be used to predict the urban expansion and the population in 2031 is 1.2 times higher than that in 2023.