The land cover classification in urban areas is described in this research work. The use of hyperspectral image analysis is growing in popularity because it performs better than conventional machine learning techniques. Hypercubes, a type of three-dimensional dataset with two spatial dimensions and one spectral dimension, make up the Hyperspectral imaging (HSI). An overview of HSI's uses in remote sensing applications and the methods for classifying it are given in this research. In the field of HSI, numerous experiments are conducted with various deep learning methods for analysis and classification. The main components of this research is convolutional neural network (CNN)and long short-term memory (LSTM) that shows to be more effective than alternative models. In this case, spectral and spatial features are extracted using CNN and LSTM, respectively, and the results are then classified using support vector machines (SVM). The datasets utilized in this study were gathered using a ROSIS sensor/spectrometer at Pavia University and Indian Pines.

LSTM CNN Gabor filter SVM Indian Pines Principal Component Analysis (PCA)