Classification of Urban Objects from HSR-HTIR data using CNN and Random forest Classifier

Detection and classification of urban objects have been to a great degree troublesome without manual help which was monotonous and tedious. In recent years, High spatial resolution hyper spectral thermal infrared (HSR-HTIR) transformed into a novel wellspring of data that wound up available for urban object detection. The classification of the HSR-HTIR image is done using Random forest and Convolutional Neural Network for the raw dataset. The classification was done with the assistance of Spectral Python, which is an unadulterated Python module for getting ready hyperspectral image data. It has capacities with regards to scrutinizing, appearing, controlling, and requesting hyperspectral imagery. Alongside Spectral Python libraries in particular Numpy, Sklearn, Keras, Tensorflow and SciPy were used. With the assistance of these tools, results for the different classification techniques were obtained, which were compared with each other and a performance assessment was made based on the level of precision.