Hyperspectral images have acquired popularity in recent years due to its efficiency in storing quality data. Due to that, deep learning has been used for predicting and obtaining more precious accuracy. Mostly this technology is used in object detection, classification, tracking, etc. In this paper, we are constructing a Convolutional neural network for classification with hyperspectral images, which would exhibit higher performance compared to other models. For this classification, we would be using benchmarking datasets and for better efficiency transfer learning and back-propagation have also been added. Moreover with transfer learning, different learning rates and optimal parameters which affect hyperspectral image classification have also been analyzed.

Convolutional neural network(CNN), Hyperspectral image classification (HSI), Deep learning