An Introduction to Artificial Intelligence and Its Applications Towards Remote Sensing

Artificial intelligence plays a major role in every field nowadays. It is a part of computer science in which machines can act like humans in terms of behaviour, thinking and decision-making. It has a tremendous impact in all sectors. In short, it replicates human intelligence. The main aim of artificial intelligence is to mimic the human brain through reasoning, learning and problem-solving. Artificial intelligence is achieved by machine learning concepts. It includes supervised, unsupervised, semi-supervised and reinforcement learning. In supervised learning, labelled data will be given for processing so that the machine can learn it easily and it either categorizes or recognize the given task. Whereas, in unsupervised algorithms, without knowing any details about the data, learning will be done automatically by using some clustering concepts. The semi-supervised algorithm will be the combination of both supervised and unsupervised algorithms. Reinforcement learning will be the most useful one in case of real environments like playing games, driving a car and playing a piano. In this case, a machine will learn on its own by getting the rewards and penalties. Additionally, ensemble algorithms like Bagging, Boosting and Stacking are also available to learn the machines. In recent years, the role of these algorithms has been marvellous in the field of remote sensing. In this chapter, a view of both machine learning and deep learning with satellite imagery for remote sensing data is given.