Home
This Title All WIREs
WIREs RSS Feed
How to cite this WIREs title:
WIREs Data Mining Knowl Discov
Impact Factor: 4.476

Knowledge discovery from remote sensing images: A review

Full article on Wiley Online Library:   HTML PDF

Can't access this content? Tell your librarian.

Abstract The development of Earth observation (EO) technology has made the volume of remote sensing data archiving continually larger, but the knowledge hidden in massive remote sensing images has not been fully exploited. Through in‐depth research on the artificial intelligence (AI)‐based knowledge discovery approaches from remote sensing images, we divided them into four typical types according to their development stage, including rule‐based approaches, data‐driven approaches, reinforcement learning approaches, and ensemble methods. The basic principles, typical applications, advantages, and disadvantages have been detailed for commonly used algorithms within each category. Conclusions include the following: (a) Rule‐based, data‐driven and reinforcement learning algorithms form a trilogy from knowledge to data, and to capabilities. (b) Rule‐based data mining algorithms can provide prior knowledge for data‐driven approaches, the knowledge discovered by data‐driven models can be as an important complement to expert knowledge and rule sets, and reinforcement learning approaches can effectively make up for the lack of training samples or small training sample in data‐driven models. (c) The traditional data‐driven machine learning approaches and their ensemble methods are the current and may be the future mainstream methods for large regional and even global scale long time series remote sensing data mining and analysis, and improving their computing efficiency is the key research direction. (d) Deep learning, deep reinforcement learning, transfer learning, and an ensemble approach of the three may be the main means for small‐area scope, short time series, and key geoscience information extraction from remote sensing images within a long time of the future. This article is categorized under: Algorithmic Development > Spatial and Temporal Data Mining Fundamental Concepts of Data and Knowledge > Big Data Mining Technologies > Artificial Intelligence
Comparison of machine learning and human brain learning
[ Normal View | Magnified View ]
The basic principles of AdaBoost
[ Normal View | Magnified View ]
The basic principles of random forest
[ Normal View | Magnified View ]
(a) The detection result generated by the RL‐CNN detection framework. (b) All nine ships can be detected by FFPN‐RL
[ Normal View | Magnified View ]
Agent and environment interaction diagram in reinforcement learning
[ Normal View | Magnified View ]
(a) Land use maps obtained by STDCNN for Kowloon and Hong Kong Island. (b) Land use maps obtained by STDCNN for Shenzhen
[ Normal View | Magnified View ]
Predicted classes over unseen aerial data. The model correctly detected the classes depicted in the images
[ Normal View | Magnified View ]
(a) Learning process of traditional machine learning. (b) Learning process of transfer learning. Traditional machine learning often attempts to learn knowledge from the beginning, while transfer learning attempts to transfer knowledge from previous tasks to new tasks when the latter has fewer high‐quality training data
[ Normal View | Magnified View ]
Long short‐term memory unit
[ Normal View | Magnified View ]
Unfolded basic recurrent neural network
[ Normal View | Magnified View ]
The typical CNN architecture. The ReLU (rectified linear unit) is the most commonly used activation function in deep learning models, corresponding to the commonly used sigmoid function in typical ANN. The main function of softmax is to map the nonstandardized output to the probability distribution on the predicted output category
[ Normal View | Magnified View ]
DNN. represents the weight vector, and represents the activation value of the jth neurons in the ith hidden layer
[ Normal View | Magnified View ]
ANN. represents the weight vector, and represents the activation value of the jth neurons in the hidden layer
[ Normal View | Magnified View ]
The basic principles of genetic algorithm
[ Normal View | Magnified View ]
The basic principles of hill‐climbing method
[ Normal View | Magnified View ]
The basic principles of SVM
[ Normal View | Magnified View ]
The basic principles of Bayesian network
[ Normal View | Magnified View ]
The basic principles of naive Bayes classifier. Conclusion is that there is a higher probability that the new sample (circle with question mark) belongs to the green class
[ Normal View | Magnified View ]
Schematic diagram of dynamic time warping algorithm. Curves C and Q on the left and above side represent the two‐time series, and the path formed by the yellow grid represents the shortest distance between Q and C
[ Normal View | Magnified View ]
The basic principles of KNN
[ Normal View | Magnified View ]
The basic principles of FP tree
[ Normal View | Magnified View ]
The basic principles of Apriori
[ Normal View | Magnified View ]
The basic principles of decision tree
[ Normal View | Magnified View ]
The basic principles of expert system
[ Normal View | Magnified View ]
AI‐based knowledge discovery methods and their commonly used algorithms
[ Normal View | Magnified View ]

Browse by Topic

Technologies > Artificial Intelligence
Fundamental Concepts of Data and Knowledge > Big Data Mining
Algorithmic Development > Spatial and Temporal Data Mining

Access to this WIREs title is by subscription only.

Recommend to Your
Librarian Now!

The latest WIREs articles in your inbox

Sign Up for Article Alerts