Co-training with Clustering for the Semi-supervised Classification of Remote Sensing Images
The collection of labeled data to train a classifier is very difficult, time-consuming, and expensive in the area of remote sensing. To solve the classification problem with few labeled data, many semi-supervised techniques have been developed and explored for the classification of remote sensing images. Self-learning and co-training techniques are widely explored for the semi-supervised classification of remote sensing images. In this paper, a co-training model with clustering is proposed for the classification of remote sensing images. To show effectiveness of the proposed technique, experiments have been performed on two different spectral views of hyperspectral remote sensing images using support vector machine as supervised classifier and semi-supervised fuzzy c-means as clustering technique. The experimental results show that co-training with clustering technique performs better than the traditional co-training algorithm and self-learning semi-supervised technique for the classification of remotely sensed images.
KeywordsCo-training Remote sensing image classification Self-learning Semi-supervised fuzzy c-means Semi-supervised learning Support vector machine
- 1.Li, J., Dias, J.M.B., Plaza, A.: Semi supervised self-learning for hyperspectral image classification. IEEE Trans. Geosci. Remote Sens. 51(7), 318–322 (2013)Google Scholar
- 13.Bensaid, A.M., Hall, L.O., Bezdek, J.C., Clarke, L.P P.: Partially supervised clustering for image segmentation. Pattern Recogn. 29, 370–379 (1996)Google Scholar