Water body extraction from Landsat ETM+ imagery using adaboost algorithm

Extracting water body information accurately from remotely sensed imagery is significant for surveying, planning and protecting water resources, and particularly flood disaster management such as monitoring, evaluation, emergency response and so on. Landsat ETM+ imagery has such features as high spatial resolution and multi-spectral resolution, which provides a rich, reliable and accurate source of fundamental data for research on water body extraction. Now, researchers have investigated lots of techniques and methods for deriving water information automatically. Previous work was often devoted to finding a sophisticated classifier to identify water body, and it is not only difficult to make full use of available features, but also easy to get into predicament of training the complex classifier. In this paper, a novel scheme for water body extraction from Landsat ETM+ imagery using the adaboost algorithm is proposed. It is based on such consideration that finding many rough rules of thumb can be a lot easier and more effective than finding a single, highly prediction rule. Adaboost is a general strategy for learning classifiers by combining simple ones. The idea of adaboost is to take a “weak classifier” - that is, any classifier that will do at least slightly better than chance - and use it to build a much better classifier, thereby boosting the performance of the weak classification algorithm. The excellent property of adaboost is the ability to integrate disparate classifiers that concentrate on different aspects of the problem, and place more weight on features that could train more accurate base classifiers. This paper, with Landsat ETM+ imagery as study object data, combines weak classifiers which are constructed by spectral information of each band, water index and relationship between spectrums to form a strong water body extraction classifier. The experimental results show that this method can achieve perfect performance, and is more effecti- - ve than traditional algorithms.