The Comparison of Classification Algorithms for Remote Sensing Images

Although a large number of remote sensing image classification algorithm is proposed, but rarely performance of them are compared with each other. In this research, we classify satellite remote sensing images using two unsupervised classification algorithm and eight supervised classification algorithm which includes a number of common algorithms twenty years, tested and compared.  Our analysis focused on satellite images of the 12 spectral bands.  In comparison of these algorithms, the number of training samples is equated. These algorithms are compared with each other in terms of complexity, accuracy and reliability. The results show that accuracy of classification directly proportion to the number of training samples. The user can also choose efficient algorithms, depending on the complexity, accuracy and reliability of each parameter.