A Bahani, Pervin Ong creator Murai, Shunji Chairperson Vilas Wuwongse Examination Committee Ochi, Shiro Examination Committee The Government of Japan Scholarship Donor text th Bangkok Asian Institute of Technology 1994 monographic eng

print

56 leaves In this study, a technique for land cover classification was proposed and implemented. The proposed method uses a combination of supervised and unsupervised classification. Split and merge clustering is applied to the selected training data and the output clusters will be used in the assignment of the unknown pixel of the image to be classified. Two existing conventional methods namely, maximum likelihood classification method (MLC) and minimum distance classification method (MDC) were also implemented. The first method is based on statistical probabilities, the second method uses distance as a discriminant to classify an unknown input pixel. Test sites were selected and the performances of the three implemented classification methods were compared. The test sites were classified into four types. The classification results showed that the proposed method can give better classification accuracy, however classification time is longer for the proposed classification method compared to both MLC and MDC. A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science, School of Engineering and Technology Thesis (M.Sc.) - Asian Institute of Technology, 1994 Remote sensing Asian Institute of Technology. http://203.159.5.9/ait-thesis/detail.php?q=B15767 http://203.159.5.9/ait-thesis/detail.php?q=B15767 080998 20260818084540.0