Mapping Deforestation and Age of Evergreen Trees by Applying a Binary Coding Method to Time-Series Landsat November Images

摘要

This paper proposes a binary coding method, a novel postclassification change detection method that indexes multitemporal satellite images into a single information layer. As a case study, this method is applied to the production of a deforestation map and a tree age map of evergreen trees. Seven images of Landsat-5 Thematic Mapper (TM) and Landsat-7 Enhanced TM Plus of Kangwon province in Korea, all obtained in November from 1984 to 2005, were used. Seven forestonforest bitmaps were produced by the normalized difference vegetation index thresholding. After the maximum value composite, the binary coding method was applied to convert the five bitmaps into a singlelayered binary index map. This simple but intuitive method produced a deforestation map and a tree age map. The deforestation map contains the history of forest fires and massive replantation while the tree age map shows the age of evergreen trees with 30 years of range, both in approximately five-year temporal resolution. Comparison of the results with a digital forest inventory, forest fire records, and field investigations showed good agreement on the south-facing sunny slopes. Errors from geometric and radiometric data processing and inaccuracies of the digital forest inventory were also identified.