|
Hokkaido University Collection of Scholarly and Academic Papers >
Graduate School of Environmental Science / Faculty of Environmental Earth Science >
Peer-reviewed Journal Articles, etc >
Hierarchical classification of land use types using multiple vegetation indices to measure the effects of urbanization
| Title: | Hierarchical classification of land use types using multiple vegetation indices to measure the effects of urbanization |
| Authors: | Shishir, Sharmin Browse this author | | Tsuyuzaki, Shiro Browse this author →KAKEN DB |
| Keywords: | Fine-scale data | | Hierarchical classification | | Decision tree | | Land use change | | Shorea forest |
| Issue Date: | Jun-2018 |
| Publisher: | Springer |
| Journal Title: | Environmental monitoring and assessment |
| Volume: | 190 |
| Issue: | 6 |
| Start Page: | 342 |
| Publisher DOI: | 10.1007/s10661-018-6714-3 |
| Abstract: | Detecting fine-scale spatiotemporal land use changes is a prerequisite for understanding and predicting the effects of urbanization and its related human impacts on the ecosystem. Land use changes are frequently examined using vegetation indices (VIs), although the validation of these indices has not been conducted at a high resolution. Therefore, a hierarchical classification was constructed to obtain accurate land use types at a fine scale. The characteristics of four popular VIs were investigated prior to examining the hierarchical classification by using Purbachal New Town, Bangladesh, which exhibits ongoing urbanization. These four VIs are the normalized difference VI (NDVI), green-red VI (GRVI), enhanced VI (EVI), and two-band EVI (EVI2). The reflectance data were obtained by the IKONOS (0.8-m resolution) and WorldView-2 sensor (0.5-m resolution) in 2001 and 2015, respectively. The hierarchical classification of land use types was constructed using a decision tree (DT) utilizing all four of the examined VIs. The accuracy of the classification was evaluated using ground truth data with multiple comparisons and kappa (kappa) coefficients. The DT showed overall accuracies of 96.1 and 97.8% in 2001 and 2015, respectively, while the accuracies of the VIs were less than 91.2%. These results indicate that each VI exhibits unique advantages. In addition, the DT was the best classifier of land use types, particularly for native ecosystems represented by Shorea forests and homestead vegetation, at the fine scale. Since the conservation of these native ecosystems is of prime importance, DTs based on hierarchical classifications should be used more widely. |
| Rights: | The final publication is available at link.springer.com |
| Type: | article (author version) |
| URI: | http://hdl.handle.net/2115/74518 |
| Appears in Collections: | 環境科学院・地球環境科学研究院 (Graduate School of Environmental Science / Faculty of Environmental Earth Science) > 雑誌発表論文等 (Peer-reviewed Journal Articles, etc)
|
Submitter: 露崎 史朗
|
