5. Conclusion
This study presented a method for forest cover classification using Landsat ETM+ data appending with time series MODIS NDVI data, and confirmed that time series NDVI features had significant effort on improving classification accuracy of fine resolution remote sensing data. The forest cover classification in North China region shown that NDVI features extracted from time series fused NDVI data could improve the overall classification accuracy approximately 5% compared to only using a single Landsat ETM+ data. This study provided an illustration of forest cover classification method integrating temporal and spatial information from different resolution remote sensing data, and this method could be expanded to more complex study of land cover classification using remote sensing data.However, only basic statistic features of time series fused NDVI data were investigated for forest cover classification, more significant features would be investigated in the future work. The fusion strategy between coarse and high spatial resolution NDVI data was another issue to further study. In conclusion, time series vegetation index data contained abundant vegetation growth information which was a helpful complementary data for land cover classification using high spatial resolution remote sensing data, especially for vegetation type classification.
5. Conclusion
This study presented a method for forest cover classification using Landsat ETM+ data appending with time series MODIS NDVI data, and confirmed that time series NDVI features had significant effort on improving classification accuracy of fine resolution remote sensing data. The forest cover classification in North China region shown that NDVI features extracted from time series fused NDVI data could improve the overall classification accuracy approximately 5% compared to only using a single Landsat ETM+ data. This study provided an illustration of forest cover classification method integrating temporal and spatial information from different resolution remote sensing data, and this method could be expanded to more complex study of land cover classification using remote sensing data.However, only basic statistic features of time series fused NDVI data were investigated for forest cover classification, more significant features would be investigated in the future work. The fusion strategy between coarse and high spatial resolution NDVI data was another issue to further study. In conclusion, time series vegetation index data contained abundant vegetation growth information which was a helpful complementary data for land cover classification using high spatial resolution remote sensing data, especially for vegetation type classification.
การแปล กรุณารอสักครู่..
