Production of the Japan 30-m Land Cover Map of 2013–2015 Using a Random Forests-Based Feature Optimization Approach
AbstractAchieving more timely, accurate and transparent information on the distribution and dynamics of the world’s land cover is essential to understanding the fundamental characteristics, processes and threats associated with human-nature-climate interactions. Higher resolution (~30–50 m) land cover mapping is expected to advance the understanding of the multi-dimensional interactions of the human-nature-climate system with the potentiality of representing most of the biophysical processes and characteristics of the land surface. However, mapping at 30-m resolution is complicated with existing manual techniques, due to the laborious procedures involved with the analysis and interpretation of huge volumes of satellite data. To cope with this problem, an automated technique was explored for the production of a high resolution land cover map at a national scale. The automated technique consists of the construction of a reference library by the optimum combination of the spectral, textural and topographic features and predicting the results using the optimum random forests model. The feature-rich reference library-driven automated technique was used to produce the Japan 30-m resolution land cover (JpLC-30) map of 2013–2015. The JpLC-30 map consists of seven major land cover types: water bodies, deciduous forests, evergreen forests, croplands, bare lands, built-up areas and herbaceous. The resultant JpLC-30 map was compared to the existing 50-m resolution JAXA High Resolution Land-Use and Land-Cover (JHR LULC) map with reference to Google Earth™ images. The JpLC-30 map provides more accurate and up-to-date land cover information than the JHR LULC map. This research recommends an effective utilization of the spectral, textural and topographic information to increase the accuracy of automated land cover mapping. View Full-Text
Scifeed alert for new publicationsNever miss any articles matching your research from any publisher
- Get alerts for new papers matching your research
- Find out the new papers from selected authors
- Updated daily for 49'000+ journals and 6000+ publishers
- Define your Scifeed now
Sharma, R.C.; Tateishi, R.; Hara, K.; Iizuka, K. Production of the Japan 30-m Land Cover Map of 2013–2015 Using a Random Forests-Based Feature Optimization Approach. Remote Sens. 2016, 8, 429.
Sharma RC, Tateishi R, Hara K, Iizuka K. Production of the Japan 30-m Land Cover Map of 2013–2015 Using a Random Forests-Based Feature Optimization Approach. Remote Sensing. 2016; 8(5):429.Chicago/Turabian Style
Sharma, Ram C.; Tateishi, Ryutaro; Hara, Keitarou; Iizuka, Kotaro. 2016. "Production of the Japan 30-m Land Cover Map of 2013–2015 Using a Random Forests-Based Feature Optimization Approach." Remote Sens. 8, no. 5: 429.
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.