- English
This dataset contains supplementary material for the journal paper by Johnson et al. (2017). In the paper, OpenStreetMap crowdsourced data was used in combination with different types of satellite imagery to identify and map land cover changes in the Laguna de Bay area of the Philippines. The dataset provides the training data (derived from the OpenStreetMap data) and validation data (based on visual interpretation of high-resolution images in Google Earth) used in the paper.
The objective is to build a predictive model using the training data, and use the model to predict land cover changes at other locations. The validation dataset allows for the accuracy of these land cover change predictions to be assessed. The training dataset is very large but error-prone due to its crowdsourced nature, making the prediction challenging.
File descriptions
OSM_training_data.xls: Table with land cover information derived from the OpenStreetMap "landuse" and "natural" polygon datasets, and spectral information derived from Landsat and ALOS-PALSAR mosaic satellite images. The spectral information can be used to help detect land cover changes.
validation_data.xls: Table with land cover information derived from visual interpretation of Google Earth images, and spectral information derived from Landsat and ALOS-PALSAR mosaic satellite images. Used for accuracy assessment.
Field descriptions
LC_2007: Land cover in 2007 (b = built-up, c = cropland/grassland, t = tree, w = water)
LC_2015: Land cover in 2015 (b = built-up, c = cropland/grassland, t = tree, w = water)
ndvi2007: Maximum normalized difference vegetation index (NDVI) value for the year 2007, extracted from Landsat 5 satellite imagery, rescaled to 0-1 range.
hv2007: PALSAR-1 mosaic gamma nought value (HV polarization) in the year 2007, rescaled to 0-1 range.
hh2007: PALSAR-1 mosaic gamma nought value (HH polarization) in the year 2007, rescaled to 0-1 range.
ndvi2015: Maximum normalized difference vegetation index (NDVI) value for the year 2015, extracted from Landsat 8 satellite imagery, rescaled to 0-1 range.
hv2015: PALSAR-2 mosaic gamma nought value (HV polarization) in the year 2015, rescaled to 0-1 range.
hh2015: PALSAR-2 mosaic gamma nought value (HH polarization) in the year 2015, rescaled to 0-1 range.
Reference
Johnson, B. A., Iizuka, K., Bragais, M., Endo, I., Magcale-Macandog, D. (2017). Employing crowdsourced geographic data and multi-temporal/multi-sensor satellite imagery to monitor land cover change: A case study in an urbanizing region of the Philippines. Computers, Environment and Urban Systems, 64, 184-193.
- English