A few months ago, in Shenzhen, China, a huge disaster happened. Shenzhen is in the south of China. Dust and ashes could be seen rising from the buildings collapsing as the landslide rushes down a hill. The death toll is 69 with 8 people reported missing. It was an industrial accident due to human negligence rather than a natural disaster. The local police had arrested some of the people involved in the irregularities of the huge waste dump which was built up in previous two years. In this post using Landsat 8 imagery data, we will conduct change detection in the area.
Bellow are some images of the accident.
The standard processing parameters of Landsat 8 OLI and TIRS according to USGS are:
|Processing:||Level 1 T- Terrain Corrected|
|Pixel Size:||OLI multispectral bands 1-7,9: 30-meters
OLI panchromatic band 8: 15-meters
TIRS bands 10-11: collected at 100 meters but resampled
to 30 meters to match OLI multispectral bands
|Data Characteristics:||GeoTIFF data format
Cubic Convolution (CC) resampling
North Up (MAP) orientation
Universal Transverse Mercator (UTM) map projection (Polar Stereographic projection for scenes with a center latitude greater than or equal to -63.0 degrees)
World Geodetic System (WGS) 84 datum
12 meter circular error, 90% confidence global accuracy for OLI
41 meter circular error, 90% confidence global accuracy for TIRS
16-bit pixel values
So the next step after downloading the imagery data from the USGS’s Earth Explorer, is to use those two images, i.e one before and one after the disaster and apply the Normalized Difference Vegetation Index (NDVI) on both. The images are from the same sensor, so it doesn’t matter to provide an atmospheric correction i.e. transforming the Landsat images from DN values to radiance values.
By subtracting those two images the final product will show the extend of the disaster. The NDVI index is a simple graphical indicator and assess whether the target being observed contains live green vegetation or not.
On the NDVI 2015 and 2016 maps with green color represent the live green vegetation and the DNDVI (i.e. Subtracting the after the accident NDVI image from the before the accident NDVI image) displays with red, the areas that have big difference before and after the disaster. With blue colors are the areas with pretty much the same vegetation. The affected area is approximately 1500 square kilometers. Also the result is a bit distorted because the images are taken on different seasons.