@DANE_Colombia /DANEColombia /DANEColombia
National Administrative Department of Statistics –DANE
Colombia
October 2015
Using satellite images to calculate land use and land cover
statistics
Background (1)
Objective
• Explore the usefulness of Big Data from satellite land images for theupdating and improvement of statistic information.
• Calculate an indicator that shows the urban agglomeration degree by usingsatellite images and population data.
• This indicator would provide additional information to the framework forthe implementation of several goals referred to in health and food security.
Data collection
Changesdetection
Ratescalculation
Indicatorcalculation
Satelliteimagescollection
Populationdata collection
Image pre-processing
Imageclassification
Populationgrowth rate
Urban landconsumptionrate
Ratio of therates
Methodology
Imagery collection
Landsat 2005 Landsat 2010Landsat 2015
Criteria:
• Availability
• Quality (percentage of cloud cover)
• Acquisitions at the same time of the year
• Spatial coverage
• Similar spectral bands
http://earthexplorer.usgs.gov/
Imagery pre-processing
Original 2005 Image Pre-processed image
Semantic net
Rules and operatorsdefinition
Coveragesgeneration
Coverages generation with object-oriented classification
Coverage map
Water
Urban/Built - Up
Grassland
Barren/Minimal Vegetation
Forest, Shrub and Scrubland
Coverage map 2005Coverage map 2010Coverage map 2015
Map Legend
Approach to the Land Consumption Rate (ULCR)
)(
)( )(
1
12
tAreaLandUrban
tAreaLandUrbantAreaLandUrbanLCR
10.807
11.288
12.178
2005 2010 2015
Barranquilla Metropolitan Area
Urban hectares
8%4%
Approach to the Population Growth Rate (PGR)
)(
)()(
1
12
tPopulation
tPopulationtPopulationPGR
1,769,191
1,898,007
2,025,071
2005 2010 2015
Barranquilla Metropolitan Area
Population
7%7%
*Population projections
* *
Proposed Indicator
RateGrowthPopulation
RatenConsumptioLandIIndicator )(
I=0.642005-2010 I=1.13
2010-2015
* *
* Preliminary data
• Big Data from satellite imagery is a powerful tool to increase thepotential applications of statistics and to improve the quality ofstatistics data. This integration allows the exploration of newmethodologies and the improvement of those already available.
• The benefits generated by the use of remote sensing, create theopportunity to make the information more traceable and allow thedesign of different types of multitemporal studies.
• This project is a good example for the use of open source software andsatellite images, which are both available on the web for free.
Comments
@DANE_Colombia /DANEColombia /DANEColombia
[email protected]@dane.gov.co