Remote sensing basics

 

On-line textbooks

Principles of remote sensing: an introductory textbook. K. Tempfli, G.C. Huurneman, W.H. Bakker, L.L.F. Janssen, et al. Fourth edition. University of Twente, Nethrelands, 2009
ISBN 978-90-6164-270-1

Available for on-line reading: http://www.itc.nl/Pub/Home/library/Academic_output/ITC-GIS-and-Remote-Sensing-Textbooks.html

http://glad.geog.umd.edu/Potapov/_Library/Principles_of_Remote_Sensing.pdf

 

Fundamentals of remote sensing. Canada Centre for Mapping and Earth Observation, Natural Resources Canada

Available as PDF: http://www.nrcan.gc.ca/earth-sciences/geomatics/satellite-imagery-air-photos/satellite-imagery-products/educational-resources/9309

http://glad.geog.umd.edu/Potapov/_Library/Fundamentals_of_Remote_Sensing.pdf

 

On-line lectures/materials

Fundamentals of Remote Sensing. NASA/ARSET

PDF lectures and video recording: https://arset.gsfc.nasa.gov/webinars/fundamentals-remote-sensing

http://glad.geog.umd.edu/Potapov/_Library/Principles_of_Remote_Sensing.pdf

http://glad.geog.umd.edu/Potapov/_Library/Fundamentals_RS_Session2_Land_Final.pdf

 

Principles of remote sensing. Centre for Remote Imaging, Sensing and Processing. National University of Singapore

On-line learning materials: http://www.crisp.nus.edu.sg/~research/tutorial/rsmain.htm

 

Science Education through Earth Observation for High Schools (SEOS) Project

On-line learning materials: http://www.seos-project.eu/home.html

 

UMD-GLAD Methodology overview

 

The UMD-GLAD Landsat-based land cover monitoring approach is an evolution of the global MODIS-based land cover monitoring. Our global forest cover change estimates based on MODIS data and stratified sampling of Landsat images were published in 2010:

http://glad.geog.umd.edu/Potapov/_Library/Hansen_PNAS_2010.pdf

http://glad.geog.umd.edu/Potapov/_Library/Hansen_PNAS_2010_SI.pdf

 

After the opening of the Landsat archive in 2008, UMD developed a system for automatic data processing and classification at national scale. The outputs of our data processing system are multi-temporal metrics that are similar to what we derived from MODIS, but at 30 m spatial resolution. The algorithm of automatic Landsat data processing/calibration, multi-temporal metrics, and metric analysis were published in Remote Sensing of Environment:

http://glad.geog.umd.edu/Potapov/_Library/Hansen_at_al_2008.pdf

http://glad.geog.umd.edu/Potapov/_Library/Potapov_at_al_2012.pdf

 

Paper describing the general Landsat-based data processing and characterization algorithm and its comparison with other approaches:

http://glad.geog.umd.edu/Potapov/_Library/Hansen_RSE_2012.pdf

 

The image classification and change detection are based on decision tree algorithm. Here are some papers describing this algorithm and its use for image classification:

http://glad.geog.umd.edu/Potapov/_Library/Hansen_Characterizing_Land_Cover.pdf

http://glad.geog.umd.edu/Potapov/_Library/DeFries1998.pdf

 

Data processing and analysis system to map forest cover change globally from 2000 to 2012. The results were published in Science in 2013:

http://glad.geog.umd.edu/Potapov/_Library/Hansen_Science_2013.pdf

http://glad.geog.umd.edu/Potapov/_Library/Hansen_Science_2013_SM.pdf

Global forest maps are available on-line:

http://earthenginepartners.appspot.com/science-2013-global-forest

Results also available through Global Forest Watch data portal:

http://www.globalforestwatch.org/

 

The global forest cover change map shows all types of tree cover dynamics (within primary forests, secondary forests, tree plantations). For deforestation assessment and forest conservation analysis, it is important to know the area of primary forest loss and degradation. Such analyses were performed for Indonesia and Central Africa:

http://glad.geog.umd.edu/Potapov/_Library/Margono_ERL_2012.pdf

http://glad.geog.umd.edu/Potapov/_Library/Margono_NCC_2014.pdf

http://glad.geog.umd.edu/Potapov/_Library/Zhuravleva_ERL_2013.pdf

 

Precise estimation of forest cover change map uncertainty is important for further use of change area to estimate carbon emissions and other ecosystem services implications. We developed and tested a system for national-wide data validation, which was prototyped in Central Africa:

http://glad.geog.umd.edu/Potapov/_Library/Tyukavina_ERL_2013.pdf

The same approach was implemented for the entire tropical biome to estimate aboveground carbon loss in natural and managed forests from 2000 to 2012:

http://glad.geog.umd.edu/Potapov/_Library/Tyukavina_et_al_2015.pdf

 

Our current activities include adaptation of our global data processing system for national monitoring agencies, as was done for Peru:

http://glad.geog.umd.edu/Potapov/_Library/Potapov_ERL_2014.pdf

 

A comprehensive system of tree cover mapping and monitoring was developed and prototyped for Bangladesh. The system is based on the national Landsat-based forest extent and change mapping and area estimation using stratified samples.

PAPER IN REVIEW, DO NOT SHARE

http://glad.geog.umd.edu/Potapov/_Library/Potapov_Bangladesh_2017.pdf

PCI Geomatica Tutorials

The UMD-GLAD methodology is using PCI Geomatica as one of the image visualization environment. PCI Geomatica is one of the leading commercial image processing tools. PCI offers a 30-day trial version of the fully functional Geomatica software (web-site registration required). It also provides a freeware data viewer: http://www.pcigeomatics.com/geomatica-freeview-download

For participants interested in learning PCI software we recommend the following training materials provided by PCI:

http://glad.geog.umd.edu/Potapov/_Library/TrainingGuide-Geomatica-1.pdf

http://glad.geog.umd.edu/Potapov/_Library/TrainingGuide-Geomatica-2.pdf

 

Statistical Sampling

Primary materials concerning the GLAD sampling approach:

Olofsson et al. (2013, 2014) - simple explanation, but it works only for yes/no sample interpretation (doesn't work for proportions for each sample, e.g. 0.5, 0.25, etc.) and only for cases when sampling strata match map classes exactly. This is a good introductory paper, but it's practical use is limited due to the issues listed above.

http://glad.geog.umd.edu/Potapov/_Library/Oloffson.%20Good%20practices%20accuracy%20assessment.pdf

http://glad.geog.umd.edu/Potapov/_Library/Oloffson.%20Good%20practices%20accuracy%20assessment.pdf

Stehman et al. (2014) - this is the paper all our area and accuracy calculations are based on. Works for any strata. Works for sample interpretation with proportions of target classes for each sample (e.g. 0.5, 0.25, etc.)

http://glad.geog.umd.edu/Potapov/_Library/Stehman_RSE_2014.pdf

Tyukavina et al. (2013) - example of approach from Stehman et al. (2014), refer to section 3.1.2. "Estimating area of forest loss and its uncertainty". The GLAD area and SE calculation script uses equations 2 and 3 from Tyukavina et al. (2013), which are based on indicator functions from Stehman et al. (2014).

http://glad.geog.umd.edu/Potapov/_Library/Tyukavina_ERL_2013.pdf

GFOI (2016) Integration of remote-sensing and ground-based observations for estimation of emissions and removals of greenhouse gases in forests. This is a comprehensive set of guidelines on calculating uncertainties

http://glad.geog.umd.edu/Potapov/_Library/MGD2.0_English.pdf

http://glad.geog.umd.edu/Potapov/_Library/MGD2.0_French.pdf

Cochran W.G. Sampling Techniques. (1977). A comprehensive textbook on sampling methods

http://glad.geog.umd.edu/Potapov/_Library/Cochran_1977_Sampling_Techniques_Third_Edition.pdf