Multispectral Analysis and Spectral Indices

How do you extract valuable information from multispectral satellite images? In this Blended Learning course, you’ll learn how to use Landsat and Copernicus/Sentinel data to analyze the characteristics and changes in vegetation, water, soil, and land cover. You’ll combine spectral bands and apply indices such as NDVI, NDWI, NBR, and SAVI in QGIS. This will teach you how to convert multispectral satellite data into useful spatial information for applications including nature conservation, agriculture, water management, fire analysis, and climate issues.

What problems can you solve using Multispectral Analysis and Spectral Indices?

How do you determine where vegetation is healthy, stressed, or damaged?
How do you map surface water, wetlands, and changes in water coverage?
How do you determine the extent and severity of fire damage using satellite imagery?
How do you distinguish between vegetation, bare soil, water, and built-up areas using spectral indices?
Which satellite source and spectral index are best suited for a specific spatial issue?

During this Blended Learning course, you’ll learn how to analyze and substantiate these issues using multispectral Earth observation data in QGIS. You’ll gain hands-on experience working with data from sources such as Sentinel-2 from the European Copernicus program and Landsat, and you’ll also be introduced to other important sources like MODIS, VIIRS, and high-resolution data such as PlanetScope. You’ll learn to combine different spectral bands and to select, calculate, and interpret indices to make the characteristics and changes in vegetation, water, soil, fire damage, and built-up areas measurable and spatially understandable.

The Theory Behind Multispectral Analysis and Spectral Indices

A good multispectral analysis begins with an understanding of how different materials reflect electromagnetic radiation. Vegetation, water, soil, and built-up areas each have their own spectral response. By combining measurements from different spectral bands, properties can be made visible and measurable that are difficult or impossible to recognize directly in a standard satellite image.

You’ll learn how visible light, near-infrared, red-edge, and short-wave infrared are used in multispectral analyses. You’ll discover, for example, why healthy vegetation responds differently in the red and near-infrared parts of the spectrum than stressed vegetation, and why water, bare soil, and built-up areas have yet other spectral characteristics.

In doing so, you’ll build on the principles of optical remote sensing and delve into how spectral differences can be converted into quantitative information. You’ll also learn that a spectral index is not an end result in and of itself. The meaning of an index depends on the landscape, the season, the sensor used, the available spectral bands, the spatial resolution, and the imaging conditions.

From Spectral Bands to the Right Index

Spectral indices combine two or more spectral bands according to a specific mathematical formula. This allows specific characteristics of the Earth’s surface to be highlighted and analyzed more effectively. You will therefore not only learn to calculate a number of commonly used indices, but above all to determine which index is appropriate for the problem at hand and whether the satellite sensor used contains the necessary spectral bands.

You will work with, among others:

NDVI for analyzing the presence, density, and condition of vegetation. This widely used vegetation index can be applied to Sentinel-2, Landsat, MODIS, VIIRS, and PlanetScope, among others.
EVI for vegetation analysis that minimizes the influence of factors such as the atmosphere and soil. EVI is widely used in large-scale vegetation monitoring with, for example, MODIS.
GNDVI for vegetation analysis using the green and near-infrared bands, for example, to investigate differences in vegetation condition.
NDWI for analyzing water and moisture-related patterns using appropriate spectral bands.
MNDWI for enhancing surface water, especially when water needs to be distinguished from buildings and other land cover.
NBR for identifying burned areas and analyzing changes and fire damage following wildfires.
SAVI for vegetation analysis in areas where bare soil has a relatively large influence on the spectral signal.
NDBI for identifying and analyzing built-up areas.
BSI for identifying bare soil and distinguishing soil from vegetation and other types of land cover.

You will learn why certain bands are combined in these indices, how to interpret the results, and what limitations each index has. This will teach you to avoid automatically applying an index without first assessing whether it is actually suitable for the specific area and problem at hand.

Working with Different Sources for Multispectral Analysis and Spectral Indices

Not every satellite sees the same way, and not every satellite source is suitable for every issue. That is why you will learn about and compare different multispectral satellite systems. You will examine available spectral bands, spatial resolution, temporal resolution, historical availability, and potential applications.

Among other things, you’ll be introduced to:

Sentinel-2 from the Copernicus program for detailed analyses of vegetation, agriculture, water, soil, and land cover. The available red-edge bands offer additional capabilities for vegetation analysis.
Landsat for multispectral analyses and investigating long-term changes in vegetation, water, land use, and fire damage, among other things.
MODIS for frequent observations of large areas and monitoring of, for example, vegetation and changes over time, with NDVI and EVI being widely used.
VIIRS for frequent Earth observation on a regional and global scale and applications such as vegetation, land surface, and fire monitoring.
PlanetScope as an example of commercial high-resolution data that enables multispectral analyses to be performed at a more detailed spatial scale.

You will learn not to think in terms of a single satellite product, but rather in terms of the spatial problem at hand. You first determine what information is needed and then select which sensor, spectral bands, and indices are most suitable for that purpose.

Selecting the Right Satellite Data

A reliable multispectral analysis starts with the right data. You’ll learn to select satellite images based on the study area, acquisition date, season, cloud cover, available spectral bands, spatial resolution, and repeat frequency.

For example, you’ll compare the detailed multispectral capabilities of Sentinel-2 with the long historical data series of Landsat and the high temporal resolution of MODIS and VIIRS. For applications requiring a high level of spatial detail, you’ll explore the capabilities of high-resolution data such as PlanetScope.

This will help you understand that the best satellite source depends on the specific problem at hand. A different sensor may be suitable for monitoring global or regional vegetation trends than for analyzing differences between individual agricultural plots.

Calculating and Analyzing Spectral Indices in QGIS

After covering the theoretical basics, you’ll get hands-on experience working with real multispectral satellite data in QGIS. You’ll learn how to load, visualize, select, and combine different spectral bands, and use raster calculations to create spectral indices.

You’ll calculate and examine indices such as NDVI, EVI, GNDVI, NDWI, MNDWI, NBR, SAVI, NDBI, and BSI. You’ll learn to visualize, classify, and analyze the resulting raster maps and examine how index values vary spatially.

This involves more than just applying a formula. You’ll learn to determine which bands from a specific sensor are needed, how to use the correct bands, and how to interpret the results in terms of their meaning.

In addition, you’ll combine different indices and other geoinformation to gain a more complete picture of an area. For example, you can combine vegetation indices with soil data, land use, or water information to better understand why certain spatial patterns emerge.

Comparing Indices and Satellite Sources

A single index or satellite source does not always provide the best answer. That is why you will learn to compare different indices and data sources with one another.

For example, you’ll investigate when NDVI provides sufficient information and when EVI, SAVI, or GNDVI can offer a better picture of vegetation. For surface water, you’ll compare NDWI and MNDWI. When addressing land cover issues, you can combine vegetation, soil, and land-use indices.

You’ll also compare results from different satellite sensors. You’ll learn to assess which differences actually provide meaningful insights into the Earth’s surface and which are caused by spectral bands, spatial resolution, the time of acquisition, or sensor characteristics.

This helps you develop an important professional skill: not automatically using a familiar index, but making a well-reasoned determination of which combination of satellite source, spectral bands, and index best addresses the issue at hand.

Analyzing Changes with Multispectral Data

An important application of multispectral satellite data is studying changes over time. You’ll learn to analyze satellite images from different periods and compare index values to make trends visible and measurable.

For example, you’ll investigate changes in vegetation condition, water surface area, fire damage, bare soil, or land development. In doing so, you’ll learn to account for season, weather conditions, cloud cover, differences between sensors, and spatial resolution.

You’ll discover why Landsat, with its long historical data series, is valuable for long-term analyses, while Sentinel-2 can provide greater spatial and spectral detail, and MODIS and VIIRS are suitable for frequent monitoring of large areas.

This also lays the foundation for the specialized topics of change detection and time-series analysis, in which longer sequences of satellite images are systematically used to investigate developments over time.

From Spectral Index to Spatial Information

During the Blended Learning program, you’ll work with realistic real-world examples and actual multispectral Earth observation data. You’ll step into the role of a remote sensing specialist and tackle issues similar to those faced by government agencies, water boards, conservation organizations, agricultural organizations, and consulting firms.

You won’t start with a specific satellite or index, but rather with the problem itself. You’ll then determine which satellite source, spectral bands, and indices are appropriate, perform the analysis in QGIS, assess the reliability of the results, and translate them into usable spatial information.

You’ll work on assignments such as:

A nature conservation manager wants to know where vegetation is deteriorating. Compare NDVI, EVI, SAVI, or another suitable vegetation index, and justify which index provides the best insight into vegetation condition.
A water management agency wants to investigate changes in surface water. Compare NDWI and MNDWI and determine which index best distinguishes the water in the study area.
A conservation organization wants to know which areas were affected by a wildfire. Use NBR and appropriate satellite imagery to analyze the extent and spatial variations of the fire damage.
An agricultural organization wants to investigate differences between agricultural parcels. Compare NDVI, SAVI, GNDVI, and, where possible, red-edge data, and determine which approach yields the most useful information.
A municipality wants to distinguish between built-up areas, bare soil, vegetation, and water. Combine NDBI, BSI, and appropriate vegetation and water indices to analyze the different types of land cover.
An organization wants to track a development over several years. Compare Sentinel-2, Landsat, MODIS, or VIIRS and justify which data source and indices best align with the desired spatial and temporal scale.

By the end of the course, you will have the knowledge and practical skills to independently analyze multispectral Earth observation data from various satellite systems in QGIS. Based on a spatial problem, you will be able to determine which satellite source, spectral bands, and indices are appropriate, calculate and interpret these indices, and assess the reliability of the results. This enables you to translate multispectral satellite data into well-founded spatial information about vegetation, water, soil, fire damage, built-up areas, and changes in the landscape.

Enroll

€395,-

Start: 1-hour online session
Self-study: Review course materials
End: 1-hour online session

Register for this course

You’ll receive 1-on-1 guidance. After signing up, our course coordinator will contact you to schedule your first session.

Learning Objectives

Explain how multispectral analysis and spectral indices are used to analyze the characteristics and changes in vegetation, water, soil, fire damage, and built-up areas.
Identify and explain the spectral characteristics of vegetation, water, soil, and built-up areas using visible light, near-infrared, red-edge, and short-wave infrared.
Compare various multispectral satellite sources, including Sentinel-2, Landsat, MODIS, VIIRS, and high-resolution data such as PlanetScope, based on their spectral, spatial, and temporal characteristics.
Based on a spatial problem, determine which satellite source, spectral bands, and resolution are most suitable for the analysis.
Explain the functioning and potential applications of spectral indices such as NDVI, EVI, GNDVI, NDWI, MNDWI, NBR, SAVI, NDBI, and BSI.
Determine which spectral index or combination of indices is most suitable for a specific issue related to vegetation, water, soil, fire damage, or land cover.
Select suitable multispectral satellite images based on the study area, acquisition date, season, cloud cover, available spectral bands, spatial resolution, and repeat frequency.
Load, visualize, and process multispectral satellite data in QGIS, and combine different spectral bands.
Calculate, visualize, and classify spectral indices in QGIS, and interpret the resulting spatial patterns in terms of their content.
Compare different indices, satellite sources, and acquisition times, and assess which combination provides the most reliable information.
Assess the influence of sensor, season, atmosphere, cloud cover, spatial resolution, and acquisition conditions on the results of multispectral analyses.
Translate multispectral analyses into substantiated spatial information for practical applications in areas such as nature conservation, agriculture, water management, fire analysis, and spatial development.

Want to know more?

Do you have questions about the course content? Or are you unsure whether the course aligns with your learning goals or preferences? Would you prefer an in-house or private course? We’d be happy to help.

FAQs on Blended Learning, Multispectral Analysis, and Spectral Indices

In multispectral analysis, measurements from different parts of the electromagnetic spectrum are combined to make properties of the Earth’s surface visible and measurable. Spectral indices are calculations that combine two or more spectral bands. Examples include NDVI and EVI for vegetation, NDWI and MNDWI for water, NBR for fire damage, SAVI and GNDVI for vegetation condition, NDBI for built-up areas, and BSI for bare soil. Which index is appropriate depends on the specific problem at hand as well as on the spectral bands made available by the satellite sensor being used.

Sentinel-2 data is available for free through the Copernicus Data Space Ecosystem. Sentinel-2 features 13 spectral bands in visible light, near-infrared, red-edge, and short-wave infrared. This makes this data source suitable for a wide range of indices, including NDVI, GNDVI, NDWI, MNDWI, NBR, SAVI, NDBI, and BSI. The red-edge bands also make Sentinel-2 particularly useful for detailed vegetation and agricultural analyses. The downloaded images and individual bands can be loaded into QGIS and combined to calculate the indices yourself.

Landsat data is available for free through USGS EarthExplorer and other data services provided by the U.S. Geological Survey. Landsat offers a long historical data series and spectral bands in visible light, near-infrared, and short-wave infrared, among others. This allows for the calculation of indices such as NDVI, EVI, NDWI, MNDWI, NBR, SAVI, NDBI, and BSI. Landsat is particularly useful when studying changes over longer periods, such as changes in vegetation, land use, surface water, or fire damage. The required bands can be downloaded and then processed in QGIS.

MODIS and VIIRS data are available free of charge through NASA Earthdata and various NASA and NOAA data portals. These satellite sensors have a coarser spatial resolution than Sentinel-2 and Landsat, but they observe large areas very frequently. This makes them particularly well-suited for regional, continental, and global monitoring, as well as for time series. For MODIS, ready-to-use products are also available in which indices such as NDVI and EVI have already been calculated. VIIRS is also used for vegetation, land surface, and fire monitoring. Depending on the specific application, you can therefore either calculate indices yourself from spectral bands or use existing derived products.

For analyses requiring greater spatial detail, PlanetScope data from Planet can be used, for example. This commercial satellite data has a higher spatial resolution than many freely available satellite sources and is therefore useful for applications such as precision agriculture, vegetation monitoring, and the analysis of local changes. Which indices can be calculated depends on the available bands of the product in question. For example, with suitable red and near-infrared bands, NDVI can be calculated. Some PlanetScope products also feature additional bands that enable the calculation of more vegetation indices. For each satellite product, you should therefore first check which spectral bands are available before determining which index you can calculate in QGIS.