Exploring Landsat Images with LandsatAnalysis Software
Purpose: Explore the spatial and spectral data available in Landsat imagery to study land cover. The software is designed to use the jpegs of Landsat images available on the United States Geological Survey’s Earthshots website: http://edcwww.cr.usgs.gov/earthshots/slow/tableofcontents
NOTE: The images that follow are not interactive. Please download the software (OSX OS9 PC) and follow along with the mini tutorial.
Background: Landsat
The Landsat satellite program has a long and significant history in monitoring our Earth. The first images of the Earth's land features were taken in 1972, and for over 40 years, the data has been used to study the many interconnected processes changing our land cover.
Background Information: Normalized Difference Vegetation Index
A Vegetation Index is a measure of the amount of healthy vegetation covering the Earth's surface.
Healthy vegetation absorbs visible light, especially red and blue light, and reflects much of the infrared.
A simple Vegetation Index is based on the difference between these intensities.
Another index, the "Normalized Difference Vegetation Index" or NDVI, provides a more consistent identification of vegetation as well as indicating the health of plants.
NDVI is the difference between the intensities of the reflected infrared and visible red light divided by the sum of the intensities of the two light measurements.
This mathematical manipulation compensates for areas experiencing hazy sunshine compared to those in clear skies as well as reduces shading and sloping topography.
Dense, healthy vegetation produces NDVI values near +1.0.
Bare soil and rock reflect similar levels of infrared and red light, so these surfaces produce NDVI values near 0.
Clouds, water, and snow reflect more visible light than infrared, which is the opposite of vegetation, and so produce NDVI values near -1.0.
Useful articles on vegetation indices using satellite data:
NASA's Earth Observatory
Remote sensing basics from the University of Arizona's RangeView
Getting Started
First, click 'Select Satellite Image'. Select a new satellite image by clicking on the 'Select New Satellite Image' button.
After selecting an image, enter the numerical value of the map scale value visible in the lower left corner of the satellite image. The program will automatically calibrate the analysis tools to measure distances and areas of features within the image.
If you make a mistake, click 'Calibrate Distance/Area Tool'.
Check the results of the calibration by measuring the scale box with the line tool. If incorrect, rerun the calibration.
Analysis Tools
1) Point Tool: Explore the reflected light intensities and a vegetation index (NDVI) at each pixel of the satellite images.
Move the cursor around a satellite image in three ways:
1) Click the mouse on an area of interest,
2) Click and drag the mouse to an area of interest,
3) Use the small up and down arrows along the upper-right edge. These arrows become visible when you select a Point Tool.
2) Line Tool: Measure distance and examine the intensities of reflected light and and NDVI values along the line (example of the Line Tool is the first illustration at the top of the page).
Place the cursor in a satellite image, click and drag the cursor to draw a line.
Adjust the end points of the line with the small up and down arrows along the upper-right edge.
These arrows become visible when you select a Line Tool.
The start and end of the line are color-coded (blue = start, red = stop).
NDVI values for pixels along the line are graphed.
Move the cursor to each satellite image to see individual lines of yearly data on the graph.
An average NDVI of pixels along the line are calculated and color-coded by the year of each image.
3) Area Tool: Measure area and examine the intensities of reflected light and and NDVI values within rectangular regions.
Place the cursor in a satellite image, click and drag the cursor to draw a rectangle.
Adjust the corners of the rectangle with the small up and down arrows along the upper-right edge.
These arrows become visible when you select a Area Tool.
Graphs of the intensities of reflected light and the NDVI values for the pixels within the rectangle are plotted.
The number of pixels within intensity increments of 5% and 0.2 NDVI values are plotted in the histogram.
Display Tools: There are 13 visual enhancements to view the satellite image.
1) RGB = standard color composite of Landsat imagery in which the intensity of the infrared measurements are displayed in the computer display's red, the visible red displayed in the computer's green, and the visible green displayed in the computer's blue.
2) B&W = black and white image of the RGB image in which the average of the intensity values the image's infrared and visible red and green is displayed equally in the computer display's red, green, and blue.
3) NDVI (Normalized Difference Vegetation Index) displays a color-coded scale of vegetation cover. For more information, see the reference articles at the top of the page.
4) Mask–Display pixels that meet intensity and NDVI thresholds you select. More information about this feature follows the Display Tool overview.
Pixels meeting your criteria are displayed as black.
White pixels represent those that did not meet the criteria.
Results may be explored with the Area Tool.
5-7) IR, Red, or Green (Gray). A gray shade image of only one set of satellite measured intensities.
For example, a gray shade image of IR displays the pixel's IR value in all three colors of the computer's display.
Gray shade images allow you to examine the intensities without biasing your sensitivity to red, green, or blue.
8-10) IR v R, IR v G, or R v G: For each pixel, display the difference between two listed satellite measurements.
The associated color of the greater value is displayed:
IR is displayed as a shade of gray, Red as red, and Green as green.
For example, using IR v R, if a pixel has 20% IR, 85% Red, and 65% Green, the difference is 65% to be displayed in the computer's Red.
The brighter the color, the greater the difference between the two measurements, and the darker the color (or closer to black), the smaller the difference between the two measurements.
11-13) Normalized versions of IR v R, IR v G, or R v G: Used to visualize the difference between two sets of satellite measurements, the basic formula is (Intensity A - Intensity B) / (Intensity A + Intensity B).
Minimizes the effects of clouds and shadowing by topography.
The color scheme is similar to those used in 8-10: shades of gray are used to indicate where IR is greater, red where Visible Red is greater, and green where Visible Green is greatest.
More Information on the Mask Display Feature: Examine the relationships between intensities of infrared and visible red and green for each pixel in the image.
You may either select a range of values to highlight (such as highlight the IR intensities between 50 and 75%) or relative intensities between two sets of measurements (such as highlight the IR values that are greater than the green values by 10 units of intensity).
To create a mask using a range of values, click any of the four check boxes next to IR, Red, Green, or NDVI or use the small arrows to change the maximum or minimum values associated with each category.
If you select more than one category, pixels meeting all criteria will be black.
A mask with IR between 50-75% and Red between 25-60%, creates black pixels that have IR between 50 & 75% AND red between 25 & 60%.
Setting criteria for a relative relationship between intensity values requires you to create the relationship with a mathematical equation.
For example, the relationship "Red = Green +/- 20" will create a mask showing pixels with red intensity greater than or equal to Green-20 and less than or equal to Green+20.
The "+/-" option acts as a filter that allows you to set the range of values when you are using the "=" option.
If you select a relationship "IR = IR + 10", a message will pop up requesting you to select non-matching variables.
To apply the relationships you have created, click the button "Mask Satellite Image using Thresholds set Above".
The criteria you have selected will be summarized in a text box below the image of the masked pixels.
Use the Area Tool to measure the number of pixels (and the area this represents) meeting your thresholds in rectangular regions
Click 'Show Original Satellite Image' and a box with the RGB image will appear.
Drag this image to any position on the computer screen
Use the point, line, and area on this image as well as the enhanced satellite image displayed on the larger window.
Images may be saved.
Getting Help: Check the 'Help" menu for information on software features when using the software.
|
