|Get on the Grid: Working with multispectral image files
||The techniques and capabilities described in this article require MapInfo Pro Advanced. This is available as a free trial download. If you already have v15.2 installed you can activate a free trial of MapInfo Pro Advanced. Go to the Backstage areas (the Pro tab) and click on Licensing.
And now, on to working with multispectral image files.
MapInfo Pro Advanced provides the ability to take advantage of the different bands of data in a multispectral image file.
For the purposes of this article, I will use a Landsat 8 imagery data set. There is a separate article that provides information on how to download Landsat 8 data along with info on the bands and how they are best used.
If you have multispectral image files from other sources then much of what is explained here may still apply.
Unpacking the Landsat 8 data
When the Landsat 8 data file is unpacked, each image band is in a separate file. It is possible to use these individually but MapInfo Pro Advanced provides a convenient way to combine them into a single multi-resolution raster (MRR) file. This makes the data easier to work with and to share.
An example of the files from a Landsat 8 downloaded tile. Each band is a separate image file.
Note that one condition of using the Combine capability is that the raster files you are combining must have the same resolution. For this example we will combine the first 7 bands. These bands are all 30 meter resolution.
Once again, more info on Landsat 8 data, including how to download it, is in this article.
The Field and Band structure in an MRR file.
You can think of the Multi-resolution raster (MRR) file format as a container. It supports having multiple rasters in a single file. The individual rasters are organised into a Field and Band structure. A field can contain one or more bands. Each band is a raster dataset and can be a continuous, classified or image raster file. These can be mixed.
The Combine capability can be used to build MRR files from different rasters.
Here are examples on how an MRR can be organised. Note that the field name and band names can be chosen. These are just examples.
Single field with multiple bands
Multiple fields with individual bands in each field
Multiple fields with one or more bands in each field
Using the Combine command
The Combine command is on the Raster tab, in the Operations group.
Handy tip: You do not have to open the images in MapInfo Pro first. You can go direct to the file system to open the images.
The screenshot below shows the steps needed to Combine bands from the Landsat 8 image into a single MRR file.
In the example above, the 7 Landsat bands are placed into one field called VictoriaB1toB7.
Once you click Process the combine operation will happen in the background. You can continue to use MapInfo Pro to do other things will the data is processed.
Upon completion, the image will display. Here is the Victoria Landsat image.
Landsat 8 image of Victoria in Australia.
Use the Grid Calculator to get rid of the blue area.
The blue color represents areas of zero values. To make the data easier to work with the next step is to use the Grid Calculator to query all the values greater than zero. This will eliminate the blue area and make some of the other calculations and visualisation more meaningful.
This article provides a very basic treatment of the grid calculator. Another article covers the grid calculator in more detail.
The Grid Calculator is in the Operations gallery.
Using the grid calculator to query the grid for all the values greater than zero.
The file name for the result must be different from the original.
The result looks like this.
Landsat 8 image of Victoria with the zero values filtered out.
Viewing the image bands
There are two modes you can view the image in. These are the RGB and Pseudo color method. This article will focus on the RGB method as it offers more power and flexibility.
The RGB method allows the choice of rendering the different bands through the red, green and blue (RGB) channels. This is a big topic for which I'll provide some useful references below. The short version is that different surfaces on the ground reflect waves from the electromagnetic spectrum in different ways. Assigning the different bands to the RGB channels results in rendering the image in ways to bring out specific characteristics.
Here is an example using the following band assignments:
Red = Band 5 (near infrared)
Green = Band 4 (red)
Blue = Band 3 (green)
Plants reflect near infrared and green light and they reflect more near infrared (NIR) rather than green. As such, when this combination is used, areas appearing red indicate vegetation. (Remember the NIR band is assigned to the red channel). The denser the plant growth, the deeper the red color.
Below is an image centered near Karumba in Queensland, Northern Australia. The bands are rendered as described above. The red areas indicate vegetation.
Landsat 8 image centered near Karumba in Queensland Australia.
Using RGB Color mode
As can be seen above, to help facilitate the use of Landsat 7 and 8 data, MapInfo Pro Advanced includes templates which assign the satellite bands to the RGB channels in commonly used ways.
The performance of MapInfo Pro Advanced is such that experimenting with different band assignments is very quick and easy.
What about Landsat 7 and earlier data?
Landsat 7 is still in operation and current imagery is being produced. It has an orbit that is juxtaposed to Landsat 8 such that between the two satellites, repeat coverage of the most of the earth is achieved every 8 days.
In addition, archives from earlier Landsat satellite missions is available as well.
Data can be downloaded from the USGS Earth Explorer web site: http://earthexplorer.usgs.gov/
An article on how to download Landsat 8 data is here.
Much more to come!
We will further explore using Landsat 8 data (as well as other remotely sensed data) in future articles. To give you a hint, we'll show how to calculate a Normalised Difference Vegetation Index.
The interpretation of image data is a very big topic. The following sources were used to help create this article and you might find them helpful too.
From the NASA Earth Observatory Web site:
Why is that Forest Red and that Cloud Blue? How to Interpret a False-Color Satellite Image
How to Interpret a Satellite Image: Five Tips and Strategies
From the United States Geological Service (USGS)
USGS Landsat Datasheet: http://pubs.usgs.gov/fs/2015/3081/fs20153081.pdf
Landsat 8: https://en.wikipedia.org/wiki/Landsat_8
False color: https://en.wikipedia.org/wiki/False_color
Do you have the latest version of MapInfo Pro?
The techniques shown in this article require MapInfo Pro Advanced. If you do not have this, a free trial is available.
Article by Tom Probert and Owen Parfrey
||Tom Probert: When not writing articles for "The MapInfo Pro" journal, Tom enjoys talking to MapInfo Pro users at conferences and events. When not working he likes to see movies with car chases, explosions and kung-fu fighting.
||Owen Parfrey: When not writing articles for "The MapInfo Pro" journal, Owen leads the team building MapInfo Pro Advanced. On his off time he can frequently be found outdoors with his family and occasionally on a farm.