GISC_381_Lab5_Instructions

docx

School

University of British Columbia *

*We aren’t endorsed by this school

Course

GISC381

Subject

Electrical Engineering

Date

Apr 3, 2024

Type

docx

Pages

Uploaded by SargentIceGerbil19

GISC 381 Lab 5 – Intro to satellite remote sensing Instructions Getting started At the start of each lab, you will need to download the data from the GISC 381 Canvas site. Data for the labs are located in the Home tab in Canvas where you can also find the lab assignment and instructions. Download the entire data folder for the given lab (e.g., Lab 1, Lab 2, etc.) and not the individual files in the folder. In doing so for this lab, select Lab5.zip and Download . The lab data will be downloaded as a zipped folder. You must unzip the folder to work with the data in ArcGIS. When in file explorer, right-click on the folder and select “7-Zip” and “Extract files…”. Beside “Extract to:” navigate to a folder of your choice to save the data (your UBC F: drive or UBC OneDrive account). DO NOT save the data or any of your results locally on the Desktop, the Documents or the C: drive. You will lose them. Introduction Remote sensing is the science of acquiring information about the Earth’s surface from a distance (i.e., without coming into contact with it), often through the use of aircraft or satellite sensors (Coops & Tooke 2017). This lab will introduce you to a few of the key remote sensing concepts used to distinguish land cover types, compute an index of vegetation productivity, the normalized difference vegetation index (NDVI), and perform some basic spatial analysis using reflectance values from Landsat 8. Assignment As you work through this document, the assignment questions will immediately precede the instructions. Please submit a separate typed lab assignment. The assignment includes 8 questions (19 marks) . The questions are included here in the Instructions and in the separate Assignment document just for reference. You do not need ArcMap for Questions 1 through 6. They will require the assigned

reading (Coops & Tooke 2017 pg. 3 - 9) and a bit of searching on the internet, so you may want to save them for last. Questions 7 and 8 require ArcMap so you will need to work through the Instructions to answer these questions. Reading: Coops, N.C., Tooke, T.R. (2017). Introduction to Remote Sensing. In: Gergel, S., Turner, M. (eds) Learning Landscape Ecology. Springer, New York, NY. https://doi.org/10.1007/978-1-4939-6374-4_1 Note: you will need jasper_ndvi for Lab 6. 1. Understanding spectral characteristics . The assigned reading (Coops & Tooke 2017) describes how passive remote sensing systems measure the amount of energy reflected by objects on the Earth’s surface. These reflectance values occur within a range of wavelengths (nm) in the electromagnetic spectrum and define what we are able to see with the human eye. While we can only see a small part of the electromagnetic spectrum, satellite sensors have been developed to record reflectance values throughout the spectrum. Using reflectance values measured across the electromagnetic spectrum, we can differentiate among different land cover types. In Figure 1 below, the y-axis represents the percent reflectance, and the x-axis represents the wavelength (nm). The values range from 400 nm (the low end of the visible part of the spectrum) to 1400 nm (short-wave infrared). Characteristic spectral curves are shown for several land cover types. Each curve can be interpreted by examining the relationship between the values on the x- and y-axis for a particular land cover type. For example, water has approximately 20% reflectance of incoming light around 425 nm and 0% reflectance (so 100% absorption of light) at approximately 900 nm.

Figure 1. Spectral characteristics of several land cover types (Source: Coops and Tooke 2017). Using Figure 1 and your reading, answer the following questions which are based on Coops and Tooke (2017). Note that ArcGIS is not required for the following four questions. Assignment Question 1 . What range of the electromagnetic spectrum has the highest reflectance for the two vegetation types? [1 mark] Assignment Question 2 . The visible region of the electromagnetic spectrum is from 400 – 700 nm. Why is there no difference between the spectral characteristics of the two vegetation types across these wavelengths? [1 mark] Assignment Question 3 . Which land cover types have similar reflectance (overlap) at both 700 nm and 1100 nm? Why would this relationship be problematic if you were creating a land cover classification using these spectral regions? [2 marks] Assignment Question 4 . Why would water with high turbidity have greater reflectance compared to clear water in the visible part of the spectrum? [1 mark]

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

2. Remote sensing imagery and GIS The Landsat 8 satellite is an example of a multispectral sensor, and the data are provided in multiple bands . The two sensors on Landsat 8 are the Operational Land Imager , which collects spectral data from the visible to shortwave infrared, and the Thermal Infrared Sensor , which collects data in the thermal wavelengths. Assignment Question 5 . It is very useful to understand what the band designations are for a satellite as it influences the visualizations and analysis we can do with the data. What are the band designations of Landsat 8 and what part (i.e., wavelengths in nanometres (nm) or micrometers μm) of the electromagnetic spectrum does each represent? [5 marks] In a GIS, multispectral imagery with many bands is called a multi-band raster and can be visualized on a computer using a combination of red, green, and blue (RGB) colours. Since we can only visualize three bands at one time, we need to assign the Landsat 8 bands we want to observe to a colour channel in a GIS. Different combinations of satellite bands in the three RGB colour channels can help us visualize and interpret different landscape features in the imagery. 2.1 True Colour Composite The most common visualization of satellite data is a true colour composite , which uses the red, green, and blue bands of the sensor in the RGB channels of a GIS to visualize satellite imagery as it would appear to the human eye. Start ArcMap and add the raster layer l8_kelowna.tif . This a Landsat 8 Operational Land Imagery (OLI) composite image for the Kelowna region for early summer 2015.

**Note: l8_kelowna.tif (and l8_jasper.tif ) only includes four of the Landsat 8 bands. ArcGIS renames them automatically in sequential order, but these names do not correspond with the Landsat 8 band names. In reality:  ‘Band_1’ in ArcMap = Landsat 8 Band 2  ‘Band_2’ in ArcMap = Landsat 8 Band 3  ‘Band_3’ in ArcMap = Landsat 8 Band 4  ‘Band_4’ in ArcMap = Landsat 8 Band 5 To create a true colour composite in ArcMap:  Right click on l8_kelowna.tif in the table of contents and select Properties .  Now select Symbology . You should see an option for RGB Composite and three colour Channels: Red, Green, and Blue (we are not going to use the Alpha channel) and the Band that is assigned to each colour channel. Now we need to assign the Landsat 8 bands to the correct channel. Remember that each satellite band represents reflectance data in a specific part of the electromagnetic spectrum. By assigning Band_3 to the Red colour channel, information in the red part of the electromagnetic spectrum will have red RGB values in the GIS and so on.

 Set the Red channel to Band_3  Set the Green channel Band_2  Set the Blue channel to Band_1 Your Landsat 8 scene is now a True Colour Composite and appears similar to an orthophoto. However, if you zoom in on downtown Kelowna you can see that our ability to discriminate different features is somewhat limited by the spatial resolution of the data (try turning on a basemap with ‘Imagery’ to compare).

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Assignment Question 6 . a) What is the spatial resolution of Landsat 8 OLI in the visible and near- infrared portion of the electromagnetic spectrum? [1 mark] b) What is the temporal resolution of Landsat 8? [1 mark] A useful way to distinguish different types of land cover or features is to use the raw spectral reflectance values stored in each pixel. Using the Go To XY tool select Decimal Degrees in the dropdown and enter Long: -119.486622 and Lat: 49.895186 and select Zoom To .

Now use the Identify tool and click on a pixel at that location (it doesn’t have to be the exact pixel but try to be close). A window will open with the RGB colour channels with the raw Landsat 8 reflectance values for that pixel. Make note of the reflectance values for the pixel. Once again in the Go To XY tool enter Long : -119.427413 and Lat : 49.851023 and select Zoom To . Using the Identify Tool select the location and record the reflectance values in the RGB channel. 2.2 False Colour Composites As you have seen, satellites like Landsat 8 collect data in various portions of the electromagnetic spectrum that are not visible to the human eye (e.g., the near- infrared, short-wave infrared etc.). We can assign these non-visible bands to colour channels in ArcGIS to create a variety of false colour composites . For example, a popular false colour composite is the colour infrared which assigns the near- infrared band to the red colour channel, the red band to the green colour channel, and the green band the blue colour channel. To create an infrared false colour composite:  Right click on l8_kelowna.tif and select Properties > Symbology .  In the colour channels set Red as Band_4 , Green as Band_3 and Blue as Band_2 .

 Before proceeding, click on the Histograms button. This shows you the distribution of the reflectance values for the band in the colour channel as well as the min, max, mean, and standard deviation of reflectance for the entire image. There are 256 values (0-255) that can be stored for each band. In the reading, radiometric resolution is described as the number of bits that can be stored by a band. Since 2 8 = 256, Landsat 8 has an 8-bit radiometric resolution .  Select Apply and OK .

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Your Landsat 8 scene is now a False Colour Composite . Red is the most prominent colour and is representing reflectance values associated with the near- infrared band.

Assignment Question 7 : a) Focusing on the red colour channel, what general types of land cover does an infrared false colour composite clearly differentiate from one another? [1 mark] b) Focusing on the red colour channel, which general type of land cover has the highest reflectance values and why would this type of land cover reflect more in the infrared compared to other land cover types highlighted by the red colour channel? [2 marks] 3. Vegetation Indices Mapping landscape features and vegetation can also be facilitated using an index that combines different bands to create a new image. A popular index in remote sensing for mapping vegetation is the Normalized Difference Vegetation Index (NDVI) which is calculated based on the difference between the near infrared (NIR) and red regions of the spectrum. NDVI values will range from -1 (no vegetation) to +1 (healthy and dense vegetation). More information on NDVI is available:  https://earthobservatory.nasa.gov/features/MeasuringVegetation/ measuring_vegetation_2.php  https://www.usgs.gov/special-topics/remote-sensing-phenology/science/ ndvi-foundation-remote-sensing-phenology The formula for NDVI is: NDVI = NIR − ¿ NIR + ¿ To calculate NDVI, start a new ArcMap session and add the raster layers l8_jasper.tif , which is an early summer 2019 Landsat 8 image (with blue, green, red, and near-infrared bands) for the Jasper, Alberta region. This time when you create the true and false colour composite (same bands as l8_kelowna.tif ) set the Stretch type to Standard Deviations and n = 0.5.

There is a little bit of cloud and noise in this image (it is very hard to get cloud free imagery in the mountains). Also, the true colour composite can look quite dark since this image has not had the reflectance values calibrated like the Kelowna area image. Using standard deviations as the stretch type increases the brightness of the image.

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Remember to think about what those higher red values (higher value = greater reflectance in the NIR) are telling you in the false composite. To calculate NDVI, we need to add the bands of l8_jasper.tif individually. Go to Add Data and double click on l8_jasper.tif . You should see the individual bands:

Add Band_3 (red) and Band_4 (NIR) to ArcMap. Each band is stored as an individual raster. The NDVI is calculated per-pixel and is just a simple ratio so we can use the Raster Calculator and Map Algebra . Open the Raster Calculator and input the following expression (B4 – B3) / (B4 + B3) which translates to (NIR – Red) / (NIR + Red) using the raster layer names in the Layers and variables window. Save the output as ndvi_jasper . It should look like this: The values should range from approximately -0.59 to 1. You will need ndvi_jasper for Lab 6.

You can see that there are only pockets of vegetation with higher NDVI values and large areas with very low NDVI. 3.1 NDVI and land cover Land cover can be useful to help understand patterns of NDVI, vegetation productivity and health. Add the lcov raster layer to ArcMap. This is a categorical raster representing land cover derived from Landsat imagery for 2019. The classes included in the Jasper region include: Raster Value Land cover class 20 Water 31 Snow & ice 32 Rock & rubble 33 Exposed barren land (also includes urban areas and infrastructure) 50 Shrubs 80 Wetlands 81 Treed wetland 100 Herbs (in this region includes a lot of alpine areas) 210 Coniferous 220 Broadleaf (i.e., deciduous) 230 Mixedwood

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Using the ndvi_japer and lcov layers calculate the mean and the standard deviation of ndvi by land cover type (hint: Zonal Statistics as Table should help). Assignment Question 8 : a) Which land cover classes have the highest and lowest mean NDVI. [1 mark ] b) What general trends do you notice about the mean NDVI values and vegetation cover in the land cover classe? [1 mark] c) Which land cover class has the highest standard deviation in NDVI? Explain why you think this land cover class had the most variable NDVI values. [2 marks] End Instructions

Related Documents

discussion8.pdf

Lab 7 - Tech 63.docx

Lab_9_Tech_63.pdf

Electrical Resistance.docx

Mate Lab Report 2.pdf

1007ICT_Module 4 Workshop Assessment.docx

EE330L - Report 2.pdf

EE330L - Report 1.pdf

NETW310-Module 5-Pietro Mariano.pptx

SOCWK 220R_ Assignment 1.docx

Sexton_Milestone 3.pptx

Build10.pdf