For this assignment, I chose to take an in depth look at the city of Toronto, Ontario as well as Vancouver-Victoria, British Columbia. Maps were obtained on the Natural Resources Canada website, GeoGratis. Using the 'Download Directory' I looked at the 'Raster Data' and 'Satellite Imagery' available for Landsat 5. I used the 'ftp' version of 'Landsat 5: Data over Major Canadian Cities' for this assignment. Geogratis can be accessed at: GeoGratis.
__________________________________________________________
The following image shows the paths taken to obtain the necessary data for my Toronto maps. The data was uploaded onto my own personal USB key because there is not enough temporary hard-drive space on the school computers.
For the Toronto dataset, I downloaded the 'toronto_BSQ.zip' file which is also available through the following link: GeoGratis - Toronto. The file was unzipped and reloaded onto my USB key in a format which I could work with in PCI Geoinformatica. The same was done for the Vancouver-Victoria dataset. I downloaded the 'vancou-victoria_BSQ.zip' which is also available through the following link: GeoGratis - Vancouver / Victoria. The following image shows the paths taken to obtain the Vancouver - Victoria data.
The image below shows the Toronto, Ontario and Vancouver-Victoria, British Columbia data required for this assignment. The image also shows the files saved onto my USB key and the unzipped versions saved as well.
__________________________________________________________
The following is the metadata provided for the map of Toronto, Ontario:
Product_id: Toronto
Satellite: Landsat5
Sensor: TM
Path: 18
Row: 29:30
Ac_date_yy_mm_dd: 1985-09-20
Spatial_resolution_meters: 25
Datum: NAD 83
Projection: UTM (zone 17)
North (north value of the image): 4957100
South (south value of the image): 4732100
East (east value of the image): 706850
West (west value of the image): 456850
No_bands: 7
Bands: 1:2:3:4:5:6:7
No_pixels: 10000
No_lines: 9000
Toronto;Landsat5;TM;18;29:30;1985-09-20;25;NAD 83;UTM (zone 17);4957100
;4732100;706850;456850;7;1:2:3:4:5:6:7;10000;9000
The following is the metadata provided for the map of Vancouver-Victoria, British Columbia:
Product_id: Victoria Satellite: Landsat5 Sensor: TM Path: 47 Row: 25:26 Ac_date_yy_mm_dd: 1987-09-21 Spatial_resolution_meters: 25 Datum: NAD 83 Projection: UTM (zone 10) North (north value of the image): 5530725 South (south value of the image): 5305725 East (east value of the image): 645750 West (west value of the image): 395750 No_bands: 7 Bands: 1:2:3:4:5:6:7 No_pixels: 10000 No_lines: 9000
Victoria;Landsat5;TM;47;25:26;1987-09-21;25;NAD 83;UTM (zone 10);5530725 ;5305725;645750;395750;7;1:2:3:4:5:6:7;10000;9000
(Obtained at: ftp://ftp.geogratis.gc.ca/landsat/l5_city/vancou-victoria/vancou-victoria_metadata.txt)
__________________________________________________________
Toronto, Ontario
The following is a global locator map of the Toronto area used in my PCI maps. The area highlighted shows the area that my PCI maps were clipped to. A zoomed in view of the highlighted Toronto area is also included below.
(Map obtained at: Google Maps)
Vancouver-Victoria, British Columbia
The following is a global locator map of the Vancouver area used in my PCI maps. The area highlighted shows the area that my PCI maps were clipped to. A zoomed in view of the highlighted Vancouver area is also included below.
(Map obtained at: Google Maps)
__________________________________________________________
Before any work could be done with any of the data that I downloaded, I first needed to clip my images to highlight an area of interest. For both major city images, I clipped the map to highlight the downtown areas. The image below shows the clipping of my Vancouver-Victoria, British Columbia map. The same was done for the Toronto, Ontario map. The 'Clipping' tool can be found in the 'Tools' drop down menu in PCI Geoinformatica - Focus.
The following maps are the True Colour Representations developed for my images of Toronto, Ontario as well as Vancouver-Victoria, British Columbia.
The True Colour Representation of these images depicts the features of the area as they are. All bodies of water are portrayed as blue while all vegetation is portrayed as green. Urban areas and buildings are shown as white and are located in the more urban areas of these cities images. Roofs of houses are shown as brown and we can see an increase in the brown areas as we move away from the urban cores and towards the suburban areas of both cities. The above images both have a coarse texture because of the forested areas and built up urban areas within them. This contrasts the smooth textures representing the bodies of water that these cities are located near.
The above representations also show a contrast in the topography of both cities. The image of Toronto appears to have a smoother surface because of the lack of shadows. Although, if we were to zoom into the image of Toronto, more shadows would appear because of all the urban areas within it. However, we know that Toronto's topography appears smoother than that of Vancouvers because of the region that it is in. In contrast to this, the image of Vancouver shows that it's topography has a little more character. This is characteristic of the region because it is located near the Rocky Mountains of Canada. Shadows highlight the elevated areas which surround the city of Vancouver. Although, like the Toronto image, if we were to zoom into the urban areas, there would also be an increase in shadows shown because of the buildings within the urban core. Therefore, if we were to take a closer look at the urban areas of both cities, the topography of both would be very similar.
The above image shows the tools used to develop the above True Colour representations of Toronto and Vancouver. Band numbers were changed using the RGB Mapping tool. Red was changed to Band 3, Green to Band 2 and Blue to Band 1. The end result of those changes are shown in the above image. For the final version my Toronto map, a 'Linear' enhancement was performed to help bring out and contrast the colours that were being portrayed. Blues were enhanced to show bluer while greens were enhanced to show greener. This aided in the True Colour analysis of the images of Toronto and Vancouver.
Above is the False Colour Representation of the image of Toronto. In this case, vegetation is appearing as bright red because of it's reflectance of the infrared portion of the spectrum. We can also tell that the vegetation in the image is relatively healthy since it is reflecting back as a bright red colour. If the vegetation was not as healthy, it's representation would appear to be more brown or dark red. The urban areas are now appearing as a pinkish colour and once again, water is being represented as blue. The above image tells us that the water is relatively clear because of its dark blue appearence. It is reflecting higher levels of the green portion of the spectrum which is why it appears as a darker blue. If the water had more sediements within it, its reflection would give its representation a more cyan colour.
The above image shows how the RGB Mapper was changed to develop the False Colour Composite of Toronto. Red was changed to Band 4, Green to Band 3 and Blue to Band 2. The changing of the above spectral bands allowed the image to portray reflectances that are not within the visible spectrum. A false colour representation can be used to help analyze the state of vegetation or sedimentation in bodies of water.
Above is the Normalized Difference Vegetation Indices (NDVI) of Vancouver-Victoria, British Columbia. In this grayscale composite, vegetation appears in a bright, white representation while urban areas, houses and bodies of water appear to be a grayish black. In the above image we can clearly see where there are any bodies of water or rivers in the Vancouver area since the black is highlighted by the brightness of the white, vegetated areas.
The above image shows how the NDVI classification was performed. Using the 'Raster Calculator' in PCI Geoinformatica - Focus, the formula for the NDVI classification was inputed into the section in the calculator labeled as 'Expression'.
( NDVI = [NIR - Red] / [NIR + Red] or NDVI = [Band 4 - Band 3] / [Band 4 + Band 3] )
The layer was then set to be resaved as the NDVI layer and the 'Running Man' was pressed in order to run the expression. In other cases of using the NDVI representation, the image could also be classified to assign the various reflectances as different colours. In the above case, we used the grayscale representation.
__________________________________________________________
The following images are the results of generating histograms for each band in the Toronto, Ontario image.
Band 1 - Toronto, Ontario
The above histogram for Band 1- Toronto, Ontario shows us that the average value of the bands pixels are at a value of approximately, 12. '0' in the pixel value represents very dark values while '255' represents very bright values. This tells us that Band 1 of the image is relatively dark because most of the pixels are between the pixel values of '0' and '50'. This also tells us that Band 1 has a very fine spatial resolution since all pixels are located in the same dark pixel value range. A similar patturn is also shown in Bands 2,
Band 2 - Toronto, Ontario
Band 2 (above) and Band 3 (below) show similar patterns to those shown in Band 1. There is not a lot of contrast between pixel values. Through this, we know that the band images are relatively dark with a fine spatial resolution.
Band 3 - Toronto, Ontario
Band 4 - Toronto, Ontario
Band 5 - Toronto, Ontario
Band 6 - Toronto, Ontario
Bands 4, 5 and 6 of the Toronto image show more of a contrast in their pixel values as most fall into the '0' to '140' range. This tells us that there is a higher number of pixels that have mixed pixel value. These bands of the Toronto image give us the details shown in the urban areas of our map. The mixed values tells us that there is an contrast of pixels in the bands where some are showing darker values and others are showing brighter values. This also tells us that Bands 4, 5 and 6 have a coarser spatial resolution than those shown in Band 1, 2 and 3.
Band 7 - Toronto, Ontario
Band 7 goes back to showing a majority of relatively low pixel values. Through this, we know that Band 7 has similar properties than those shown in Band 1, 2 and 3. Therefore, we know that Band 7 is relatively dark with a fine spatial resolution.
The following images are the results of generating histograms for each band in the Vancouver-Victoria, British Columbia image.
Band 1 - Vancouver-Victoria, British Columbia
Band 1's mean pixel value of '66' tells us that this band has a mix of pixels were some have brighter values, while some have lower values. Through this, we know that Band 1 has a mixed fine and coarse spatial resolution.
A majority of the pixel values in Band 2 (above) are relatively low with a mean pixel value of approximately '24'. This tells us that the layer is relatively dark with a fine spatial resolution. Band 3 (below) shows similar values to those shown in Band 2.
Band 3 - Vancouver-Victoria, British Columbia
Band 4 - Vancouver-Victoria, British Columbia
Band 4 (above) and Band 5 (below) show more of a contrast in pixel values to those shown in Bands 1, 2 and 3. Through this, we know that there is a mix of pixels with a range of values contrasting between dark and moderately bright values. This also tells us that these bands have a moderately fine and moderately coarse spatial resolution. These bands help to give the image of Vancouver-Victoria, British Columbia it's details.
Band 5 - Vancouver-Victoria, British Columbia
Band 6 - Vancouver-Victoria, British Columbia
Band 6 (above) and Band 7 (below) also show similar patterns in their spatial resolution. Both have a majority of pixel values that are relatively low so we know that a majority of the pixels in the bands have darker values. Since a majority of the pixels have aprroximately the same value, we also know that these bands both have fine spatial resolutions.
Band 7 - Vancouver-Victoria, British Columbia
__________________________________________________________
Works Cited
Campbell, James B. (2008). Introduction to Remote Sensing. New York, NY: The Guilford Press.
Computer Language Company. (2010). Spatial resolution - computer dictionary defanition. Retrieved from http://computer.yourdictionary.com/spatial-resolution
Dictionary.Com. (2011). Dictionary.Com. Retrieved from http://www.dictionary.reference.com/
Liew, S.C. (2001). Principles of remote sensing: centre for remote imaging, sensing and processing. Retrieved from http://www.crisp.nus.edu.sg/~research/tutorial/opt_int.htm
Natural Resources Canada, Initials. (2007, April 03). Geo gratis - download directory. Retrieved from http://geogratis.cgdi.gc.ca/geogratis/en/download/index.html
Band 2 - Vancouver-Victoria, British Columbia
