Below you will find 3 images representing three different scaling factors. Using the raster calculator, we were able to compute the scaling factors x=2,8,16 using the equation below. Each factor represents different levels of edge detection detail. Running this calculation on the image helps us smooth out any noise while enhancing edges. It needs to be kept in mind, that when we run smoothing filters and calculations, we lose some information from the original image, hence the reduction in noise. When looking at the different scaling factors, we can notice some differences.
The actual cell values of the images in each scaling factor vary. The higher the scaling factor, the more differentiation you'll get in regards to cell values. Looking at the figures below, we see three different cell value tables. Each table represents a different scaling factor.
Attempting to pick similar cells amongst all three scaling factors, we can see a difference in the cell values. These images have DN ranging 0-255. In the selected cells, x=2 has a cell value of 115, x=8 has a cell value of 146, and x=16 has a cell value of 234. The relationship between scaling factors and cell values is as follows: as the scaling factor increases, so does the cell value. This makes searching for features, such as roads, more effective.
Scaling factor x=2: In the above image we can clearly define roads and see a variation in land cover. The image is smooth, and there is very little noise.
Scaling factor x=8: In the above image we can see that the image looks sharper, and roads become more visually dominant while landcover becomes more blended.
Scaling factor x=16: In the image above, we can see a significant different in the sharpness of lines and edges. There is a significant more amount of noise as landcover pixels become more blended. This scaling factor was used to extract the roads because roadways appeared at a brighter cell value than all other pixels in the image.
Below you can see a detailed step-by-step description of the process I took to apply the scaling factors:
1. Load and extract the newly filtered band 3 (TM3)
2. Next, you need to select a filter for the filtered band 3. Right click the filtered band 3 layer and select filter.
3. Once you have selected the filter option, you need to ensure the following in the window that will pop up. Leave all fields as default except:
a) Mask: Entire Raster
b) Filter size: x = 3, y = 3 (3x3 window)
c) Use low pass filtering for average
4. You will then apply the filter, and name the new file Fuzzy. You will need to create this image in order to run the calculation provided above. Once you have saved this file, you can exit the filtering window.
5. Under the tools menu in the tool bar, you will need to select raster calculator. This is where you can input the calculations provided above. Run the calculation for each scaling factor (x = 2, x = 8, and x = 16). Your end goal will be three images that have been enhanced by three different scaling factors.
6. To view the new enhanced images in other band combinations other than grayscale, you need to transfer the layers into the original A2Data file.
7. Once the layers have been transferred, you can remove the files from the side, and open up the A2Data file that the layers were transferred into.
8. You can now open the RGB Mapper tool and select the appropriate bands you would like to see. Applying a linear enhancement once the bands are selected will help correct the images accordingly. To see the scaled images in true or false colour, click HERE.
1. Load and extract the newly filtered band 3 (TM3)
2. Next, you need to select a filter for the filtered band 3. Right click the filtered band 3 layer and select filter.
3. Once you have selected the filter option, you need to ensure the following in the window that will pop up. Leave all fields as default except:
a) Mask: Entire Raster
b) Filter size: x = 3, y = 3 (3x3 window)
c) Use low pass filtering for average
4. You will then apply the filter, and name the new file Fuzzy. You will need to create this image in order to run the calculation provided above. Once you have saved this file, you can exit the filtering window.
5. Under the tools menu in the tool bar, you will need to select raster calculator. This is where you can input the calculations provided above. Run the calculation for each scaling factor (x = 2, x = 8, and x = 16). Your end goal will be three images that have been enhanced by three different scaling factors.
6. To view the new enhanced images in other band combinations other than grayscale, you need to transfer the layers into the original A2Data file.
7. Once the layers have been transferred, you can remove the files from the side, and open up the A2Data file that the layers were transferred into.
8. You can now open the RGB Mapper tool and select the appropriate bands you would like to see. Applying a linear enhancement once the bands are selected will help correct the images accordingly. To see the scaled images in true or false colour, click HERE.