Question 3: Image Enhancement

[Part 1]Comparing non-Filtered image and Filtered image

Analysis: 

 - By filtration, I was able to enhance an true colour image with less noise. When I compare those two images above, filtered true colour image has less noise with clear view.

 - The original true colour image has red dots all over the image that TM3 (red band) was containing. Moreover, I am able to see a long thin stripe on the top that TM 2 had. The white band that TM 1,2,3 had can be clearly seen in the Original true colour image.

 - On the other hand, the second true colour image that I create after filter operation has less noises. I operate all different and the best filtering system on each band. I used, FMO for TM1, FME for TM2, and MNFNR for TM3. (Too see the results, see Question 2) This filtered True Colour image does not have the thin stripe on the top of the image. Also,
 By Using Maximum Noise Fraction Base Noise Removal (MNFNR), which is another way of removing bands affected by noise in a dataset, Dots all over the image (made by Band 3) are completely eliminated.  Moreover,  Even though I can still see the white band on the bottom-left of the image, the white band is much lighter than one in the original image.

 Analysis: 

- As those two false colour images above shows, the original false colour image and the filtered false colour image are very different. I these are false colour images that has NIR band as red colour, TM3 (Red) as Green colour, and TM2 (Green) as Blue colour.  

- The original false colour image also has red dots all over the image that TM3 (red band) was containing. Moreover, I am able to see a long thin stripe on the top that TM 2 had. The white band that TM 1,2,3 had can be clearly seen in the Original false colour image.

 -  On the other hand,  the second false colour image that I create after filter operation has less noises. Dots

On the other hand, the second true colour image that I create after filter operation has less noises. In order to get the filtered & combined image, I operate all different and the best filtering system on each band. The same as True Colour image, I used, FMO for TM1, FME for TM2, and MNFNR for TM3. (Too see the results, see Question 2). In the result image, there was no the thin stripe on the top of the image. Moreover, Maximum Noise Fraction Base Noise Removal (MNFNR), which is another way of removing bands affected by noise in a dataset, nicely worked well to remove little red dots in Band 3. Also, unlikely true colour image, the white band  on the bottom-left is barely seen.

 

[Part 2] MAKING ENHANCED IMAGE with FUZZY Band

Step: 

In order to enhance the roads image only,

a) we should make a Fuzzy image that represents the Red band after processing with a 3×3 low pass filter  (I used Median Filter)

b) for the Red band, I used filtered image MNFNR

b)  Run the enhancement formula 

 Red + [ (Red - Fuzzy) * x ]     ;   x = 2, 8 or 16

 *** x represents a scaling factor to emphasize the effect of enhancement. 

c) run the formula with all three scaling factors (x)

Analysis:

- In the histogram, x axis represents the pixel value, and y axis represents the numbers of pixel. Base on observing histograms produced by different scale factors, the larger enhancement scale tends to spread out the pixel distribution. In the image with enhancement scale factor 2, the range of pixel value is between -45 to 461, and most of pixels accumulate at the pixel value 35.
  In the image with enhancement scale factor 8, the range of pixel value is wider than the image with scale factor 2. The range of pixel value is in between -353 and 1139. Most of pixels accumulates at the pixel value 62.771 (Mean Value).
Moreover, for the image with enhancement scale factor 16, the range of pixel value is even wider than the image with scale factor 8. The range of pixel value is in between-785 and 2043. The mean value was 82.5024.
  To sum up, as the scaling factor is higher, the mean value gets higher and the range tends to be wider. In the enhancement scaling factor 2 image, the shape of histogram tends to be smoother than the image with enhancement factor 8 and 2. In addition, there is a significant difference between the scale 2 image and the other two images, which there is small amount of pixels in scale factor 8 and 16 images lie on the region where the pixel value is over 100, and this does not happen in scale factor 2 image.  

 

[Road Extraction image with Scaling Factor 2]

[Road Extraction image with Scaling Factor 8]

[Road Extraction image with Scaling Factor 16]

Conclusion: 
The image with scale factor 2 provides the best performance on enhancing the roads. It has the most clear road shape among three images. Also, base on the visual observation, there is a significant difference between the scale factor 2 image and the other two images.

--------------------------------------------------------------------------

To see Screenshots click here