H10+V4+Spatial+Enchancements

Below are images of the study area that show the use of Spatial Filters. In this exercise, both a smoothing filter and edge detector were applied to the images to help better understand what the image is showing. The image used is a 2000 x 2000 pixel subset of the study area from July 7th 2000 that is located in the lower left corner of the main area. This area was picked due to the different features that can be found here. While several areas within the image can be studied to show how the filters work, the best is that of the snowed covered mountains located on the right side of the image.

For the smoothing filter I choose to apply a High Pass Filter to help sharpen the image. The High Pass filter serves two purposes, by removing the low frequency components it not only sharpens but also enhances the edges between different regions. In this exercise, I applied two different kernels, 3x3 & 7x7, with different center values. The default setting is 3x3 with a center value of 8. This setting in my opinion did not produce the desired affect I was looking for so I changed the center value to 12. As a result I obtained a sharper image that with the default value. Next I increased my kernel size to 7x7 which had a default center value of 48. I choose to increase my center value to 72, which is the same ratio as the increase for the 3x3 kernel size. By using the same ratio of center values for both kernel sizes, the comparison of the two images I feel is more accurate. When comparing the two kernel sizes one notices the areas of snow and ice on the mountains better than without the filter. With the 3x3 kernel the snow covered areas appear as different shades which could indicate differences in snow depth. While the 7x7 kernel enhances the snow areas even more identifying the areas that appear to be ice as opposed to those that are just snow.

For the detection I applied a Sobel Filter to help identify the edges of the different regions in the image. Since this filter has a preset kernel of 3x3, no changes could be made. As can be seen in the image, once applied once can identify the the smaller areas of different regions better. An example of this is the river channel in the center part of the image. It is easier to see the size and shape of the small streams to the right of the river.