For Washington D.C. I collected 7 histograms, one for each band of the image.
As we can see from these histograms, the data tends to trend in the area of being a low intensity image, meaning the image is darker or contains several dark areas. Although there are two histograms that are trending in the mid-tonal intensity range. In order to improve this and make the image more easy to analyze and distinguish different feature, would be to apply a contrast stretch. This stretches the image data along the x axis allowing for more even tones.
For the image statistics I chose to use scatter plot graphs.
From these scatter plats we can see that the areas where the blue line is more concentrated is in the lower intensity range. Again, if a contrast stretch is done then the image would not be as dark and features could be distinguished slightly better.
Canmore, Alberta Image Histograms and Statistics
The histogram trends are similar to Washington, but in this case there is a lot more concentration at the low intensity area of the graph, meaning there is quite a lot of darkness in the image. In this case again, applying a contrast stretch would allow those low intensity areas to be stretched out over the x-axis and stretches the values along the full range of pixel values. This was the resulted image is less harsh and the values are not stretched too far where the image would not long be able to distinguish features.
I would chose a contrast stretch for both of these images, rather than a histogram equalization that only apply a linear scaling function to the pixel values, but provides a harsh image as a result.
Scatter Plots
Once again these scatter plats are showing, from the concentration of the blue line, that these are low level intensity images and again, a contrast stretch would help to distribute the pixel values along the x-axis.
These images are interesting because they show the differences between an older well established city, and a city that is experiencing growth in a extremely difficult environment, that being in a valley of the rocky mountains. More importantly I wanted to focus on the actual satellite imagery and the fact that TM images don't allow one to look at the fine details of cities, much like I had anticipated but was disappointingly disproved. I learned that if I were to be looking at vegetation or needed to measure mountain ranges then this imagery would be useful, but for looking closely at cities, especially if I were to be tracking rate of development and growth, then I would need finer and more powerful satellite imagery.
References:
[All satellite imagery retrieved from]
Tm-earthsat-orthorectified. (6/14/06, 7/29/02). Retrieved from http://glcfapp.glcf.umd.edu:8080/esdi/ftp?id=12736 & ftp://ftp.glcf.umd.edu/glcf/Landsat/WRS2/p015/r033/p015r33_5t870516.TM-EarthSat-Orthorectified/
[information concerning histogram analysis retrieved from]
Fisher, R., Perkins, S., Walker , A., & Wolfart, E. (2004).Intensity histograms. Retrieved from http://homepages.inf.ed.ac.uk/rbf/HIPR2/histgram.htm
Lillesand, T.M., R.W. Kiefer, and J.W. Chipman. 2008. Remote sensing and image
interpretation (6th edition). Hoboken, NJ: Wiley. 756 p.
[Google Earth Images retrieved from]
Source; "Banff National Park/Canmore." 51°05’00.65”N, 115°20’47.49”W. Google Earth. 2012. January 18, 2012.
No comments:
Post a Comment