Remote Sensing Lab 4

Ethan Nauman

10/30/14

The goal of this lab was to introduce us to skills in image preprocessing. A few of the skills we learned while completing this lab were: learning how to pick an area of interest from a larger image, demonstrate how enhancing an image can help the purpose of being able to visually interpret an image, learn radiometric enhancement techniques of optical images, learn how to link satellite images with google earth, and to be introduced to different methods of resampling. 

Part 1: Image Subsetting

For the first part of the lab we learned how to select and area of interest through an inquire box on a satellite image. This method is quite simple but it does pose a problem. That problem is that usually an area of interest isn't in she shape of a square or rectangle, so this technique has limitations. Upon opening the Eau Claire image in Erdas, we selected the raster tool set. These tools allow us to open the inquire box. The inquire box came onto our Eau Claire image and we repositioned it so that it was over the Chippewa and Eau Claire area. By clicking on the outer edge of this box you can either make it smaller or bigger. After I put the inquire box over these two areas, I needed to capture this area. This is done by using the subset and chip tool under the raster tools. Upon completing this step and saving the image, I uploaded the saved image into another viewer. The image that appeared was the area that was in the subset box, that image looked like this. 

Part 1 Section 2: Subsetting with the use of area of interest

This process is used when your area of interest isn't in the shape of a rectangle or box. This technique is quite useful. 

For this process I began with the original Eau Claire image. The next step to this process was to upload the Eau Claire and Chippewa counties shape file on top of this image. Once I completed this I then had to select both of the counties, this was done by holding down the shift key and selecting just the two counties. After selecting the two counties I then had to paste from selected image. This left a dotted line around my area of interest which in turn would allow for me to save the image. Upon completion of this step my final image appeared as this: 

Part 2 Image Fusion

In this portion of the lab, I was able to fuse a finer resolution image with a coarse resolution image allowing for a more clear picture which allowed for better utilization. I began by opening the original Eau Claire image in one viewer and opened a panchromatic image in a second viewer. The panchromatic image was set at 15 meters while the original image was set at 30 meters, I would be using the panchromatic image to 'pan-sharpen' the original image. The pan sharpen tool is located under the raster tool set. Once I clicked on the pan sharpen tool I then went down to the resolution merge tool. Upon clicking on the resolution merge tool, a resolution merge box then appeared. This had many parameters that I needed to go through before completion. The input file was the panchromatic image of the Eau Claire area, the multispectral input file was the original Eau Claire image. I had to create a folder to save the pan sharpened image in and put that folder as my output file. Under the method portion of the resolution merge box, I had to check the multiplicative box. Under the resampling techniques box I used the nearest neighbor tool. After I completed all these steps, the tool was ready to run. The final image that I received after completing this was a sharpened image, that image looked like this: 

Part 3: Simple radiometric enhancement techniques

In this section I learned some simple radiometric enhancement techniques that would enhance image spectral and radiometric quality. The first section of this part dealt with haze reduction. The image I used for this section was in our radiometric folder. I used the raster tool set and used the haze and reduction tool. This in turn brought up the haze reduction window, for the input file I used the Eau Claire image. I also had to create another folder which I named haze reduction. For my output file on the haze reduction window I used this file. Upon filling out these two simple steps I ran the program and was quite amazed at what happened. The original image had clouds over the southeast portion of the image and after running the tool my final image appeared as this: 

There is no more clouds over any portion of the image. 

Part 4: Linking image viewer to google earth

I once again started with the original Eau Claire image and fit it to frame. I then clicked on the google earth icon on the upper left portion of Erdas, and then connect to google earth. I then moved the google earth portion over to the other desktop for easier viewing. I want to show the Eau Claire image on the google earth portion so I had to connect the two. To do this I clicked on the match GE to view on the Erdas interface. Next I wanted to sync the two images. I clicked on the sync GE to view on the Erdas interface. Upon completing this portion I then could zoom in on the original Eau Claire image and it would direct google earth to the same area as that on my Eau Claire image. I found this to be very useful.

Part 5: Resampling

Resampling is the process of changing the size of the pixels on the image which lead to a much clearer image. I could also go the other way with resampling, which would increase the size of the pixels, depending on what I am trying to portray in the image. The image for this section was located under our resampling folder. I brought our Eau Claire image into Erdas and fit it to frame. I then clicked on the raster toolset and used the spatial tool. This brought a drop down menu which I used the resample pixel size tool. I first resampled for nearest neighbor, then I resampled for bilinear interpolation. They used the same process I only had to change from nearest neighbor to bilinear interpolation. The images of both of these were very similar and did make the image much clearer. Upon clicking on the resample pixel size tool, this then brought up the resample window. The original Eau Claire image was the input file. I had to create another folder on my desktop named resample outlet which would allow for me to save the image when done. Under the resample method portion of the window I clicked on bilinear interpolation. I changed the output cell size from 30x30 micrometers to 20x20 micrometers. I then accepted all other parameters and ran the tool. My final image after resampling the pixel sizes looked like this:

Upon completion of this lab I reflected on the techniques I learned and realized that these in turn can prove quite useful. My favorite technique I learned was the haze reduction tool. If you have problems with your satellite image with clouds or poor quality and run this tool, it can prove quite effective and solve your problems.