Geometric Correction

Geometric Correction

Goals:

This lab was an introduction to the pre processing method used when dealing with images that need to be geometrically corrected. This lab focused on image to map rectification using a reference map to find points on the landscape and image to image registration which compared two images.  

Methods:


Part 1: Image to Map Rectification

The first part focused on geometrically correcting an image in Erdas using the image to map rectification method.  To do this, I first opened both the image to be rectified and the reference image in Erdas Imagine. Using processing tools and the Create GCP tool we created ground control points (GCPs) which is part of the geometric correction process. This first section dealt with a 1st order polynomial transformation of images. For this 1st order polynomial transformation a minimum of three GCPs were needed. We added points until the model solution was current. These GCPs needed to be spread out over the images in order to prevent distortion. We then checked the accuracy by evaluating the Root Mean Square (RMS) error. Our goal was to see this number below 2.0 or more ideally 0.5. Using the Natural Neighbor re sampling method we created an output image from this process


Part 2: Image to Image Registration 


The second part focused on geometrically correcting an image in Erdas using the image to image registration method. This process was similar to the process taken in part 1 but this time we were dealing with a  3rd order polynomial transformation which required a minimum of ten GCPs. After the points were added we evaluated the geometric accuracy through the RMS total error. When getting to the geometric correction processing step we used the Bi linear Interpolation re sampling method which produced our final image for this lab. 

Results:
 Figure 1. Images with GCPs

Figure 2. List of GCPs with corresponding color

Figure 3. Solved model solution with RMS error below 2.0

Sources:

Satellite Images: Earth Resources Observation and Science Center, U.S. Geological Survey
(DRG): Illinois Geo spatial Data Clearing House

Comments

Post a Comment

Popular posts from this blog

Spectral Signature Analysis & Resource Modeling

Image
Goals: The goal of this lab was measure levels of spectral reflectance from various surface and materials. We also evaluated natural resources using remote sensing bands. We focused on Eau Claire and Chippewa counties for this lab. Methods: Part 1: Spectral Signature Analysis We used Erdas and ArcMap to process the data needed for this lab. Part 1  of this lab focused on measuring and plotting the spectral reflectances of 12 surface types. In order to identify different surface areas with more ease I reference the satellite image to google earth. We used the polygon tool in Erdas to draw an area within the desired surface. Then we used the signature editor tool to view the area's spectral signature mean plot. We collected all 12 surfaces and displayed their spectral reflectance signature mean plots together for further analysis (Figure 13). We analyzed vegetation and moisture content mostly. Figure 1. Spectral Signature Mean Plot for airport runways Figure 2. Spectra
Read more