Lab 6: Geometric Correction

Introduction & Goal

This lab introduces an image preprocessing method known as geometric correction. This lab generally helps students understand the two different types of this method that are performed on satellite images prior to gathering details or data on biological, physical, sociology, and cultural information from these images.

Methods

In this lab, we used a United States Geological Survey (USGS) 7.5-minute digital raster graphic (DRG) image of the Chicago metropolitan statistical area and surrounding areas to correct a Landsat TM image of the same area. We then utilized a Landsat TM image for eastern Sierra Leone, Africa to rectify a geometrically distorted image of that area.

By opening Erdas Imagine 2020 and displaying the Chicago raster graphic (DRG) into the viewer, we first compared images side by side as well as utilizing the swipe tool to observe the distortion of the two images.

Figure 1) Display of Sierra Leone Africa from the second part of the lab with the Swipe tool active to observe distortion between the DRG and Landsat image.

We then utilized the multispectral tool to activate the raster processing tools for multispectral imagery. To start correcting the image, the Control Points tool is used to collect ground control points (GCPs). GCPs are locations on the earths surface that identify an area/specific location and used to correct the image geometric errors in association with appropriate mathematical models (we use 1st -3rd order polynomials to develop the transformation coefficient that is used to rectify the image).  There are several ways to collect GCPs, but in this lab we utilized Image Layer (new viewer). After completing these processes, maximize the multipoint geometric correction window to start georeferencing. It is important to add a specific number of GCPs when performing a 1st - 3rd polynomial transformation. In this case, you only need 3 pairs of GCPs, however, it is always a good idea to collect more than the required GCPs in geometric correction in order for the output image to have a good fit. 

Once the model has obtained a solution, or the minimum number of GCPs required to run the polynomial transformation, you only need to add additional GCPs to only one of the images, as the program will automatically add the corresponding GCP to the second image on the right pane. To evaluate if the image is lined up correctly, we evaluated the GCPs by looking at the Root Mean Square (RMS) Error. By receiving a total RMS error of 0.5 and below, you can achieve the closes geometric correction output. After clicking compute transformation matrix, you can Display resample Image with Nearest neighbor as the resampling method to save and name your output image. A second portion of this exercise is dedicated towards working with the Sierra Leone DRG and Landsat Image to practice utilizing a 3rd polynomial transformation and resampling method changed to Bilinear interpolation.

Results

Figure 2) Georeferencing the graphic and image of the Chicago region.

Figure 3) Georeferencing the graphic and image of the Sierra Leone region. 

Figure 4) Final product utilizing bi-linear interpolation resampling method instead of nearest neighbor executed in part one of this lab.

The data for this lab was provided by:
- The Geography Department at UWEC
- Satellite images: Earth Resources Observation and Science Center, USGS
- And DRG from Illinois Geospatial Data Clearing House

Comments

Post a Comment

Popular posts from this blog

Lab 5: Spectral Reflectance Signature Analysis & Resource Monitoring

Image
Introduction & Goal The goals of this lab is to gain experience on the interpretation of spectral reflectance signatures of multiple surface features on Earth. The data is captured by satellite images to be used in performing monitoring and basic delineations of Earth's Resources. Including monitoring the health of vegetation and soils utilizing remote sensing band ratio techniques. This lab post is intended for future reference and documentation of the correct image classification skills, and includes collecting spectral reflectance signatures from remotely sensed images, graph them, and perform analysis on them by the end of this lab. Methods By using a Landsat ETM + image that covers the Chippewa Valley area and other regions of WI and MN, we will gather and analyze the spectral reflectance signatures of various Earth surface features. Twelve materials and surfaces from the image had data collected including spectral reflectance signatures in Erdas, although spectral signatu
Read more

Lab 7: LiDAR Remote Sensing

Image
Introduction & Goal The objective of this laboratory exercise is to help students develop skills in  Erdas Imagine 2020 and  ESRI ArcGIS Pro  software that include essential basic knowledge on LiDAR data structure and processing. This lab generally helps students understand processing and retrieval of surface and terrain models and processing and creation of intensity image and other products that derive from the point cloud. Methods & Data When opening Erdas Imagine, we displayed individually LAS files and examined whether there are overlapping points at the boundaries of the tiles, because Lidar data tiles mostly have overlaps at the tile boundary to prevent void areas. We added all the lidar point cloud files provided by the UWEC Geography department (.las). After loading the . las files into the viewer, we were able to zoom into any tile of choice to see individual lidar points. We then utilized ArcGIS Pro for the remainder of this lab because it contains easier workflows c
Read more