Georeferencing a Satellite Image Using Another Georeferenced Image in ERDAS Imagine

Music credit:
Author: Kevin MacLeod
freemusicarchiv....
Georeferencing a Satellite Image Using Another Georeferenced Image in ERDAS Imagine
To create and evaluate a list of GCPs that will be used to register one image to another in a first order transformation
Define source and reference images and the geometric model
Use GCP editor to delineate GCPs
Use computed RMS error from Check points to evaluate GCPs.
Use computed transformation matrix to resample the unrectified data
Verify accuracy of rectification using Inquire Cursor
Instructions:
Open Erdas Imagine 8.3.1, follow the instructions, and answer the following questions. You may have to use the Rectification Online Manual to help you through this exercise.
1. From the Main Icon Panel select DataPrep/Image Geometric Correction. In the dialog box that appears, select the From Image File radio button and then select tmAtlanta.img. Click OK. This is the image that has not been georeferenced.
2. You know have a choice of geometric models. From the list of Geometric Models, select Polynomial and click OK.
3. The Polynomial Models Properties dialog will now appear along with the Geo Correction Tools at the top of the screen. We will use a first order transformation, so leave the Polynomial Order set at 1. Answer Question 1.
A minimum number of GCPs is necessary to calculate the transformation, depending on the order of the transformation. This number of points is:
(t + 1)(t + 2)
__________
2
where t is the order of the transformation
Select the Projection Tab. Since a projection has not yet been defined, no projection will appear. Select the Set Projection from GCP Tool button. A dialog box will appear that gives you choices for the reference input. In this exercise you will be performing an image to image rectification, so select Image Layer (New Viewer) and click OK. Select PanAtlanta.img as your reference image and click OK. Click OK to the Reference Map Projection dialog. Your window will organize itself with Chip Viewers and the GCP Tool.
4. Click Apply and Close to the Polynomial Model Properties dialog box. You would only need to save the model parameters if you were performing repeated rectification.
5. If you have not changed your GCP Tool preferences, the Automatic Edit Mode icon will be selected. This is useful for point prediction and matching.
6. Move both link boxes so that they cover a common identifiable point feature. This feature should now be seen in both chip viewers. From the GCP Editor tool bar click the Keep Current Tool icon. Then click the Create GCP icon. Now digitize the common point in each of the chip viewers. Once this is done you will need to click on the Select GCP icon to click and drag the linboxes to other identifiable features.
7. In order to make your GCPs stand out visually, you may want to change their display color. From the GCP Tool Color column, click and select a desired color for all or individual GCPs.
8. Repeat step 4 until you have selected 8 GCPs. Remember to distribute the GCPs evenly throughout the image. Editing out unwanted GCPs is possible through the CellArray. First select the point and then using hidden functions in the first column, select Delete Selection.
9. In order to evaluate your transformation matrix, you will now convert some of the GCPs from Control points to Check points. The difference is that the transformation matrix is not based on Check points. Therefore you can use Check points as an independent means of evaluating the rectification by looking at the RMS error.
Select any 4 points in the CellArray. While they are highlighted, select Edit/Set Point Type/Check. The Type column will update.
10. Select the Compute Error For Check Points icon. The Control Point Error will now change to be the Check Point error. In order to see the Error for the Control points again, select the Solve Geometric Model with Control Points icon and the RMS will change. Answer Question 2.
11. If you have a high RMS error, then delete a control point with the highest Contribution value. This can be done by selecting the point so that you turn the row yellow and then using the hidden functions from the Point # column and selecting Delete Selection. You can add and delete control points until you have an acceptable RMS value. A First Order Polynomial requires a minimum of 3 GCPs so remember to maintain this minimum number with an even geographical distribution.

[@tamilarcgisuniversity]

Good video, well explained...

[@gisworld_2017]

Thanks 🥰