FilterTiePointsByGlobalTransform Task

This task uses the global transform to filter tie points. For orthorectified images, nadir, or near-nadir images, the transformation model between the first and second image fits an RST transform. When the scene is rather flat and the sensor is very far from the scene, the transformation model between the two images fits a first-order polynomial transform. Global transform is the most common filtering method.

The following diagram shows where this task belongs within an image-to-image registration workflow:

References:

Jin, Xiaoying. ENVI automated image registration solutions. NV5 Geospatial Solutions, Inc. whitepaper (2017). ENVI automated image registration solutions can be downloaded from our website.

Jin, Xiaoying, and Robert Schafer. Method and system for automatic registration of images. Exelis, Inc., assignee; now owned by NV5 Global, Inc. U.S. Patent No. 9,245,201 (issued January 26, 2016).

Example

This example uses sample images from the Image Registration tutorial. The files are available from our ENVI Tutorials web page. Click the Image Registration link to download the .zip file to your machine, then unzip the files. Update the file references in the example with the correct locations.

; Start the application

e = ENVI()

; Open input rasters

File1 = 'quickbird_2.4m.dat'

File2 = 'ikonos_4.0m.dat'

Raster1 = e.OpenRaster(File1)

Raster2 = e.OpenRaster(File2)

; Get the auto tie point generation task from the catalog of ENVITasks

Task = ENVITask('GenerateTiePointsByCrossCorrelation')

; Define inputs

Task.INPUT_RASTER1 = Raster1

Task.INPUT_RASTER2 = Raster2

; Run the task

Task.Execute

; Get the output tie points

TiePoints = Task.OUTPUT_TIEPOINTS

; Get the tie point filter task from the catalog of ENVITasks

FilterTask = ENVITask('FilterTiePointsByGlobalTransform')

; Define inputs

FilterTask.INPUT_TIEPOINTS = TiePoints

; Run the task

FilterTask.Execute

; Get the output tie points

TiePoints2 = FilterTask.OUTPUT_TIEPOINTS

; Get the image-to-image registration task from the catalog of ENVITasks

RegistrationTask = ENVITask('ImageToImageRegistration')

; Define inputs

RegistrationTask.INPUT_TIEPOINTS = TiePoints2

RegistrationTask.WARPING = 'Triangulation'

; Run the task

RegistrationTask.Execute

; Get the output raster

WarpedRaster = RegistrationTask.OUTPUT_RASTER

; Get the collection of data objects currently available in the Data Manager

DataColl = e.Data

; Add the output to the Data Manager

DataColl.Add, WarpedRaster

; Display the input rasters

View = e.GetView()

Layer1 = View.CreateLayer(Raster1)

Layer2 = View.CreateLayer(Raster2)

; Display the result

Layer3 = View.CreateLayer(WarpedRaster)

Syntax

Result = ENVITask('FilterTiePointsByGlobalTransform')

Input parameters (Set, Get): ERROR_THRESHOLD, FITTING_TRANSFORM, INPUT_TIEPOINTS, OUTPUT_TIEPOINTS_URI

Output parameters (Get only): OUTPUT_TIEPOINTS

Parameters marked as "Set" are those that you can set to specific values. You can also retrieve their current values any time. Parameters marked as "Get" are those whose values you can retrieve but not set.

Input Parameters

ERROR_THRESHOLD (optional)

Specify the maximum error to allow for each tie point. The tie point with the largest error distance from the predicted location is iteratively removed until no tie points have an error greater than the this value. Multiview images of a 3D scene contain local geometric differences due to variation in local terrain relief and difference in imaging view angle. If image misalignment is mainly due to local geometric differences, increase this value.

The default value is 5.0. A higher value keeps more tie points with larger fitting errors. If you prefer to keep most tie points, enter a high value, such as 100.

FITTING_TRANSFORM (optional)

Specify the transform to use when automatically filtering tie points. The default is First-Order Polynomial.

First-Order Polynomial: A first-order polynomial warp includes an XY interaction term to account for image shear.
RST: Rotation, scaling, and translation; this is the simplest method. It does not allow for shearing in the image warp.