![]() |
Most of this examples represent challenge for registration algorithm. Multi-temporal images are really hard to register and later to evaluate the match quality. That's why the result of several examples is evaluated automatically as uncertain fit.
Example 1:This example demonstrates registration of highly distorted images with high rotation. |
|
|
Example 2: |
|
|
Example 3:Registration of two images with high temporal changes. |
|
|
Example 4:Two rain forest images with two year difference where it's possible to see deforestation progress. |
|
|
Example 5:Registration of two radar images with two year difference. |
|
|
Example 6:Registration of two Landsat images with four year difference and associated rotation. |
|
|
Example 7:Registration of two different sensor images. » Urban SPOT band 3 (08/08/95) + TM band 4 (06/07/94), 256x256. |
|
Example 8:Here we demonstrate the registration result by mosaicking images using equalization, minimum error blend and multi-resolution spline. » Handheld digital camera with lightning changes 640x480 + 640x480. |
|
|
Example 9:This example demonstrates the application of minimum error blend and multi-resolution spline without equalizing images, this method preserves as much as possible original image information and yet eliminates any possible seams. |
|
|
Example 10:Demonstration of minimum-error blending technique to improve visual quality of erroneous distortion modeling. » Handheld digital camera with perspective distortion 640x480 + 640x480. |
|
|