EXAMPLES

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.

Registration technology examples

Example 1:

This example demonstrates registration of highly distorted images with high rotation.

» Highly JPEG compressed optical 306x386 + 335x472.

ucsb1.jpg ucsb2.jpg

Example 2:

» High resolution Ikonos optical 1024x1024 + 1064x1172.

ucsb_ikonos1.jpg ucsb_ikonos2.jpg

Example 3:

Registration of two images with high temporal changes.

» Multi-temporal optical 405x350.

trail1.gif trail2.gif

Example 4:

Two rain forest images with two year difference where it's possible to see deforestation progress.

» Forest TM band 5 (06/07/92) + (07/15/94), 512x512.

tm92.gif tm94.gif

Example 5:

Registration of two radar images with two year difference.

» JERS-1 radar (10/10/95) + (08/13/96), 256x256.

sj_1.gif sj_2.gif

Example 6:

Registration of two Landsat images with four year difference and associated rotation.

» Agricultural TM band 5 (09/09/90) + (07/18/94), 512x512.

2_90.gif 2_94.gif

Example 7:

Registration of two different sensor images.

» Urban SPOT band 3 (08/08/95) + TM band 4 (06/07/94), 256x256.

1_spot3df.gif

1_tm4df.gif

Mosaicking technology examples

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.

window_view_1.jpg window_view_2.jpg

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.

» Airborne optical camera 640x480 + 640x480.

sbairport01.jpg sbairport05.jpg

Example 10:

Demonstration of minimum-error blending technique to improve visual quality of erroneous distortion modeling.

» Handheld digital camera with perspective distortion 640x480 + 640x480.

lab_1.jpg lab_2.jpg

VRL > Webreg
Home | Instructions | Examples | Statistics