Systems and methods for super registration and preserving spectral purity of images
Provided are systems and methods for registration and carrying out fine image adjustments of aerial or satellite images to obtain super registered images. Also provided are systems and methods for registration where resampling is minimized or avoided to reduce, minimize, or prevent spectral degradation.
1. A method for registration of an ortho corrected image to a reference ortho corrected image, the method comprising the steps of:
a) obtaining the ortho corrected image and the reference ortho corrected image;
b) matching the ortho corrected image to the reference ortho corrected image to generate a first set of match points;
c) analyzing magnitudes of (x,y) shifts of one or more local windows of data and filtering from the first set of match points one or more blunder match points being outside of a threshold level of misalignment in orientation and/or length with respect to their corresponding points on the reference image to generate a second set of match points containing one or more locations with missing match points;
d) populating the one or more locations by interpolation from surrounding valid match points to generate a third set of filtered match points, the valid match points being at or within the threshold level of misalignment; and
e) using the third set of match points to resample the ortho corrected image to the reference ortho corrected image to generate a super registered image.
2. The method of claim 1 , further comprising the step of applying Gaussian filtering to the second set of match points between steps c) and d).
3. The method of claim 1 , wherein the matching algorithm matches the ortho corrected image to a grid of every n'th column and n'th row of the reference ortho corrected image to generate the first set of match points, where n is greater than 1, and the third set of match points is interpolated to have the same dimensions as the reference ortho corrected image.
4. The method of claim 1 , wherein Fast Fourier Transform phase matching is used in step b).
5. The method of claim 1 , wherein analyzing the magnitudes of the (x,y,) shifts in step c) comprises continuously recomputing statistical averages and standard deviations of length and orientation on the one or more local windows of data, and the threshold level is a statistical threshold.
6. The method of claim 1 carried out N times to obtain a stack of N super registered images, using the same reference image, where N is an integer greater than 1.
7. A computing device comprising one or more processors and at least one non-transitory computer readable storage medium, the at least one non-transitory computer readable storage medium containing non-transitory computer-readable instructions for execution by the one or more processors to cause the one or more processors to perform a set of operations, the operations comprising:
a) obtaining the ortho corrected image and the reference ortho corrected image;
b) matching the ortho corrected image to the reference ortho corrected image to generate a first set of match points;
c) analyzing magnitudes of (x,y) shifts of one or more local windows of data and filtering from the first set of match points one or more blunder match points being outside of a threshold level of misalignment in orientation and/or length with respect to their corresponding points on the reference image to generate a second set of match points containing one or more locations with missing match points;
d) populating the one or more locations by interpolation from surrounding valid match points to generate a third set of filtered match points, the valid match points being at or within the threshold level of misalignment; and
e) using the third set of match points to resample the ortho corrected image to the reference ortho corrected image to generate a super registered image.
8. The computing device of claim 7 , wherein the operations further comprise the step of applying Gaussian filtering to the second set of match points between steps c) and d).
9. The computing device of claim 7 , wherein the matching algorithm matches the ortho corrected image to a grid of every n'th column and n'th row of the reference ortho corrected image to generate the first set of match points, where n is greater than 1, and the third set of match points is interpolated to have the same dimensions as the reference ortho corrected image.
10. The computing device of claim 7 , wherein Fast Fourier Transform phase matching is used in step b).