System and method for image fusion

• US 7,340,099 B2
• Filed: 01/17/2003
• Issued: 03/04/2008
• Est. Priority Date: 01/17/2003
• Status: Active Grant

• Current Assignee: University of New Brunswick
• Original Assignee: University of New Brunswick
• Inventors: Zhang, Yun
• Primary Examiner(s): Bella; Matthew C.
• Assistant Examiner(s): Rosario; Dennis
• Application Number:
  US10/345,932
• Publication Number:
  US 20040141659A1
• Time in Patent Office: 1,873 Days
• Field of Search: 382/284, 382/294, 382/169, 382/168, 382/191, 348/33
• US Class Current: 382/191
• CPC Class Codes: G06T 2207/10036 Multispectral image; Hyperspectral image
• CPC Class Codes: G06T 2207/10041 Panchromatic image
• CPC Class Codes: G06T 2207/20221 Image fusion; Image merging
• CPC Class Codes: G06T 2207/30181 Earth observation
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction

Abstract

A method of fusing a low resolution multispectral image with a high resolution panchromatic image comprising the step of generating a new synthetic pan image using multispectral bands from the multispectral image.

Claims

1. A method of fusing a low resolution multispectral image having image bands with a high resolution panchromatic image having one image band, comprising the steps of:

selecting image bands from the multispectral image according to the spectral coverage of the panchromatic image;

determining the full grey value range and the common mean wherein the full grey value range is equal to the maximum grey value range of all the original image bands used in the fusion and the common mean is located at the centre of the full grey value range;

performing a histogram standardization on the selected multispectral bands and the panchromatic band; and

generating a new synthetic panchromatic image using the selected multispectral bands.

2. A method according to claim 1 wherein the new synthetic panchromatic image is generated using the equation:

${\mathrm{NewPan}}_{\mathrm{Syn}}=\sum <br><br><br><br>{\mathrm{\phi <br><br>}}_k<br><br>{\mathrm{Multi}}_k$

where NewPan_syn is the new synthetic panchromatic image, Multi_k are the selected multispectral bands whose spectral ranges are overlapped by that of the panchromatic image and φ

_k are coefficients for simulating the NewPan_syn image.

3. A method according to claim 2 wherein the coefficients φ

_k are calculated using the equation;

${\mathrm{Pan}}_{\mathrm{Orig}}=\sum <br><br><br><br>{\mathrm{\phi <br><br>}}_k<br><br>{\mathrm{Multi}}_k$

where Pan_Orig is the original panchromatic image being fused.

4. A method according to claim 1 wherein the step of performing a histogram standardization on the selected image bands comprises the steps of:

(a) calculating the means and standard deviations of the selected bands; and

(b) stretching the histograms of the selected image bands to a common full grey value range with a common mean according to their individual means and standard deviations.

5. A method according to claim 4, wherein the image bands in the stretching step lie within a standard deviation of ±

2σ

or greater.

6. A method according to claim 1, wherein the spectral range of the bands selected from the multispectral image overlaps the spectral range of the panchromatic image.

7. A method according to claim 6, wherein at least one of the bands selected from the multispectral image is different.

8. A method according to claim 6, wherein the images being fused are optical images selected from the group consisting of images from a satellite sensor, an airborne sensor and a digital camera.

9. A method of fusing a low resolution multispectral image with a high resolution panchromatic image comprising the steps of:

(a) selecting from the multispectral image, multispectral bands having spectral ranges which overlap the spectral range of panchromatic image;

(b) calculating means, deviations, and a covariance matrix from all the image bands;

(c) performing a histogram standardization on the selected multispectral bands and the panchromatic image;

(d) standardizing all the bands in the multispectral image and the panchromatic image using the means and deviations;

(e) determining coefficient values using the selected multispectral bands and the panchromatic image;

(f) calculating band weights for simulating a synthetic panchromatic image from the covariance matrix;

(g) generating a synthetic panchromatic image using the coefficient values and the selected multispectral bands;

(h) generating a synthetic panchromatic image using the band weights and the selected standardized bands; and

(i) performing a Product-Synthetic Ratio using the standardized panchromatic band, the standardized multispectral bands and the synthetic panchromatic image;

(j) generating one or more fused image bands using the standardized panchromatic image, one or more of the standardized multispectral bands, and the synthetic panchromatic image.

10. A method according to claim 9 wherein the synthetic panchromatic image has a grey value variation similar to that of the panchromatic image.

11. A method according to claim 9, wherein the selected multispectral bands together cover a spectral range close or equal to that of the panchromatic image within the range from visible to infrared.

12. A method according to claim 9, wherein the images being fused are optical images selected from the group consisting of images from a satellite sensor, an airborne sensor and a digital camera.

13. A method according to claim 12, wherein for IKONOS and QuickBird images, the selected multispectral bands are bands 1, 2, 3 and 4.

14. A method according to claim 12, wherein for Landsat 7 images, the selected multispectral bands are bands 2, 3 and 4.

15. A method according to claim 13, wherein the fused image bands are any three multispectral bands.

16. A method according to claim 13, wherein the fused image bands are one or more bands selected from the group comprising bands 1, 2, 3 and 4.

17. A method according to claim 16, including the step of fusing bands 1, 2, 3 and 4 at the same time.

18. A method according to claim 14, wherein the fused image bands are one or more bands selected from the group comprising bands 1, 2, 3, 4, 5 and 7.

19. A method according to claim 18, including the step of fusing bands 1, 2, 3, 4, 5 and 7 at the same time.

20. A method according to claim 9, wherein the coefficient values are generated using a computational method selected from the group comprising regression and least squares.