Image alignment method and apparatus

Patent metadata

Field	Value
Publication number	US-10127679-B2
Application number	US-201715448799-A
Country	US
Kind code	B2
Filing date	Mar 3, 2017
Priority date	Sep 5, 2014
Publication date	Nov 13, 2018
Grant date	Nov 13, 2018

Abstract

An image alignment method and apparatus, where the method and apparatus include obtaining image information of two to-be-aligned images, determining, using a cross-correlation measurement model, first coordinate offset according to the image information of the two images, where the first coordinate offset are used to indicate position deviations of to-be-aligned pixels between the two images in the coordinate system, and aligning the two images according to coordinates of pixels in the first image in the coordinate system and the first coordinate offset.

First claim

What is claimed is:

1. A computer-implemented image alignment method, comprising:
obtaining image information of two to-be-aligned images, image information of a first image comprising coordinates of pixels in the first image in a selected coordinate system, pixel values of the pixels in the first image, and a pixel value gradient of the pixels in the first image, image information of a second image comprising pixel values of pixels in the second image and a pixel value gradient of the pixels in the second image, and the two images being located in the selected coordinate system;
determining, using a cross-correlation measurement model based on a color cross-correlation and a weighted gradient cross-correlation, first coordinate offset according to the image information of the two images, the first coordinate offset indicating position deviations of to-be-aligned pixels between the two images in the selected coordinate system; and
aligning the two images according to the coordinates of the pixels in the first image in the selected coordinate system and the first coordinate offset.

2. The method according to claim 1 , wherein before determining the first coordinate offset, the method further comprises:
obtaining image information of a third image, the image information of the third image comprising pixel values of pixels in the third image and a pixel value gradient of the pixels in the third image, the third image being located in the selected coordinate system, and the first image and the third image being to-be-aligned original images;
determining, using the cross-correlation measurement model, a coordinate transformation matrix according to image information of the to-be-aligned original images, the coordinate transformation matrix indicating a spatial position relationship of to-be-aligned pixels between the to-be-aligned original images in the selected coordinate system;
determining second coordinate offset according to the coordinate transformation matrix, the second coordinate offset being used to indicate position deviations of the to-be-aligned pixels between the to-be-aligned original images in the selected coordinate system; and
obtaining the second image according to the second coordinate offset and the pixel values of the pixels in the third image.

3. The method according to claim 2 , wherein determining the coordinate transformation matrix comprises determining the coordinate transformation matrix by calculating a minimum value of

E
1

⁡

(
H
)

=

∑
p

⁢

⁢

E
2

⁡

(

p
,

w
p

)

,

where E 2 (p,w p )=ρ(1−|Φ I (p,w p )|)+τρ(1−|Φ ∇I (p,w p )|),

Φ
I

⁡

(

p
,

w
p

)

=

(

I

1
,
p

-

I

1
,
p

′

)

T

⁢

(

I

3
,
p

-

I

3
,
p

′

)



I

1
,
p

-

I

1
,
p

′



⁢



I

3
,
p

-

I

3
,
p

′



,

Φ

∇
I

⁡

(

p
,

w
p

)

=

Assignees

• Huawei Tech Co Ltd

Inventors

• Shen Xiaoyong
• Jia Jiaya
• Hung Kwok Wai

Classifications

• G06T5/50

  using two or more images, e.g. averaging or subtraction · CPC title

• G06T7/32 Primary

  using correlation-based methods · CPC title

• G06T3/0075

  Physics · mapped topic

• G06T7/13

  Edge detection · CPC title

• G06T2207/10036

  Multispectral image; Hyperspectral image · CPC title