Granted Patent B2

US 9,934,585 · App. 14/872,642 · Granted Apr 3, 2018

Abstract

Claims (48)

1. An apparatus for registering images, the apparatus comprising at least one processor configured to implement:

a first transform model estimator configured to generate a first transform model based on corresponding feature points between a first image and a second image;

a second transform model estimator configured to generate a second transform model based on a first partial image of the first image and a second partial image of the second image, the second partial image being generated based on the first partial image; and

a registrator configured to register the first image and the second image by transforming the first partial image using the first transform model and the second transform model,

wherein the second transform model estimator comprises:

a second partial image generator configured to generate the second partial image from the second image, the second partial image being a segmentation of the second image corresponding to a first transform image obtained by transforming the first partial image using the first transform model; and

an estimator configured to generate the second transform model based on corresponding feature points between the first transform image and the second partial image.

2. The apparatus of claim 1 , wherein the second transform model estimator comprises

a first partial image generator configured to generate the first partial image, the first partial image being a segmentation of the first image which comprises an object detected from the first image.

3. The apparatus of claim 2 , wherein the first partial image and the second partial image comprise a region of the object and a surrounding region larger than the region of the object by a predetermined factor.

4. The apparatus of claim 2 , wherein the first partial image generator is further configured to determine the segmentation based on whether the object included in the segmentation satisfies preset conditions.

5. The apparatus of claim 4 , wherein the preset conditions indicate at least one of object temperature and object motion.

6. The apparatus of claim 1 , wherein the registrator is further configured to transform the first partial image into a first transform image using the first transform model, transform the first transform image into a second transform image using the second transform model, and register the second transform image and the second image.

7. The apparatus of claim 1 , wherein the first image is a thermal image, and the second image is a visible light image.

8. An apparatus for registering images, the apparatus comprising at least one processor configured to implement:

a first estimator configured to generate a first transform model based on corresponding feature points between a first image and a second image;

a partial image generator configured to generate a first partial image by extracting an image comprising an object from a preliminary transform image obtained by transforming the first image using the first transform model, and generate a second partial image by extracting an image corresponding to the first partial image from the second image;

a second estimator configured to generate a second transform model based on corresponding feature points between the first partial image and the second partial image; and

a registrator configured to register the first image and the second image by transforming the first partial image using the second transform model.

9. The apparatus of claim 8 , wherein the first partial image and the second partial image comprise a region of the object and a surrounding region larger than the region of the object by a predetermined factor.

10. The apparatus of claim 8 , the registrator is further configured to transform the first image into the preliminary transform image using the first transform model, transform the first partial image into a final transform image using the second transform model, and register the final transform image and the second image.

11. The apparatus of claim 8 , wherein the first image is a thermal image, and the second image is a visible light image.

12. A method of registering images, the method comprising:

generating, by at least one processor, a first transform model based on corresponding feature points between a first image and a second image;

generating, by the processor, a second transform model based on a first partial image of the first image and a second partial image of the second image, the second partial image being generated based on the first partial image; and

registering, by the processor, the first image and the second image by transforming the first partial image using the first transform model and the second transform model,

wherein the generating of the second transform model comprises:

generating the second partial image from the second image, the second partial image being a segmentation of the second image corresponding to a first transform image obtained by transforming the first partial image using the first transform model; and

generating the second transform model based on corresponding feature points between the first transform image and the second partial image.

13. The method of claim 12 , wherein the generating of the second transform model comprises

generating the first partial image, the first partial image being a segmentation of the first image which comprises an object detected from the first image.

14. The method of claim 13 , wherein the first partial image and the second partial image comprise a region of the object and a surrounding region larger than the region of the object by a predetermined factor.

15. The method of claim 12 , wherein the registering of the first image and the second image comprises:

transforming the first partial image into a first transform image using the first transform model;

transforming the first transform image into a second transform image using the second transform model; and

registering the second transform image and the second image.

16. The method of claim 12 , wherein the first image is a thermal image, and the second image is a visible light image.

17. A method of registering images, the method comprising:

generating, by at least one processor, a first transform model based on corresponding feature points between a first image and a second image;

generating, by the processor, a first partial image by extracting an image comprising an object from a preliminary transform image obtained by transforming the first image using the first transform model, and generate a second partial image by extracting an image corresponding to the first partial image from the second image;

generating, by the processor, a second transform model based on corresponding feature points between the first partial image and the second partial image; and

registering, by the processor, the first image and the second image by transforming the first partial image using the second transform model.

18. The method of claim 17 , wherein the first partial image and the second partial image comprise a region of the object and a surrounding region larger than the region of the object by a predetermined factor.

19. The method of claim 17 , wherein the registering of the first image and the second image comprises:

transforming the first image into the preliminary transform image using the first transform model;

transforming the first partial image into a final transform image using the second transform model; and

registering the final transform image and the second image.

20. The method of claim 17 , wherein the first image is a thermal image, and the second image is a visible light image.

Assignments (8)