IMAGE PROCESSING APPARATUS

• US 20080232715A1
• Filed: 12/03/2007
• Published: 09/25/2008
• Est. Priority Date: 06/08/2005
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Shimizu, Masayoshi, Nakano, Yasuhiko, Watanabe, Yuri, Miyakawa, Ayu
• Granted Patent: US 8,401,333 B2
• Time in Patent Office: Days
• US Class Current: 382/284
• CPC Class Codes: G06T 2207/20016 Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: G06T 7/33 using feature-based methods
• CPC Class Codes: G06T 7/35 using statistical methods
• CPC Class Codes: H04N 1/3876 Recombination of partial images to recreate the original image

Abstract

When different images are combined, the corresponding parts are to be aligned accurately. To thus combine images, the feature point of each pixel is calculated from each original image using an operator (filter). The pixels are divided into blocks that the resolution matches that of the reduced image of each original image, and an expectation value map for registering the expectation value of the feature point of each block is drawn. Feature points are expected from the expectation values which the expectation value map holds, and reduced images are superposed one on the other. The feature points of the original images corresponding to the feature points between which the correspondence is derived from the reduced images are extracted, and the original images are superposed one on the other using the extracted feature points.

Claims

1. An image processing apparatus which combines two images by superposing one on another, comprising:

a reduced image generation means for generating reduced images from original images to be superposed one on another other;

an expectation value map generation means for calculating a feature value using a predetermined operator filter for each pixel of an original image of an image to be superposed, dividing a feature value of each pixel into blocks corresponding to the resolution of the reduced image, and generating an expectation value map in which an expectation value of a feature value is registered;

a feature point extraction means for extracting a feature point from the expectation value registered in the expectation value map; and

a superposition means for superposing reduced images of original images to be superposed one on the other using the feature points, extracting a feature point in an original image corresponding to the feature point from a result of the superposition, and superposing the original images one on the other.

2. The apparatus according to claim 1, wherein

the reduced image is generated for a plurality of resolutions, a reduced image of lowest resolution is superposed, a reduced image of sequentially higher resolution is superposed, and the original image is finally superposed.

3. The apparatus according to claim 1, wherein

the feature point extraction means removes an area in which it is determined that an object in an image is moving, and a feature point is extracted.

4. The apparatus according to claim 1, wherein

the expectation value is calculated only for a portion including an edge of an image.

5. The apparatus according to claim 1, wherein

an expectation value for a portion including no edge of an image is set to 0.

6. The apparatus according to claim 1, wherein

after matching brightness levels of two images to be superposed one on another, a superposing process is performed.

7. The apparatus according to claim 3, wherein

an area obtained by performing an expanding process on an area in which the object moves is removed from an area for use in the process.

8. The image processing apparatus according to claim 1, wherein

the operator filter is made of a Moravec operator.

9. An image processing method of combining two images by superposing one on another, comprising the steps of:

generating reduced images from original images to be superposed one on another other;

calculating a feature value using a predetermined operator filter for each pixel of an original image of an image to be superposed, dividing a feature value of each pixel into blocks corresponding to the resolution of the reduced image, and generating an expectation value map in which an expectation value of a feature value is registered;

extracting a feature point from the expectation value registered in the expectation value map; and

superposing reduced images of original images to be superposed one on the other using the feature points, extracting a feature point in an original image corresponding to the feature point from a result of the superposition, and superposing the original images one on the other.

10. A program used to direct a computer to realize an image processing method of combining two images by superposing one on another, comprising the steps of:

generating reduced images from original images to be superposed one on another other;

calculating a feature value using a predetermined operator filter for each pixel of an original image of an image to be superposed, dividing a feature value of each pixel into blocks corresponding to the resolution of the reduced image, and generating an expectation value map in which an expectation value of a feature value is registered;

extracting a feature point from the expectation value registered in the expectation value map; and

superposing reduced images of original images to be superposed one on the other using the feature points, extracting a feature point in an original image corresponding to the feature point from a result of the superposition, and superposing the original images one on the other.