Image processing apparatus, image processing method and storage medium

• US 8,908,991 B2
• Filed: 05/29/2012
• Issued: 12/09/2014
• Est. Priority Date: 06/22/2011
• Status: Active Grant

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Bamba, Yasuo
• Primary Examiner(s): Couso, Yon
• Application Number:
  US13/482,824
• Publication Number:
  US 20120328152A1
• Time in Patent Office: 924 Days
• Field of Search: 382/103, 382/254, 382/274, 382/275, 382/299
• US Class Current: 382/275
• CPC Class Codes: G06T 3/12 Panospheric to cylindrical image transformations

Abstract

A high-resolution image obtained by an image sensing operation by an image sensing unit, and a low-resolution image having a resolution lower than the high-resolution image are acquired. An object which satisfies a predetermined condition is detected from the low-resolution image, and an object recognition processing for a region corresponding to the object in the high-resolution image is performed, thus correcting geometric distortions of the region.

Claims

1. An image processing apparatus comprising:

an acquisition unit that acquires a high-resolution image obtained from a broad-range image sensing unit, and a low-resolution image which corresponds to the high-resolution image and has a resolution lower than the high-resolution image, wherein each of the high-resolution image and the low-resolution image contains geometric distortions;

a detection unit that detects an object in the low-resolution image;

a decision unit that decides a region including the detected object within the low-resolution image as an object region and decides a region corresponding to the object region within the high-resolution image as a target region;

a correction unit that corrects the geometric distortion of a partial image being an image of the target region, within the high-resolution image; and

a recognition unit that performs an object recognition processing for the partial image of which the geometric distortion has been corrected.

2. The apparatus according to claim 1, wherein said detection unit detects the object in the low-resolution image sensed at an image sensing timing corresponding to an image sensing timing of the high-resolution image acquired by said acquisition unit.

3. The apparatus according to claim 1, wherein said acquisition unit acquires the low-resolution image by decimating pixels from the high-resolution image.

4. The apparatus according to claim 1, wherein the broad-range image sensing unit has an omni-directional mirror having a circular hyperboloidal, circular paraboloidal, or conical shape, or a fish-eye lens.

11. The apparatus according to claim 1, wherein said detection unit detects a moving object in the low-resolution image.

12. The apparatus according to claim 1, wherein said correction unit converts the partial image having a fan shape, into a partial image having a rectangular shape.

5. An image processing method executed by an image processing apparatus, comprising:

an acquisition step of acquiring, by a processor of the image processing apparatus, a high-resolution image obtained from a broad-range image sensing unit, and a low-resolution image which corresponds to the high-resolution image and has a resolution lower than the high-resolution image, wherein each of the high-resolution image and the low-resolution image contains geometric distortions;

a detection step of detecting, by the processor of the image processing apparatus, an object in the low-resolution image;

a decision step of deciding, by the processor of the image processing apparatus, a region including the detected object within the low-resolution image as an object region and deciding a region corresponding to the object region within the high-resolution image as a target region;

a correction step of correcting, by the processor of the image processing apparatus, the geometric distortion of a partial image being an image of the target region, within the high-resolution image; and

a recognition step of performing, by the processor of the image processing apparatus, an object recognition processing for the partial image of which the geometric distortion has been corrected.

6. The method according to claim 5, wherein in the detection step, the object is detected, by the processor of the image processing apparatus, in the low-resolution image sensed at an image sensing timing corresponding to an image sensing timing of the high-resolution image.

7. The method according to claim 5, wherein in the acquisition step, the low-resolution image is acquired, by processor of the the image processing apparatus, by decimating pixels from the high-resolution image.

8. A non-transitory computer-readable storage medium storing a program for controlling a computer to execute:

an acquisition step of acquiring, by a processor of an image processing apparatus, a high-resolution image obtained from a broad-range image sensing unit, and a low-resolution image which corresponds to the high-resolution image and has a resolution lower than the high-resolution image, wherein each of the high-resolution image and the low-resolution image contains geometric distortions;

a detection step of detecting, by the processor of the image processing apparatus, an object in the low-resolution image;

a decision step of deciding, by the processor of the image processing apparatus, a region including the detected object within the low-resolution image as an object region and decides a region corresponding to the object region within the high-resolution image as a target region;

a correction step of correcting, by the processor of the image processing apparatus, the geometric distortion of a partial image being an image of the target region, within the high-resolution image; and

a recognition step of performing, by the processor of the image processing apparatus, an object recognition processing for the partial image of which the geometric distortion has been corrected.

9. The medium according to claim 8, wherein in the detection step, the object is detected in the low-resolution image sensed at an image sensing timing corresponding to an image sensing timing of the high-resolution image.

10. The medium according to claim 8, wherein in the acquisition step, the low-resolution image is acquired by decimating pixels from the high-resolution image.

13. An image processing apparatus comprising:

an acquisition unit that acquires a high-resolution image obtained from a broad-range image sensing unit, and a low-resolution image which corresponds to the high-resolution image and has a resolution lower than the high-resolution image, wherein each of the high-resolution image and the low-resolution image contains geometric distortions;

a detection unit that detects an object in the low-resolution image;

a decision unit that decides a region including the detected object within the low-resolution image as an object region and decides a region corresponding to the object region within the high-resolution image as a target region; and

a correction unit that extracts an image of a region, from the partial image of which the geometric distortion has been corrected.

14. The apparatus according to claim 13, wherein said correction unit converts the partial image, having a fan shape, into a partial image having a rectangular shape.

15. An image processing method executed by an image processing apparatus comprising:

an acquisition step of acquiring, by a processor of the image processing apparatus, a high-resolution image obtained from a broad-range image sensing unit, and a low-resolution image which corresponds to the high-resolution image and has a resolution lower than the high-resolution image, wherein each of the high-resolution image and the low-resolution image contains geometric distortions;

a detection step of detecting, by the processor of the image processing apparatus, an object in the low-resolution image;

a decision step of deciding, by the processor of the image processing apparatus, a region including the detected object within the low-resolution image as an object region and deciding a region corresponding to the object region within the high-resolution image as a target region; and

a correction step of extracting, by the processor of the image processing apparatus, a region from the high-resolution image of which the geometric distortion has been corrected.

16. The method according to claim 15, wherein in the correction step, the partial image, having a fan shape, is converted, by the processor of the image processing apparatus, into a partial image having a rectangular shape.

17. A non-transitory computer-readable storage medium storing a program for controlling a computer to execute:

an acquisition step of acquiring, by a processor of an image processing apparatus, a high-resolution image obtained from a broad-range image sensing unit, and a low-resolution image which corresponds to the high-resolution image and has a resolution lower than the high-resolution image, wherein each of the high-resolution image and the low-resolution image contains geometric distortions;

a detection step of detecting, by the processor of the image processing apparatus, an object in the low-resolution image;

a decision step of deciding, by the processor of the image processing apparatus, a region including the detected object within the low-resolution image as an object region and decides a region corresponding to the object region within the high-resolution image as a target region; and

a correction step of extracting, by the processor of the image processing apparatus, an image region, from the image of which the geometric distortion has been corrected.

18. The medium according to claim 17, wherein in the correction step, the partial image, having a fan shape, is converted into a partial image having a rectangular shape.

19. An image processing apparatus comprising:

an acquisition unit that acquires a high-resolution image obtained from a broad-range image sensing unit, and a low-resolution image which corresponds to the high-resolution image and has a resolution lower than the high-resolution image, wherein each of the high-resolution image and the low-resolution image contains geometric distortions;

a detection unit that detects an object in the low-resolution image;

a decision unit that decides a region including the detected object within the low-resolution image as an object region and decides a region corresponding to the object region within the high-resolution image as a target region; and

a recognition unit that extracts an image region, from the image of which the geometric distortion has been corrected and performs an object recognition processing for the extracted image of the region.

20. The apparatus according to claim 19, wherein said detection unit detects a moving object in the low-resolution image.

21. The apparatus according to claim 19, wherein the broad-range image sensing unit has an omni-directional mirror having a circular hyperboloidal, circular paraboloidal, or conical shape, or a fish-eye lens.

22. An image processing method executed by an image processing apparatus comprising:

an acquisition step of acquiring, by a processor of the image processing apparatus, a high-resolution image obtained from a broad-range image sensing unit, and a low-resolution image which corresponds to the high-resolution image and has a resolution lower than the high-resolution image, wherein each of the high-resolution image and the low-resolution image contains geometric distortions;

a detection step of detecting, by the processor of the image processing apparatus, an object in the low-resolution image;

a decision step of deciding, by the processor of the image processing apparatus, a region including the detected object within the low-resolution image as an object region and deciding a region corresponding to the object region within the high-resolution image as a target region; and

a recognition step of extracting, by the processor of the image processing apparatus, an image of a region, from the image of which the geometric distortion has been corrected.

23. The method according to claim 22, wherein in the detection step, a moving object is detected, by the processor of the image processing apparatus, in the low-resolution image.

24. A non-transitory computer-readable storage medium storing a program for controlling a computer to execute:

an acquisition step of acquiring, by a processor of an image processing apparatus, a high-resolution image obtained from a broad-range image sensing unit, and a low-resolution image which corresponds to the high-resolution image and has a resolution lower than the high-resolution image, wherein each of the high-resolution image and the low-resolution image contains geometric distortions;

a detection step of detecting, by the processor of the image processing apparatus, an object in the low-resolution image;

a decision step of deciding, by the processor of the image processing apparatus, a region including the detected object within the low-resolution image as an object region and decides a region corresponding to the object region within the high-resolution image as a target region; and

a recognition step of extracting, by the processor of the image processing apparatus, an image region, from the image of which the geometric distortion has been corrected.

25. The medium according to claim 24, wherein in the detection step, a moving object is detected in the low-resolution image.