Apparatus and method for detecting horizon in sea image

• US 8,275,170 B2
• Filed: 12/07/2007
• Issued: 09/25/2012
• Est. Priority Date: 12/08/2006
• Status: Expired due to Fees

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: Lee, Jae Kwang, Park, Chang Joon, Lee, In Ho
• Primary Examiner(s): Carter, Aaron W
• Application Number:
  US11/952,811
• Publication Number:
  US 20080137960A1
• Time in Patent Office: 1,754 Days
• Field of Search: 382/103, 382/199, 382/203, 382/282, 701/21, 348113-124, 348143-149
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/30184 Infrastructure
• CPC Class Codes: G06V 10/25 Determination of region of interest [ROI] or a volume of interest [VOI]

Abstract

Provided is an apparatus and method for detecting a horizon which is necessary to detect a camera movement when compositing sea images in a marker-free sea-image camera tracking system. In the method, an ROI is selected near a horizon in a still image of the moving image of the sea, and each maximum point corresponding to the maximum brightness difference is detected from brightness differences between two pixels of each pair having two symmetrical pixels in each column of the ROI. The horizon is detected through a line-fitting using the maximum points. Therefore, the result of horizon detection can be very stable and a horizon can be easily detected in a sea scene having hundreds of frames.

Claims

1. An apparatus, comprising:

a region-of-interest (ROI) detection unit that detects coordinates of an ROI selected in a still image corresponding to a frame of a moving sea image; and

a control unit that

detects a horizon corresponding to a line-fitting in the ROI of the still image determined by the detected coordinates,

calculates parameters a and b as Eigenvalues of the following matrix A,

2. The apparatus of claim 1, wherein the still image is adjusted in contrast and is Gaussian-filtered.

3. The apparatus of claim 2, wherein the still image is adjusted in contrast according to a previously set value.

4. The apparatus of claim 2, wherein the still image is adjusted in contrast under control of the control unit.

5. The apparatus of claim 1, wherein the ROI detection unit detects coordinates of the ROI in a first still image among a plurality of still images corresponding to a plurality of frames of the moving sea image.

6. The apparatus of claim 1, wherein the line-fitting is performed by

for a plurality of pixels in the ROI, each pixel having two pixels disposed in vertical symmetry thereabout, computing respective brightness differences between the two pixels, and

detecting a maximum point corresponding to a maximum brightness difference in a column of the ROI including the plurality of pixels.

7. The apparatus of claim 6, wherein the maximum point is detected for each of a plurality of columns in the ROI.

8. The apparatus of claim 1, wherein the control unit assigns a same file name to the still images corresponding to the moving sea image, and assigns a serial number to each of the still images.

9. The apparatus of claim 1, wherein the coordinates detected by the ROI detection unit are obtained from a left-upper-point and a right-lower-point.

10. A method, comprising:

designating an ROI (region-of-interest) in a still image corresponding to a frame of a moving sea image;

detecting a horizon corresponding to a line-fitting in the ROI of the still image;

calculating parameters a and b as Eigenvalues of the following matrix A,

11. The method of claim 10, further comprising adjusting the still image in contrast and Gaussian-filtering the still image.

12. The method of claim 10, further comprising selecting a first still image from among a plurality of still images each corresponding to a frame of the moving sea image.

13. The method of claim 10, further comprising determining the ROI using coordinates obtained from a left-upper-point and a right-lower-point.

14. The method of claim 10, wherein performing the line-fitting includes

for a plurality of pixels in the ROI, each pixel having two pixels disposed in vertical symmetry thereabout, computing respective brightness differences between the two pixels, and

detecting a maximum point corresponding to a maximum brightness difference in a column of the ROI including the plurality of pixels.

15. The method of claim 14, wherein performing the line-fitting further includes detecting the maximum point for each of a plurality of columns in the ROI.

16. The method of claim 15, wherein the line-fitting further comprises:

obtaining a center among the maximum points; and

shifting the maximum points to the center.