Normalization method for eliminating false detections in side scan sonar images

• US 5,493,619 A
• Filed: 03/11/1994
• Issued: 02/20/1996
• Est. Priority Date: 03/11/1994
• Status: Expired due to Term

• Current Assignee: Northrop Grumman Systems Corporation (Northrop Grumman Corporation)
• Original Assignee: Northrop Grumman Corporation
• Inventors: Gilmour, John E., Haley, Paul H.
• Primary Examiner(s): Boudreau, Leo H.
• Application Number:
  US08/209,095
• Time in Patent Office: 711 Days
• Field of Search: 382/1, 382/14, 382/15, 382/22, 382/54, 395/21, 348/163, 367/131, 367/88, 367/93
• US Class Current: 382/199
• CPC Class Codes: G01S 15/8902 Side-looking sonar
• CPC Class Codes: G01S 7/539 using analysis of echo signal for target characterisation; Target signature; Target cross-section

Abstract

A method for eliminating false detections at the water column boundary in side scan sonar images is provided. Each image is filtered by computing the median greylevel value of a pixel and its four adjacent pixels to produce a median image. The median image is filtered with a first linear filter to produce an edge image. The edge image is then filtered with a second linear filter to produce an edge-smoothed image. The edge of the image is detected by searching each row of the edge-smoothed image for the maximum greylevel. A transition region is defined between the detected edge and a preselected right boundary and the average greylevels over the rows and columns in this transition region is computed. The sonar image is then normalized by computing the image mean grayness level of each pixel in the image.

Claims

1. A method for eliminating false detections at the water column boundary in side scan sonar images consisting of a matrix of pixels, each pixel having a known greyness level, comprising the steps of:

a. filtering each image by computing the median greylevel value of a pixel and its four adjacent pixels to produce a median image;

b. filtering said median image with a first linear filter to produce an edge image;

c. filtering said edge image with a second linear filter to produce an edge-smoothed image having a plurality of rows and columns;

d. detecting the edge by searching each row of said edge-smoothed image for the maximum greylevel;

e. defining a transition region between said detected edge and a preselected right boundary, said transition region having a plurality of rows and columns, and computing the average greylevels over the rows and columns in said transition region; and

f. normalizing the image by computing the image mean greyness level of each pixel in the image.

2. The method of claim 1 wherein at least three iterations of said median filtering are performed before said median image is filtered by said first linear filter.

3. The method of claim 1 wherein segments are defined in the transition region between said detected edge and said right boundary and the average greylevel of the segments are computed.

4. The method of claim 3 wherein said image normalization is performed over three separate regions of said image.

5. The method of claim 4 wherein one of said three separate regions of said region is a water column region and the water column region of said image is normalized to have a constant greylevel mean.

6. The method of claim 4 wherein the greylevel of said segments between said water column boundary and said right boundary are normalized by a sliding mean that interpolates the greylevel means of said segments.

7. The method of claim 4 wherein the region to the right of said right boundary is normalized by a greylevel mean computed from a moving average.

8. The method of claim 7 wherein said moving average is computed using an odd number of columns.

9. A system for eliminating false detections at the water column boundary in side scan sonar images consisting of a matrix of pixels, each pixel having a known greyness level, comprising:

a. a median filter which computes the median greylevel value of a pixel and its four adjacent pixels and produces a median image;

b. a first linear filter which receives said median image and produces an edge image;

c. a second linear filter which receives said edge image and produces an edge-smoothed image having a plurality of rows and columns;

d. edge detecting means which searches each row of said edge-smoothed image for the maximum greylevel;

e. means for defining a transition region between said detected edge and a preselected right boundary, said transition region having a plurality of rows and columns and computing the average greylevels over the rows and columns in said transition region; and

f. means for normalizing the image by computing the image mean greyness level of each pixel in the image.

10. The system of claim 9 wherein at least three iterations of said median filtering are performed before said median image is filtered by said first linear filter.

11. The system of claim 9 including means for defining segments in the transition region between said detected edge and said right boundary and computing the average greylevel of the segments.

12. The system of claim 11 wherein said image normalization is performed over three separate regions of said image.

13. The system of claim 12 wherein one of said three separate regions of said region is a water column region and the water column region of said image is normalized to have a constant greylevel mean.

14. The system of claim 12 wherein the greylevel of said segments between said water column boundary and said right boundary are normalized by a sliding mean that interpolates the greylevel means of said segments.

15. The system of claim 12 wherein the region to the right of said right boundary is normalized by a greylevel mean computed from a moving average.

16. The system of claim 15 wherein said moving average is computed using an odd number of columns.