Method and apparatus for locating and identifying an object of interest in a complex image

• US 5,848,190 A
• Filed: 05/14/1996
• Issued: 12/08/1998
• Est. Priority Date: 05/05/1994
• Status: Expired due to Term

• Current Assignee: Northrop Grumman Systems Corporation (Northrop Grumman Corporation)
• Original Assignee: Northrop Grumman Corporation
• Inventors: Mandeli, Steven, Kleehammer, Robert J.
• Primary Examiner(s): CHANG, JON CARLTON
• Application Number:
  US08/645,937
• Time in Patent Office: 938 Days
• Field of Search: 382/209, 382/217, 382/218, 382/260, 382/262, 382/278, 382/291, 382/290, 364/728.03
• US Class Current: 382/218
• CPC Class Codes: G06T 7/74 involving reference images or patches
• CPC Class Codes: G06V 10/507 Summing image-intensity values; Histogram projection analysis

Abstract

A method of location and identification of an object of interest in a complex image frame involves defining the object of interest as a reference template. The ranks of grey shades in the complex image frame and reference template are computed. The ranks of grey shades in the reference template are compared with the ranks of grey shades of the complex image, thereby determining closeness of fit. A threshold for detection is set by utilizing selected decision data.

Claims

1. A method of locating and identifying an object of interest in a complex image frame, comprising the steps of:

a. selecting a section of the complex image frame which defines a number of data points in a data set for the selected section, wherein each data point defines a data level of grey shade;

b. computing ranks of grey shade for the data points in the data set for the selected section by ranking each sample data point in the data set for the selected section with respect to the number of other data points in the entirety of the data set for the selected section, the grey shade data levels of which are exceeded by the grey shade data level of the sample data point;

c. defining an object of interest as a single reference template having a number of data points in a data set for the reference template, wherein each data point defines a data level of grey shade;

d. computing ranks of grey shade in the reference template by ranking each sample data point in the data set for the reference template with respect to the number of other data points in the entirety of the data set for the reference template, the grey shade data levels of which are exceeded by the grey shade data level of the sample data point;

e. comparing the ranks of grey shade of said reference template with the ranks of grey shade of the selected section of said complex image, thereby determining closeness of fit; and

f. detecting an object of interest when the comparison exceeds a threshold for detection.

2. A method as claimed in claim 1, wherein the complex image frame has N×

M data points, and the reference template has n×

m data points, which is less than the N×

M data points, and the reference template is iteratively and incrementally compared with successive different sets of data points within the complex image frame N×

M set of data points.

3. A method as claimed in claim 2, wherein the reference template is iteratively and incrementally compared with successive different sets of data points within the complex image frame by iteratively incrementing an address of a set of data points, having the same number of data points as the reference template, within the N×

M set of data points.

4. An apparatus for locating and identifying an object of interest in a complex image frame, comprising:

a. an object model data base means for defining a plurality of objects of interest, each object of interest being defined as a single reference template having a number of data points in a data set, wherein each data point defines a data level of grey shade;

b. sampler means, receiving a complex image frame having a number of data points in a data set, greater than the number of data points in the reference template, wherein each data point defines a data level of grey shade, for providing a sampled digitized image frame, having the same number of data points as the reference template, and having an address defining the location of the sampled digitized image frame within the complex data frame;

c. incrementer means for incrementally and interactively controlling and moving the location of the sampled digitized image frame within the complex image frame, which is defined by an address, to interactively and incrementally compare the reference template with different sets of data points within the complex image frame;

d. rank generator means for replacing grey shade information in the sampled digitized image frame and also in the reference template with ranks, by ranking each sample data point in the data set for the sampled digitized image frame with respect to the number of other data points in the entirety of the data set for the sampled digitized image frame, the grey shade data levels of which are exceeded by the grey shade data level of the sample data point, and by ranking each sample data point in the data set for the reference template with respect to the number of other data points in the entirety of the data set for the reference template, the grey shade data levels of which are exceeded by the grey shade data level of the sample data point, thereby providing a ranked reference template and a ranked sampled digitized image frame;

e. correlating means for correlating the ranked template with the ranked sampled image frame and producing a correlation output signal indicative of the extent of correlation;

f. threshold generator/comparator means for providing a correlation threshold and comparing the correlation output signal with the correlation threshold, and producing an output signal indicating when the correlation output signal does not exceed the correlation threshold; and

g. controller means for controlling said object model data base means and receiving an output signal from said threshold generator/comparator means, and responsive thereto said controller means causes said incrementer means to increment the address of said sampled image frame, and wherein said incrementer means informs said controller means when a sampled digitized image frame has been completely incrementally scanned over the entire data set of the complex image frame, and said controller means controls said object model data base means to replace the reference template with another reference template for another model of interest.

5. Apparatus as claimed in claim 4, further comprising report generator means for generating an object identification report containing identification data identifying the reference template from said object model data base means and location data from said incrementer means.

6. Apparatus as claimed in claim 4, wherein said object model data base means has an input said from controller means, a first output connected to said sampler means, said rank generator, and said report generator means, and a second output connected to said controller means.

7. Apparatus as claimed in claim 4, wherein said sampler means has a first input from said object model data base means, a second input from said incrementer means, and an output connected to said rank generator means and said threshold generator/comparator means.

8. Apparatus as claimed in claim 4, wherein said incrementer means has a first output connected to said controller means and a second output connected to said sampler means and said report generator means.

9. Apparatus as claimed in claim 4, wherein said rank generator means has a first input from said object model data base means, a second input from said sampler means, and first and second outputs connected to said correlating means.

10. Apparatus as claimed in claim 4, wherein said correlating means has a first input from a first output of said rank generator means, a second input from a second output of said rank generator means, and an output connected to said threshold generator/comparator means.

11. Apparatus as claimed in claim 4, wherein said threshold generator/comparator means has a first input from said sampler means, a second input from said correlating means, and an output coupled to said controller means.

12. Apparatus as claimed in claim 4, wherein said controller means has a first input from said incrementer means, a second input from said threshold generator/comparator means, a third input from said object model data base means, and an output connected to said object model data base means.

13. Apparatus as claimed in claim 4, wherein said report generator means has a first input from said object model data base means, a second input from said incrementer means, a third input from said threshold generator/comparator means, and an output for providing said object identification report.

14. Apparatus as claimed in claim 4, further comprising a first input terminal means for inputting an image frame update data signal to said object model data base means and to said controller means, a second input terminal means for inputting digitized complex image frame data to said sampler means, and a report generator means for generating an object identification report containing identification data identifying the reference template from said object model data base means and location data from said incrementer means, and an output terminal means for outputting the object identification report.

15. Apparatus as claimed in claim 4, wherein said object model data base means has a first output for outputting identification data on the reference template, and a second output for outputting a data base exhausted signal indicating that all reference templates for all objects of interest have been completely incrementally scanned over the entire complex image frame, said sampler means has an output for outputting sampled image frame data, said incrementer means outputs an image scan complete signal at a first output, indicating that a sampled digitized image frame has been completely incrementally scanned over the entire complex image frame, and location data at a second output, said rank generator means outputs ranked reference template data at a first output and ranked sampled digitized image frame data at a second output, said correlator means outputs correlation output signal data at an output, said threshold generator/comparator means outputs identification status data at an output, and said controller means outputs a signal indicating the selection of a new object of interest at an output coupled to said object model data base means.