Multi-resolution object classification method employing kinematic features and system therefor

• US 6,393,137 B1
• Filed: 06/17/1999
• Issued: 05/21/2002
• Est. Priority Date: 06/17/1999
• Status: Expired due to Term

• Current Assignee: Raytheon Company (Rtx Corporation)
• Original Assignee: Raytheon Company (Rtx Corporation)
• Inventors: Chen, Hai-Wen, Riddle, Jack G., Schmitt, Harry A.
• Primary Examiner(s): Johns, Andrew W.
• Assistant Examiner(s): AZARIAN, SEYED H
• Application Number:
  US09/335,427
• Time in Patent Office: 1,069 Days
• Field of Search: 382/132, 382/128, 382/103, 382/240, 382/219, 382/248, 382/209, 382/217, 382/218, 382/190, 382/191, 342/90, 378/62, 607/27, 600/509, 600/578
• US Class Current: 382/103
• CPC Class Codes: G06V 10/431 Frequency domain transformation; Autocorrelation

Abstract

A multi-resolution feature extraction method and apparatus. In the illustrative embodiment, the feature extractor includes circuitry for receiving and transforming a time variant data signal into a multi-resolution data signal. The multi-resolution data signal is compared to each of a plurality of object templates. The system then generates a feature vector based on a correlation of the multi-resolution data signal to one of the object templates. The multi-resolution feature extraction method employs object templates formed by transforming time variant image data for each of a plurality of objects into a respective multi-resolution template and averaging all templates for each respective object. The method includes steps for transforming an incoming time variant data signal into a multi-resolution data signal, comparing the multi-resolution data signal to each of the object templates, and generating a feature vector when the multi-resolution data signal correlates to one of the object templates. In a more specific implementation, the method further includes the steps of calculating a confusion matrix (CM), classifying the feature vectors as one of the objects to thereby produce classified objects responsive to the CM, and selecting a target from the classified objects. A multi-resolution feature extractor according to the present invention employs object templates formed by transforming time variant image data for each of a plurality of objects into a respective multi-resolution template and averaging all templates for each respective object to thereby generate object templates.

Claims

1. A multi-resolution feature extraction method employing object templates formed by transforming time variant image data for each of a plurality of objects into a respective multi-resolution template and averaging all templates for each respective object to thereby generate object templates, said method comprising the steps of:

transforming an incoming time variant data signal into a multi-resolution data signal;

comparing the multi-resolution data signal to each of the object templates; and

generating a feature vector when the multi-resolution data signal correlates to one of the object templates.

2. The multi-resolution feature extraction method as recited in claim 1, wherein the transforming step performs a discrete wavelet transform step.

3. The multi-resolution feature extraction method as recited in claim 1, wherein the transforming step performs a discrete wavelet packet step.

4. The multi-resolution feature extraction method as recited in claim 1, wherein the transforming step converts the incoming time variant data signal into a multi-resolution data-signal which is localized in both time and frequency.

5. A multi-resolution feature extraction method comprising the steps of:

transforming time variant image data for each of a plurality of objects into a respective multi-resolution template;

averaging all templates for each respective object to thereby generate object templates;

transforming an incoming time variant data signal into a multi-resolution data signal;

comparing the multi-resolution data signal to each of the object templates; and

generating a feature vector each time the multi-resolution data signal correlates to one of the object templates.

6. The multi-resolution feature extraction method as recited in claim 5, wherein each of the transforming steps performs a discrete wavelet transform step.

7. The multi-resolution feature extraction method as recited in claim 5, wherein each of the transforming steps performs a discrete wavelet packet step.

8. The multi-resolution feature extraction method as recited in claim 5, wherein the transforming steps converts the incoming time variant image data and incoming time variant data signal into a multi-resolution template and a multi-resolution data-signal, respectively, each of which is localized in both time and frequency.

9. A method for operating a multi-resolution seeker system, comprising the steps of:

converting time variant image data for each of a plurality of objects into a respective multi-resolution template;

averaging all templates for each respective object to thereby generate object templates;

transforming an incoming time variant data signal into a multi-resolution data signal;

comparing the multi-resolution data signal to each of the object templates; and

generating a feature vector each time the multi-resolution data signal correlates to one of the object templates;

calculating a confusion matrix (CM);

classifying the feature vectors as one of the objects to thereby produce classified objects responsive to the CM; and

selecting a target from said classified objects.

10. The multi-resolution feature extraction method as recited in claim 9, wherein the transforming step performs a discrete wavelet transform step.

11. The multi-resolution feature extraction method as recited in claim 9, wherein the transforming step performs a discrete wavelet packet step.

12. The multi-resolution feature extraction method as recited in claim 9, further comprising:

computing a probability-of-error vector (PV) when there are more than one of the objects in the incoming time variant data signal and

selecting a target from said classified objects responsive to the PV.

13. A multi-resolution feature extractor employing object templates formed by transforming time variant image data for each of a plurality of objects into a respective multi-resolution template and averaging all templates for each respective object to thereby generate object templates, comprising:

a memory which stores the object templates and

a feature extractor which transforms an incoming time variant data signal into a multi-resolution data signal, compares the multi-resolution data signal to each of the object templates, and generates a feature vector each time the multi-resolution data signal correlates to one of the object templates.

14. A multi-resolution seeker system, comprising:

a memory;

a feature extractor operatively connected to the memory which converts time variant image data for each of a plurality of objects into a respective multi-resolution template, averages all templates for each respective object to thereby generate object templates, transforms an incoming time variant data signal into a multi-resolution data signal, comparing the multi-resolution data signal to each of the object templates, and generates a feature vector each time the multi-resolution data signal correlates to one of the object templates;

a processor operatively coupled to the memory, which processor calculates a confusion matrix (CM);

a classifier operatively connected to the memory which classifies the feature vectors as one of the objects responsive to the CM to thereby produce classified objects; and

a selector operatively coupled to the memory which selects a target from said classified objects.

15. The multi-resolution seeker system as recited in claim 14, wherein:

the processor includes means for computing a probability-of-error vector (PV) when there are more than one of the objects in the incoming time variant data signal and

the selector selects the target from said classified objects responsive to the PV.

16. A feature extractor comprising:

first means for receiving and transforming a time variant data signal into a multi-resolution data signal,

second means for comparing the multi-resolution data signal to each of a plurality of object templates, and

third means for generating a feature vector based on a correlation of the multi-resolution data signal to one of the object templates.

17. The invention of claim 16, wherein the object template is formed by transforming time variant image data for each of a plurality of objects into a respective multi-resolution template and averaging all templates for each respective object.