ORIENTATION INVARIANT OBJECT IDENTIFICATION USING MODEL-BASED IMAGE PROCESSING

• US 20110102419A1
• Filed: 11/12/2007
• Published: 05/05/2011
• Est. Priority Date: 11/13/2006
• Status: Active Grant

• Current Assignee: Cybernet Systems Corporation
• Original Assignee: Cybernet Systems Corporation
• Inventors: Cohen, Charles J., Haanpaa, Douglas, Jacobus, Charles J., Beach, Glenn J.
• Granted Patent: US 8,150,101 B2
• Time in Patent Office: Days
• US Class Current: 345/419
• CPC Class Codes: G06F 3/048 Interaction techniques based on graphical user interfaces [GUI]
• CPC Class Codes: G06T 15/00 3D [Three Dimensional] image rendering
• CPC Class Codes: G06T 15/10 Geometric effects
• CPC Class Codes: G06T 15/20 Perspective computation
• CPC Class Codes: G06T 2207/10012 Stereo images
• CPC Class Codes: G06T 2207/30212 Military
• CPC Class Codes: G06T 2207/30248 Vehicle exterior or interior
• CPC Class Codes: G06T 3/20 Linear translation of whole images or parts thereof, e.g. panning
• CPC Class Codes: G06T 3/60 Rotation of whole images or parts thereof
• CPC Class Codes: G06T 7/20 Analysis of motion motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 20/13 Satellite images

Abstract

A system for performing object identification combines pose determination, EO/IR sensor data, and novel computer graphics rendering techniques. A first module extracts the orientation and distance of a target in a truth chip given that the target type is known. A second is a module identifies the vehicle within a truth chip given the known distance and elevation angle from camera to target. Image matching is based on synthetic image and truth chip image comparison, where the synthetic image is rotated and moved through a 3-Dimensional space. To limit the search space, it is assumed that the object is positioned on relatively flat ground and that the camera roll angle stays near zero. This leaves three dimensions of motion (distance, heading, and pitch angle) to define the space in which the synthetic target is moved. A graphical user interface (GUI) front end allows the user to manually adjust the orientation of the target within the synthetic images. The system also includes the generation of shadows and allows the user to manipulate the sun angle to approximate the lighting conditions of the test range in the provided video.

Claims

1. A method of identifying an object, comprising the steps of:

a) storing geometric information about a plurality of candidate objects;

b) imaging a target object to be identified with a camera;

c) rendering stored geometric information to a simulated image utilizing pose and environment information;

d) comparing the image of the target object to the stored geometric information;

e) rotating and moving the image of the target object in three-dimensional space;

f) repeating steps c), d) and e) to determine the best match or matches between the target object and the candidate objects.

2. The method of claim 1, wherein:

the target object is imaged at a distance; and

it is assumed that the target object is positioned on relatively flat ground and that the camera roll angle stays near zero.

3. The method of claim 1, wherein:

the target object is imaged at a distance; and

three dimensions of motion (distance, heading, and pitch angle) are used to define the space in which the target object is moved.

4. The method of claim 1, further including the step of thresholding the images of the candidate objects and target object.

5. The method of claim 1, further including the step of providing a graphical user interface (GUI) allowing a user to manually adjust the orientation of the target object.

6. The method of claim 1, further including the step of generating and manipulating shadows during the comparisons between target object and the candidate objects.

7. The method of claim 1, wherein a graphics processor unit is utilized to perform rendering and comparison to greatly enhance processing speed.

8. A method of identifying a land vehicle, comprising the steps of:

a) storing geometric information about a plurality of candidate vehicles;

b) imaging a target vehicle to be identified with a camera;

c) rendering stored geometric information to a simulated image utilizing pose and environment information;

d) comparing the image of the target object to the stored geometric information;

e) rotating and moving the image of the target object in three-dimensional space;

f) repeating steps c), d) and e) to determine the best match or matches between the target object and the candidate objects.

9. The method of claim 8, wherein:

the target vehicle is imaged at a distance; and

it is assumed that the target vehicle is positioned on relatively flat ground and that the camera roll angle stays near zero.

10. The method of claim 8, wherein:

the target vehicle is imaged at a distance; and

three dimensions of motion (distance, heading, and pitch angle) are used to define the space in which the target vehicle is moved.

11. The method of claim 8, further including the step of thresholding the images of the candidate vehicles and target vehicle.

12. The method of claim 8, further including the step of providing a graphical user interface (GUI) allowing a user to manually adjust the orientation of the target vehicle.

13. The method of claim 8, further including the step of generating and manipulating shadows during the comparisons between target vehicle and the candidate vehicles.

14. The method of claim 8, wherein the target vehicle is a military vehicle.

15. The method of claim 8, wherein the target vehicle is a tank.

16. The method of claim 8, wherein a graphics processor unit is utilized to perform rendering and comparison to greatly enhance processing speed.