Geolocation of remotely sensed pixels by introspective landmarking

• US 8,942,421 B1
• Filed: 11/07/2012
• Issued: 01/27/2015
• Est. Priority Date: 11/07/2012
• Status: Active Grant

• Current Assignee: Harris Corporation (L3Harris Technologies, Inc.)
• Original Assignee: Exelis Incorporated (L3Harris Technologies, Inc.)
• Inventors: Virgilio, Vincent Nicholas
• Primary Examiner(s): NAKHJAVAN, SHERVIN K
• Application Number:
  US13/670,974
• Time in Patent Office: 811 Days
• Field of Search: None
• US Class Current: 382/103
• CPC Class Codes: G06V 20/13 Satellite images

Abstract

A method of geodetically locating pixels of a captured image of a planetary body comprises the steps of: detecting an object on the planetary body using an imaging sensor viewing the planetary body; matching the object to a predetermined landmark on the planetary body; and updating, at a time t_k, a state vector representing kinematics of the imaging sensor, and t_k representing a present update time. Updating the state vector at the present time occurs if, and only if, the matching step is successful. In addition, the method includes computing a line-of-sight (LOS) vector from the imaging sensor to the planetary body, based on observations of the planetary body and the kinematics of the state vector; and geodetically locating the pixels of the captured image, based on the LOS vector. The LOS vector is based only on (a) the predetermined landmark and (b) a position command from a ground processing segment to the imaging sensor.

Claims

1. A method of geodetically locating pixels of a captured image of a planetary body, the method comprising the steps of:

detecting an object on the planetary body using an imaging sensor viewing the planetary body;

matching the object to a predetermined landmark on the planetary body;

updating, at a time t_k, a state vector representing kinematics of the imaging sensor, and t_k representing a present update time,

wherein updating the state vector at the present time occurs if, and only if, the matching step is successful;

computing a line-of-sight (LOS) vector from the imaging sensor to the planetary body, based on observations of the planetary body and the kinematics of the state vector; and

geodetically locating the pixels of the captured image, based on the LOS vector;

wherein the LOS vector is based only on (a) the predetermined landmark and (b) a position command from a ground processing segment to the imaging sensor.

2. The method of claim 1 wherein detecting and matching the object to the predetermined landmark includes

collecting an image from the imaging sensor of the detected object,

predicting an image of the detected object, and

correlating the collected image with the predicted image.

3. The method of claim 2 wherein correlating includes

finding a peak value resulting from correlating the collected image with the predicted image,

determining if the peak value is above a predetermined threshold value, and

flagging the matching step as successful, if the peak value is above the predetermined threshold value.

4. The method of claim 2 wherein updating the state vector includes

determining an offset vector between the collected image and the predicted image,

multiplying the offset vector by a Kalman gain, at the present update time, to obtain a state increment, and

updating the state vector using the state increment.

5. The method of claim 1 wherein the state vector includes the following parameters:

attitude parameters of the imaging sensor,

orthogonality of the imaging sensor, and

orbit parameters of the imaging sensor.

6. The method of claim 5 wherein

the attitude parameters include roll, pitch and yaw, representing a combined attitude of an orbiting platform and the imaging sensor disposed in the platform, and

the attitude parameters are in a fixed grid reference frame (FGF).

7. The method of claim 5 wherein

the orbit parameters include Kamel parameters representing an orbit'"'"'s deviation from the nominal orbit.

8. The method of claim 5 wherein

the orbit parameters include Kepler parameters describing an orbit'"'"'s deviation from the nominal orbit.

9. The method of claim 5 wherein

the attitude and orthogonality parameters are used to compute the LOS vector to a pixel in the captured image, and

the orbit parameters are used to compute an orbit vector from the imaging sensor to a center of the planetary body.

10. The method of claim 9 wherein

the LOS vector and the orbit vector are used to compute a geodetic location of each of the pixels in the captured image, the geodetic location provided in latitude and longitude of the planetary body.

11. The method of claim 1 wherein

the state vector is a state vector used in a Kalman filter, and

predicted observations of the planetary body are provided by an observation matrix used in the Kalman filter, and

the observation matrix is parameterized by scanning angles of the imaging sensor, and a current estimate of yaw in the state vector.

12. The method of claim 11 wherein

the observation matrix is further parameterized by an ideal curvature of the planetary body.

13. The method of claim 11 wherein the scanning angles are based on nominal FGF coordinates of the predetermined landmark, during the time t_k, and

a measured LOS in a line-of-sight reference frame (LRF), during a time greater than or equal to t_k−

1 and less than t_k,

wherein t<sub>k−

1</sub> is a time of a previous state vector update, based on a successful match.

14. A method of assigning geodetic locations to pixels in an image taken by an imaging sensor of a planetary body, the method comprising the steps of:

receiving a collected image including pixels in a neighborhood of an observed feature;

forming a predicted image including pixels in a neighborhood of a predetermined landmark;

flagging the collected image as likely including the predetermined landmark;

calculating an offset between the predicted and collected images, at a present time of t=t_k;

updating a state vector representing kinematics of the imaging sensor, based on the offset; and

assigning a geodetic location to a pixel captured by the imaging sensor, using a geodetic location of the predetermined landmark, at the present time of t=t_k.

15. The method of claim 14 wherein

assigning a geodetic location to a pixel captured by the imaging sensor includes using an observed location based on a scanning angle of the imaging sensor, at a previous time of t greater than or equal to t_k−

1 and less than t_k, where t<sub>k−

1</sub> is a previous time of updating the state vector.

16. The method of claim 15 wherein

the scanning angle of the imaging sensor is based on telemetry data from a ground station.

17. The method of claim 15 including the steps of:

performing star measurements to obtain another scanning angle of the imaging sensor, and

assigning a geodetic location to a pixel captured by the imaging sensor includes using the other scanning angle based on the star measurements.

18. The method of claim 15 wherein

the geodetic location of the pixel is assigned using only (a) the observed location and (b) the predetermined landmark, and

the geodetic location of the pixel is not assigned using star measurements.