UtopiaCompression Corporation

Passive Range Detection Using Monocular Video

UC’s Passive ranging technology is being built for Unmanned Aerial Vehicles such as the Global Hawk to aid them in obstacle avoidance and path planning

Range estimation of remote targets is a challenging problem. The use of stereo cameras on UAS is (or may be) impractical due to the need for a platform that supports large baseline separation between the two cameras that make up the stereo pair. Structure from Motion may not yield accurate depth information when the target size of the perceived objects is small. Likewise, maneuvering the sensor aircraft can be disruptive to the mission and cause inconvenience to both operators and regulators. Though Bearings measurement has proved to be a good range estimator, it might not be useful in certain scenarios. For example, when the bearing angle of an approaching intruder is too small, or is changing very slowly, it may be necessary for the observer to perform maneuvers to introduce relative movement between the observer and the target.

In order to overcome this problem, UtopiaCompression Corporation (UC) developed a novel Interacting Multiple Models - Particle Filter method for achieving target range estimation using a single camera. The algorithm fuses information obtained from particle filters using bearing angle measurements - in both azimuth and elevation, and a general model of the variation of target size with observer - target kinematics. It is worth noting that we do not make any assumption about the size, shape, type or any other features of the target. Furthermore, as this approach does not depend solely on the bearing measurement, the sensor vehicle does not need to make deliberate maneuver to obtain a significant bearing angle measurement.

Sense and Avoid Program

Range and bearing estimation from 2-D imagery critical to a number of defense related applications. Two such applications of great interest to AFRL/SN and DoD in general are Sense and Avoid (SAA) for Unmanned Aerial Systems (UAS) and missile warning systems (MWS) for manned aircraft. Robust and quick intruder location estimation using a single electro-optical/infrared sensor is extremely valuable to safely operate in the National Air Space (NAS).

Performance of UC’s Passive Ranging Algorithm on video captured from Beech Bonanza aircraft (200 frames)

Actual initial Range = 3.5 nmi

Error in initial guess = -7.9991 nmi

Final range error = 0.3319 nmi

Average error = 0.5089 nmi