Low-Resolution Target Classification from a Staring Infrared Sensor.
Abstract
A unique method to extract and classify aircraft targets from low-resolution staring infrared data is presented which uses sequences of sensor frames to obtain target signature samples. The classifier is based on a target feature description consisting of X and Y energy projections of the target. The target information is first extracted from sensor data through sampling windows consisting of contiguous rows and columns of detectors in the path of the expected target. The samples are then reduced to remove degradations from noise and sensor optics. The classifier assumes that approximate target position and velocity are available from a target detection and track subsystem. Tests were performed using a software sensor simulation that includes optical blur and sensor motion models. The classifier is shown through simulation to successfully extract and classify various transport and fighter class aircraft with sensor footprint resolution of 11 to 45 meters. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1984
- Accession Number
- ADA151690
Entities
People
- J. N. Tilley
Organizations
- Air Force Institute of Technology