Developments occurring in VLSI, VHSIC, and mosaic focal plane technologies are spawning an improved generation of intelligent infrared sensors that will be able to detect and track airborne vehicles from a spacecraft positioned in a high altitude orbit about the earth. These targets appear in the focal plane of a sensor as faint unresolved targets moving against an intensely cluttered background. Clutter can seriously degrade the detection of targets and results from phenomena such as drift and jitter in the sensor's line-of-sight and time-varying components in the background.; An algorithm for detecting faint unresolved targets moving against cluttered backgrounds is developed. The algorithm is founded on the principle that targets in a sequence of images form well-defined patterns often referred to as target tracks while false alarms caused by noise and clutter are more random. Target tracks generally appear as straight lines for nonmaneuvering targets and as curves for maneuvering targets. The target detection algorithm incorporates third-order frame-to-frame differencing, the concept of a fading target track map, the Hough transform, and M-ary signal detection using matched filters to detect nonmaneuvering targets and to estimate the current position, direction, and speed of these targets. The performance of the algorithm on a synthetic data base is discussed in addition to an implementation of the algorithm using special purpose hardware.; Two models of an unresolved target moving against a cluttered background are also discussed. An analytical model is used to study the processes responsible for the formation and degradation of infrared imagery from an earth-staring mosaic sensor. A simulation model is used to generate a data base with known characteristics for evaluating the performance of the target detection algorithm.
展开▼
