Methods and apparatus for detecting anomalies in images according to various aspects of the present invention operate in conjunction with an optical How system. The optical flow system may receive image data for a first image and a second image and generate How vectors corresponding to differences between the first image and the second image. An analyzer may be coupled to the optical flow system and analyze the flow vectors to identify anomalies in the second image.
Systems and methods of tracking a beam-aspect target are provided. In embodiments, a target is tracked with a Kalman filter while detections are received. After a detection is missed, the Kalman filter may be concurrently propagated with a blind-zone particle filter until a probability that the target is in a blind zone exceeds a threshold. When the probability exceeds the threshold, the Kalman filter may refrain from further propagating. After a gated detection is received, the blind-zone particle filter and an unrestricted-zone particle filter may be concurrently propagated while a probability that the target is in an unrestricted zone exceeds a threshold. The system may return to tracking with the Kalman filter when a covariance of the unrestricted-zone particle filter falls below a predetermined covariance.
Systems and methods of tracking a beam-aspect target are provided. In embodiments, a target is tracked with a Kalman filter while detections are received. After a detection is missed, the Kalman filter may be concurrently propagated with a blind-zone particle filter until a probability that the target is in a blind zone exceeds a threshold. When the probability exceeds the threshold, the Kalman filter may refrain from further propagating. After a gated detection is received, the blind-zone particle filter and an unrestricted-zone particle filter may be concurrently propagated while a probability that the target is in an unrestricted zone exceeds a threshold. The system may return to tracking with the Kalman filter when a covariance of the unrestricted-zone particle filter falls below a predetermined covariance.