Target Detection through Robust Motion Segmentation and Tracking Restrictions in Aerial FLIR images

摘要

An efficient automatic moving target detection and trackingudsystem in airborne forward looking infrared (FLIR) imageryudis presented in this paper. Due to camera ego-motion, theseuddetection and tracking tasks are challenging problems.udBesides, previously proposed techniques are not suitable forudaerial images, as the predominant regions are non-textured.udThe proposed system efficiently estimates not only theudcamera motion but also the target motion, by means of anudaccurate motion vector field computation and robust motionudparameters estimation technique. This information allowsudaccurately to segment each target, and tracking them withudego-motion compensation. Verification of trackingudrestrictions helps detecting true targets while reducing very significantly the false alarm rate. Excellent results have been obtained over real FLIR sequences.