This paper investigates real-time target determination,
recognition and tracking with an airborne video system.
The real-time target determination, recognition and tracking
with an airborne video system has been played in movies
and dramas. However, it was not realized in the past. The
difficulty lies on the determination of color threshold and
shape matching. In this research, we proposed an algorithm
to solve this difficulty. Experiments also verify that the
successful rate of this algorithm is higher than 80 %. The
processing rate can reach 20 to 30 Hz even with an onboard
computer, implying that this algorithm is applicable to a real
airborne system. Our algorithm is potentially applicable to
airborne video tracking system in a UAV.