Research On The Detection Technology For Dim Moving Point Targets In Image Sequences

"Study on Detection and Tracking Dim Point Moving Target from Image Sequences ",regarded as a new branch in the field of active visual research,has nowadays become a research focus because of its more and more applications. Moreover, the dim point target's own characteristics, low contrast and random motion and the like make it more difficult to search and detect targets in images.First, the related theories of the detection of dim point target, spatial-temporal diversity in mobile communication field, and TDB detection algorithm are described in detail. In this paper, we apply these theories and algorithms to the detection of dim point Target from Image Sequences, and make researches on the technologies of target searching and trajectory estimation based on region division in combined frame image, Probability Weights of target group, and inter-frame mutual relations. This technology first combines image sequences in time domain to eliminate the background effect, then mesh the combined frame image to estimate the number of targets and the targets'existent domain. So each target can be detected accurately while the computational complexity is reduced, and then each target's track is initialized by using the mutual relations characteristics of inter-frame targets.In this paper, simulation is made for the algorithm by using MATLAB. Simulation results and experimental data have proved the algorithm's feasibility and effect ,and have certain application value.