Research On The Spatial-time Searching Based Real-time Detection Technology Of Dim Moving Point Targets In Ir Image Sequences

The detection of infrared dim moving point targets is a hot issue nowadays. With the rapid development of Signal and information processing technology, the infrared target detection technique has been used in many applications, such as military, meteorology-satellite and medical science etc.With the characteristics of dim moving point targets in background and noise, no obvious shape, size and grain etc, the technology of detecting dim moving point targets is difficult in infrared image detection.The detection of infrared dim moving point targets can be divided in to DBT and TBD.It is difficult to guarantee performance of detecting dim moving point targets by using technology of Detect-Before-Track. We make use TBD to detect the dim moving point targets.The technology of TBD is applied for two methods. One is that the dim moving point targets are detected by projecting the spatial-temporal image sequence in to plane image; another is that detecting the dim moving point targets is the spatial-temporal image spaces. The former lose the information; reduce calculation amount and calculation time greatly; the latter is on the contrary. The researches on the dim moving point targets are the spatial-temporal image spaces in the thesis, and also reduce calculation amount and calculation time greatly.The two technologies of dim moving point targets are the spatial-temporal image spaces with TBD in the thesis. The first one, the moving message of targets is adopted which reduce calculation amount, calculation time greatly and structure the candidate trajectories .The energies along the candidate trajectories are accumulated and compared with the threshold .Angle restriction and inter-frame association is used ,and gets the possible targets more accurately. The other, the technology based on likelihood ratio function and the constant false alarm probability detect the dim moving point targets, and delete the trajectories of noises at the same time which calculation time greatly. The implementation of the algorithm and the simulation are finished based on MATLAB tools. Theoretical analysis and simulation results show the effectiveness of these two algorithms.