Based on a series of operations such as detecting candidate target trajectories

the optimal algorithm of maximum likelihood ratio judgment and post-processing

etc.

this paper presents a kind of adaptive algorithms that can automatically recognize the point targets making straight movement with uniform velocity in the field of view from low SNR image sequence in which CN- and UCN- sequences are all suppressed. The theoretical analyses and optimal algorithms of various kinds of operations are expounded in the paper. The experimental results have proved the correctness of theoretical analyses and the feasibility of the adaptive algorithms. The lowest distinguishable SNR is 0.5

and the optimal algorithms are compatible with infrared- and television image sequences. The quantity of data processing can be also greatly reduced by the optimal algorithms.