Abstract: The imaging characteristics of midcourse ballistic closely spaced objects (CSO) on an infrared focal plane array (IR FPA) are analyzed, and it is pointed out that space-based infrared sensors should super-resolute CSO while tracking on them. A novel model-based algorithm of joint super-resolution and ballistic trajectory estimation for CSO is presented which combines models of IR FPA imaging and midcourse ballistic dynamics, and establishes an objective function of joint super-resolution and trajectory estimation based on the least squares criteria from space and time information. In addition, CSO initial state parameters are selected as model parameters by analysis. To cope with the high-dimensional and nonlinear characteristics of the original objective function, a dimension-decreased function is inferred, and then quantum-behaved particle swarm optimization is included to estimate directly the model parameters, and subsequently the trajectory and radiant intensity of the objects are calculated, thus ultimately realizing midcourse CSO joint super-resolution and ballistic trajectory estimation. Simulation results confirm the effectiveness of the algorithm. Meanwhile, the results also show that, in contrast to traditional methods, the new algorithm not only avoids complex data association but also gains stronger resolution capability and higher precision in location.

Key words: closely spaced objects, super-resolution, ballistic trajectory estimation, particle swarm optimization, infrared multi-sensor, midcourse ballistic