Compressive moving target tracking with thermal light based on complementary sampling

We present a method to directly reconstruct background subtracted images using compressive sensing (CS) theory and complementary modulation technique. Since the moving objects of interest only occupy a small portion of the field of view, i.e., they are sparse in the spatial domain, and the complementary modulation strategy makes the values of patterns in the suitable range for CS reconstructions, thus, with this method, we can retrieve object silhouettes and trajectory with high image quality and strong robustness against the changing illumination and the noise. To demonstrate the performance of the proposed protocol, we make a comparison of our strategy with traditional methods. This protocol may attract general interest and be instructive for the fields of surveillance systems, teleconferencing, and even search and rescue missions.