Bayesian superresolution

Well-known principles of physics explain why resolution restrictions occur in images produced by optical diffraction-limited systems. The limitations involved are present in all diffraction-limited imaging systems, including acoustical and microwave.; In most circumstances, however, prior knowledge about the object and the imaging system can lead to resolution improvements. In this dissertation I outline a method to incorporate prior information into the process of reconstructing images to superresolve the object beyond the above limitations. This dissertation research develops the details of this methodology. The approach can provide the most-probable global solution employing a finite number of steps in both far-field and near-field images. In addition, in order to overcome the effects of noise present in any imaging system, this technique provides a weighted image that quantifies the likelihood of various imaging solutions. By utilizing Bayesian probability, the procedure is capable of incorporating prior information about both the object and the noise to overcome the resolution limitation present in many imaging systems.; Finally I will present an imaging system capable of detecting the evanescent waves missing from far-field systems, thus improving the resolution further.