Whilst the underlying mechanisms of atmospheric turbulence are complex, the observed effects on imaging can be described in simpler terms. In this thesis, I address the effects seen as geometric distortions in anisoplanatic imaging and propose new digital restorations techniques that are real-time capable and predictive. The anisoplanatic problem arises in wide-field telescopic imaging and in new ventures of astronomy such as giant telescopes that process wide-field imagery. The methods proposed here, both digital and digital-optical hybrid, remove the position dependent distortions as a precursor to image analysis. Previous existing digital restoration techniques have used a prototype formed by averaging an image time sequence for image registration where valuable high frequencies information is lost due to the low-pass filtering effect of averaging. The proposed techniques are capable of using any arbitrary frame in the sequence as prototype, thus circumventing the low pass filtering effect and also allowing real-time implementation. Furthermore, these techniques are made predictive by the use of Kalman filtering. The predictive capabilities of these techniques open a new path to the combination of digital processing and adaptive optics that can result in hybrid systems. The key to adoption of hybrid systems is to reduce the complexity and expense of the optics and couple this with digital processing prediction. To this end I also propose a new type of inexpensive and fast piezoelectric deformable mirror based on the vibration modes of circular PVDF membranes that exhibit striking similarities to Zernike polynomials. It requires only two electrodes for actuation and a very simple driving signal generator, therefore constituting an inexpensive and viable alternative to existing deformable mirrors. With the emergence of multi-conjugate adaptive optics (MCAO) and multiobject adaptive optics (MOAO) in astronomy, and the more demanding correction required for long range surveillance imaging, this inexpensive deformable mirror and the real-time capable digital algorithms are promising building blocks for a hybrid solution to the anisoplanatic imaging problem.