G. Pisinger (Germany)
point spread function, sensor modeling, circle fitting, sub pixel precision
In this article an approach to the image restoration problem is presented. After the description of the general model, a method for the approximation of the point spread function with tensor product B-splines is derived. Normally, images are treated as point samples. Here, image values are consid ered as area samples generated by overlapping integrators, necessary to model the inter-pixel effects of a CCD cam era. In addition the approximation problem was adapted to the case where a suitable noise model was considered. Using the image restoration model, a new technique for model based circle fitting is derived. This method includes the effects of the point spread function. Tests on real and simulated data indicate the effectiveness of the proposed algorithm.
Important Links:
Go Back
