This paper considers the use of soft computing methods in solving the problem of super-resolution, i.e. the reconstruction of a high-resolution image from a series of low-resolution video frames. By means of carefully selected juxtaposition of frame expansion, registration, re-mapping and fusion techniques, it is possible to obtain a high-quality high-resolution image summarizing a given video stream. A hybrid evolutionary algorithm used in the enlargement process is shown to contribute much to the overall acceptability of the super-resolved result. Our superresolving system is self-contained. Each component operator and each process is well delineated and easy to code. The applicability of the proposed system is shown by experimental results.
展开▼