Image Processing has enabled large improvements in the performance of radio telescopes. Non-linear deconvolution and adaptive (self-) calibration techniques were developed in radio astronomy in the seventies and eighties. These processing techniques are now part of the standard tool kit used by any radio astronomer. In addition, these techniques have influenced the design of new telescopes. In the last decade, computing technology has enabled the development of similar but substantially more powerful image processing techniques. Moore’s Law has enabled larger problems to be solved, and object-oriented software techniques have enabled more physically realistic descriptions of radio telescopes to be used as a basis for image processing. I describe how these changes are being exploited to allow new processing methods on an unprecedented scale.