Abstract: Computational photography is an integrative technology which combines the computer and software technology with modern sensors and modern optics to create new imaging systems and image applications. In succession on the basis of existing photography, computational photography improves the components, workflow and even the principle of photography to break out of the limitation of traditional photography. It becomes a new revolution of photography. Computational photography is a cross-research area of many disciplines which can solve a problem through different perspectives and approaches. To fully understand this new area, computational photography is classified to computational scenes, computational optics, computational sensors and computational processing. In each subclass, some hot issues and typical projects are introduced. Finally, the existing circumstances of computational photography and trend in development are discussed.