Abstract
The digital camera is an excellent example of the current evolution of cyber-physical systems, i.e. systems intimately coupling mechanics, physics, electronics and software. It is also a wonderful example, accessible to all, of the power of computer science methods compared with those of physics and mechanics alone. The lecture will present the panoply of algorithms embedded in modern cameras and post-production software, then discuss the major impact they have on the design of cameras and lenses, which is currently undergoing a complete overhaul, and the impact they have on professional and amateur photographers alike.
Silver-based photography, which has a long history, progressed only slowly over the course of the 20th century: slow improvements in film and paper, the introduction of automatic exposure calculated analogically from photo cells, rangefinder or reflex aiming, all took decades. On the contrary, from the moment the first digital camera came onto the market in 1990, digital photography evolved extremely rapidly. By 2003, you could find decent semi-professional cameras, and by 2009, high-quality SLRs at affordable prices. Now there's a whole range of cameras of various sizes, all capable of delivering high-quality images. Even telephones have become very good photo and video cameras, mainly thanks to the algorithms they implement. As they can do many other things, such as immediately send images to the Internet, they are displacing the old small compact cameras and serving as the sole equipment for occasional photographers and for everyone in countries where film photography was unaffordable for the locals. The logic of digital photography has thus become very different from that of film, although the latter is still favored by some artists.