A simpler camera with embedded intelligence sounds like a good idea for machine vision. In fact, as editor Andy Wilson writes in his My View column in this issue, many vendors of machine-vision cameras are already heading in this direction, adding FPGAs, DSPs, and GPUs to their products so that their customers can build ever more sophisticated systems.
But wait, it seems that a group of 10-year-old kids is working on the same idea. Actually, it’s not quite the same idea since the kids are working on a truly simple camera called BigShot (www.bigshotcamera.org). The creator of BigShot is Shree Nayar, chairman of Columbia University’s computer science department and director of the Computer Vision Laboratory.
BigShot is a build-it-yourself digital camera from a kit with fewer than 20 parts that easily snap or screw together. Labels visible through a transparent cover point out the microprocessor, the memory chip, and other components that let this device capture, store, and reproduce images. Using an ingenious lens wheel lets the camera take normal, panoramic, and even stereoscopic pictures.
Nayar says, “The real point is to use the camera as an excuse to expose the kids to as many science and engineering concepts as possible.” It’s also more than that since in test sites in New York City, Bengaluru, India, and Vung Tao, Vietnam, the camera has served as a means for children of very different social and economic backgrounds to communicate and express themselves.
What can vendors and integrators of machine-vision products learn from such an undertaking? One lesson, perhaps, is that imaging is a very powerful and ubiquitous technology, with many markets and implications we are just beginning to understand. Furthermore, it is critical to educate young people in science and engineering and encourage some to follow these career paths. And finally, simplicity and transparency help make technology a tool rather than a barrier to development, whether it’s in manufacturing, security, biomedical research, or human relations.
In this issue, for example, an article by Andy Wilson shows that the development of new standards for high-speed camera-to-computer interfaces will simplify the task of machine-vision system integrators in creating applications that were not possible even a few years ago. Other articles describe developments in a range of applications improved by machine-vision systems, including the inspection of solar cell wafers, the sorting by color of plastic identification tags, and the image reconstruction of human physiology.
In ways we may not yet recognize, the future success of machine-vision and image-processing systems like these is being secured by the interest, enthusiasm, and energy of 10-year-olds fiddling with a do-it-yourself camera.