The eyes have it

In many a primer on machine vision and image processing systems, the camera is often compared with the human eye and the computer to the human brain. In reality, of course, the human eye/brain combination is far more sophisticated than any currently available camera/computer-based system.

Indeed, today's commercially available image sensors and the cameras that incorporate them bear little resemblance to the complex biological structure of the human eye. Similarly, although certain types of computer programs such as neural networks attempt to emulate the way the brain functions, the CPUs that execute them with their Harvard and von Neumann architectures cannot be compared to the structure of neurons in the brain.

To develop sophisticated systems that emulate the human visual system requires circuits that mimic neuro-biological architectures. Such neuromorphic engineering is, at present, the domain of researchers such as Professor Tobi Delbruck at the Institute of Neuroinformatics, who has developed a vision sensor that in some respects emulates the workings of the human retina as you will read on page 11 of this issue.

In future, such research may lead to image sensors and camera systems that fully emulate the capability of the human eye. While such sensors may be useful in certain applications such as particle tracking, however, they may not be widely adopted by systems integrators looking to choose sensors and cameras to meet a range of machine vision applications.

In high-speed inspection, for example, vision systems can be deployed to inspect products as they are produced at rates greater than hundreds of parts per minute. In medical applications, infrared cameras may be used to analyze physiological functions related to skin temperature. More sophisticated multi-spectral systems that combine multiple frequencies are especially useful in applications such as food inspection, medical, and military imaging systems.

In such applications, the speed at which objects must be inspected or the wavelengths required to make such measurements will far outperform the capabilities of any human being. In these and many other types of vision systems, currently available camera/computer systems are sufficient to achieve the desired result.

To choose the types of cameras used to perform such tasks requires evaluating the specific types of cameras, speeds, sensors and interfaces that are available. To help you do so, Vision Systems Design is proud to publish our eleventh Camera Directory issue.

Within these pages, you will find numerous companies categorized by name, interface and other specifications. Readers who prefer to search online can find this Camera Directory at www.vision-systems.com, where additional vision products, manufacturers and specifications can be found.