Mark Twain once wrote, “You cannot depend on your eyes when your imagination is out of focus.” His observation on humanity has even more significance for the problem of computer vision. Although cameras can somewhat replicate the function of the eyes, computer vision cameras can allow systems to see and respond to what appears in the field of vision. While some cameras may be directly affixed to circuit boards, others come in more conventional shapes and sizes, depending on their technical requirements and capabilities. These might include typical digital or Web camera designs as self-contained units affixed with mounting ports.
Such designs might be incorporated as onboard personal computer (PC) and camera systems. Others may be installed in a vehicle. They may even be used on a remote-controlled platform or robot.
Usually considered a branch of artificial intelligence, computer vision relates to a form of machine pattern recognition — that is, teaching computers how to perceive not only pixels, but people, objects, terrains, and more. The objective of such processes is the simulation of visual intelligence, as defined by machine capabilities. Sight depends a great deal upon interpretation — natural for people and animals, but a challenge for computers.
To simulate vision, computers break down images into constituents, such as light and shadow, edges and fields. Other sensing techniques can include infrared (IR) or heat vision, and night vision, in addition to conventional digital imaging technologies typically using charge-coupled devices (CCD). Further varieties of computer vision cameras may distinguish still images, or compare video frames to detect or analyze motion differences between them.
Vehicle-mounted computer vision cameras may be used for personal digital video recording. These can track automobile movement in day or night environments. Some cameras feature light-emitting diode (LED) technology to aid with sensing, as well as motion detection and built-in microphones. Cameras attached to security systems may activate an alarm or transmit audio and video to a remote monitor, computer or smart phone. Usually, cameras differ according to resolution, scan field, and trigger modes. Some are designed for discrete or hidden placement, as with security systems.
Outdoor security cameras may be mounted in larger, weatherproof housing to deter crime. They may be programmed with various scanned fields for automatic or manual motion control. Submersible cameras are waterproofed for use in extreme environmental conditions.
Considered active vision technologies, such cameras can aid in navigation and dynamic orientation of a moving object within a moving environment — for example, in unmanned aerial vehicles (UAV). They are found in applications ranging from high-tech precision systems to personal uses. Computer vision cameras perform many duties with a powerful range of sensing and imaging technologies.
These products often come as part of complex security and monitoring systems. Accessories can include various cables, mounting hardware, adapters, and analyzing software. Wireless antennas can also permit remote placements. As part of smart systems for motion sensing, analysis, and trigger functions, computer vision cameras can perceive whole vistas of opportunities and applications.