A type of artificial intelligence, computer vision robotics equips electronic or mechanical devices with sight for the purpose of achieving a particular task. The technology combines hardware and software that captures images, interprets the data, and performs a programmed response. The computer vision technology may involve nothing more than a video camera connected to a computer containing interpretive software. More elaborate forms of computer vision may include a network of infrared lasers and sensor receptors transmitting information to analytical software. Commonly used computer vision applications include industrial inspection, medical imaging, and surveillance.
The required software generally contains computer vision algorithms often originating from the Open Source Computer Vision Library, or Open CV. The library is a free database developed by the Intel Corporation for software designers. Using these algorithms, programmers create software that learns object recognition, receives visual data, and often initiates a mechanical response. The software might translate the information into images or record information of a particular area.
Computer vision robotics used by industrial manufacturers often includes high-speed video cameras or laser beams as well as sensors that can detect products traveling down a conveyor belt. This information passes to a computerized system that counts the items or inspects the quality of finished product. If an error occurs, the software might respond by triggering an alarm. Companies often use image recognition combined with mechanical interaction for sorting and packaging.
Converting two dimensional into three-dimensional images requires computer vision robotics. Cameras positioned around an object or location capture images from different angles. Software then compiles these images and renders a multidimensional depiction. Similarly, computerized technology converts energy signals emitted by medical imaging devices into multidimensional images that display the internal structures of the human body.
Video surveillance systems combine video cameras, monitoring a specific territory, with computer vision robotics software that is capable of analyzing motion tracking in real time. The software initially learns the difference between moving and stationary objects. The program then differentiates between human and nonhuman movement. Finally, programs distinguish between normal and abnormal walking patterns or peculiar limb movement.
Modern gaming systems incorporate computer vision robotics that recognizes human movement and translates the information into movements performed by animated figures onscreen. The Kinect® technology consists of a device that emits an infrared laser grid into the playing field. Motion sensors contained in the device detect beams reflected back toward the unit. The software interprets these signals in real time and in three-dimensional patterns. The program then transfers these movements onto human like representations on-screen.