RoboSight
RoboSight by RNIS offers an affordable, but powerful vision recognition system, ideal for robotic platforms and small embedded applications. Powered by the CogniMem CM-1K neural network chip developed by General Vision, complex visual objects are readily detected and categorized from patterns previously trained by the user or the embedding system. The recognized visual categories can then be fed to the higher-level functions of the controlling system to guide goal decisions, object localization, platform movement and navigation, and environment manipulation.
The RoboSight neural network vision system consists of a PIC18F4620 microcontroller interfaced with an Omnivision OV6630 CMOS camera module and General Vision's Cognimem CM-1K silicon neural network chip. The board can be interfaced to most computers, microcontrollers, or robots using the I2C or RS-232 serial interfaces. Once properly trained on a set of images (eg objects), the RoboSight is capable of recognizing these patterns and will output a match on the I2C read-only interface. Training of RoboSight's neural network can be done completely standalone using the segmented LED display, switch, and push buttons on the RoboSight's HID (human interface device) or for more advanced users, though the serial interface.
RoboSight by RNIS offers an affordable, but powerful vision recognition system, ideal for robotic platforms and small embedded applications. Powered by the CogniMem CM-1K neural network chip developed by General Vision, complex visual objects are readily detected and categorized from patterns previously trained by the user or the embedding system. The recognized visual categories can then be fed to the higher-level functions of the controlling system to guide goal decisions, object localization, platform movement and navigation, and environment manipulation.
The RoboSight neural network vision system consists of a PIC18F4620 microcontroller interfaced with an Omnivision OV6630 CMOS camera module and General Vision's Cognimem CM-1K silicon neural network chip. The board can be interfaced to most computers, microcontrollers, or robots using the I2C or RS-232 serial interfaces. Once properly trained on a set of images (eg objects), the RoboSight is capable of recognizing these patterns and will output a match on the I2C read-only interface. Training of RoboSight's neural network can be done completely standalone using the segmented LED display, switch, and push buttons on the RoboSight's HID (human interface device) or for more advanced users, though the serial interface.