|
RADOVAN
ROBOTICS SCENE RECOGNITION |
Embedded Low Resolution
Scene Recognition
|
Radovan Robotics is proud to introduce, as one of our premiere technologies, Recognizer (c) embedded low-resolution scene recognition. Recognizer (c) technology is implemented on a very small embeddable circuit board. It combines DSP processing with a slightly reduced array resolution (256 x 244) to produce real time or near real time output at greatly reduced size, weight, implementation time and cost. Recognizer technology is much more reliable than PC based technology during storage, transport and usage. Recognizer (c) decisions are based on factory programmed algorithms, which are generated by Radovan Robotics to be specific to your custom application. It is easily embedded and it does not require software installation or imaging expertise on the part of the user. Compared with PC based imaging systems, the Recognizer (c) board offers size and weight reductions on the order of 100:1 and a recurring cost reduction on the order o f 10:1. It is designed to interface up to 8 standardized NTSC 640 x 488 pixel low cost video cameras or modules. Camera selection and analysis approach are predetermined by onboard firmware. Communication with a recognizer (c) board is simple. All of the high-speed activities: frame grabbing, filtering and analysis are performed by the on board hardware and DSP processor. The collected data is reduced on board as necessary for the application. Generally, the output data is target identification, X distance, Y distance and rotation. This data is presented through a single serial port and some TTL bits as needed for time-critical applications. Because of this design, the recognizer (c) can be treated as a sensor rather than a traditional image processor. Therefore, it can interface easily to anything from a single chip controller to a PC. While the recognizer (c) is an excellent choice for embedded low to medium resolution recognition tasks, its performance can be further enhanced by using it interactively with a larger system. For example, if a large number of different types of objects are to be identified, a recognizer board can outline one of the objects with chain coded data and present that data to a PC. The PC can then compare that chain with a database of object chains to make the full identification. These modes can be combined to form a stand-alone system that can be upgraded as needed. Please call the number below to request assistance in developing your application. |
|
Thank You and Best Regards, Phone: 805-375-7059 |
|
|
|
|
|
|
|
