FPGA board for real time vision development systems

Abstract

Describes an FPGA based system for real time vision applications at a robotics vision laboratory. It contains an ALTERA FPGA (20K100), an interface with a digital camera, three VRAM memories to contain the data input and some output memories (a VRAM and a EDO) to contain the results. The aim of this system is to develop and test vision algorithms, such as image compression, optical flow calculation with differential and correlation methods and to hold applications that use previous calculated optical flow, like image invariants detection. The designed system is able to connect the digital camera, or the FPGA output (results of algorithms) to a PC, through its ports USB, IEEE 1394 or CENTRONICS. It also has an expansion connector to make possible the expansion of the system, with some elements of the process, for instance one system would calculate the optical flow of the input images and another system would calculate the focus of expansion in the image and the time to impact in the scene.