Image Processing and Vehicles – Using FPGA to Reduce Latency of Time Critical Tasks

Abstract

When automating complex vision-controlled objects such as self-driving cars, large amounts of data need to be processed accurately, quickly and reliably at video frame rates. In this paper we propose the use of an Intel Cyclone V FPGA to process image data in a parallel form, building a safe real-time system. We discuss the hardware used to build the physical prototype and the control algorithms to build the control architecture that controls the prototype.