Hardware implementation of a virtual blind cane on FPGA

This paper presented an FPGA based real-time virtual blind cane system for the visually impaired. This system consists of a Xilinx ZYNQ-7000 device, a CMOS camera, an inertial measurement unit (IMU) and a line laser. With this system, the visually impaired can recognize the distance and shape of the barrier ahead in real time to avoid it. The computational complexity of processing images in real time can be very high. In addition, for most embedded systems, there is a payload limitation, high performance computer and GPU is not suitable for embedded system. In this research, FPGA was chosen as its computing platform. Contrastive experimental results are presented to show the speed and robustness.