Embedded intelligent camera algorithm based on hardware IP

Abstract

This article designed an embedded smart camera based on the ZYNQ-7000 development board, which has an ARM+FPGA architecture. The system functions are divided into three parts: image acquisit ion, image processing and image display, each of them adopting a modular design. The image acquisition part adopts OV7725 camera. The image processing section uses the VIVADO HLS development kit to design hardware IP-based image processing algorithms. Compared with the traditional HDL-based IP core design method, HLS method that use C/C++ to design IP reduces the development difficulty and shortens the development cycle. Hardware IP implemented through HLS includes image preprocessing, spot-based centroid positioning algorithm, edge detection and corner detection. For complex image processing algorithms, it is still used software to implement, while preprocessing are accomplished through hardware IP. The VDMA channel realizes high-speed data interaction between hardware and software. The image display part is based on the embedded Linux system, OpenCV, and cross -platform C++ graphical user interface development framework QT to implement the GUI interface. The entire embedded smart camera system finally achieves accelerat ion processing based on hardware IP, such as spot location, sobel edge detection, and corner detection.