电子测量与仪器学报

戴伟杰,王衍学,李昕鸣,王祎颜.面向 FPGA 部署的改进 YOLO 铝片表面缺陷检测系统[J].电子测量与仪器学报,2023,37(9):160-167

面向 FPGA 部署的改进 YOLO 铝片表面缺陷检测系统

YOLO aluminum profile surface defect detection system for FPGA deployment

中文关键词: FPGA YOLOv2 算法高层次综合设计 PYNQ 异构计算

英文关键词:FPGA YOLOv2 algorithm high-level synthesis PYNQ heterogeneous computing

基金项目:国家自然科学基金(51875032,52275079)、北京建筑大学研究生创新项目(PG2023131)资助

作者	单位
戴伟杰	1.北京建筑大学机电与车辆工程学院
王衍学	1.北京建筑大学机电与车辆工程学院
李昕鸣	1.北京建筑大学机电与车辆工程学院
王祎颜	1.北京建筑大学机电与车辆工程学院

Author	Institution
Dai Weijie	1.School of Mechanical-Electronic and Vehicle Engineering, Beijing University of Civil Engineering and Architecture
Wang Yanxue	1.School of Mechanical-Electronic and Vehicle Engineering, Beijing University of Civil Engineering and Architecture
Li Xinming	1.School of Mechanical-Electronic and Vehicle Engineering, Beijing University of Civil Engineering and Architecture
Wang Yiyan	1.School of Mechanical-Electronic and Vehicle Engineering, Beijing University of Civil Engineering and Architecture

摘要点击次数: 963

全文下载次数: 1262

中文摘要:

在工业生产中,产品缺陷的智能检测是至关重要的。现场可编程门阵列(FPGA) 是一种具有算力强、功耗低等特点的嵌入式设备,能够将小型卷积神经网络部署其中。本文基于 Xilinx Zynq 系列 FPGA 设计了一套改进 YOLOv2 目标检测算法,在模型框架中增加重排序层,对切片图进行并行计算处理后再重组,完成铝片表面缺陷的检测。该算法经过高层次设计(HLS) 后,进行 RTL 转换与 IP 核封装,并导入到工程项目中完成 SoC 设计。通过综合、布局布线后生成比特流文件,导入至 PYNQ 镜像中,完成对铝片表面的工业缺陷检测。实验结果表明,本系统能够准确地检测出缺陷,并将功耗降低至 2. 494 W。

英文摘要:

In industrial production, intelligent detection of product defects is crucial. Field-programmable gate arrays ( FPGAs) are embedded devices with features such as high arithmetic power and low power consumption that enable small convolutional neural networks to be deployed in them. In this paper, a set of improved YOLOv2 target detection algorithm is designed based on Xilinx Zynq series FPGAs, and a reordering layer is added to the model framework to complete the detection of surface defects on aluminum sheets by parallel computing processing of the slice map before reorganisation. The algorithm is designed at a high level ( HLS), then RTL converted and IP cores are packaged and imported into the project to complete the SoC design. Generate bitstream files through comprehensive layout and wiring, import them into PYNQ images, and complete industrial defect detection on the surface of aluminum sheets. The experimental results show that this system can accurately detect defects and reduce power consumption to 2. 494 W.

查看全文查看/发表评论下载PDF阅读器