摘要
针对组合仪表检测特点,设计并实现一种以图像处理单元(GPU)为核心的机器视觉仪表检测方案。通过分析通用PC和图像处理专用机的性能和特点,提出采用PC+GPU运算的新型架构,使系统能够兼顾柔性和实时性。进行仪表自动检测时由PC同时驱动下位机和摄像头,并将从摄像头传回的数据发送给GPU,经GPU进行图像处理后由CPU做出判断并输出检测结果。最后给出了实验验证结果,证明了方案的可行性。
According to the properties of the inspection of clusters, a new project of machine vision for automated visual inspection of clusters is designed and realized, which is based on GPU. Having analyzed and compared the characteristics of PC and the image processing machine, the system using an innovative architecture of PC with GPU is proposed, which ensures the flexibility and realtime performance. During the automatic inspection, PC drives the subordinate machine and the camera at the same time, and transfers the data that from camera to GPU, then CPU analyzes the result and gives the conclusion after the image processing by GPU. Finally, the experiment result is shown, demonstrating the feasibility of the project.
出处
《微计算机信息》
2011年第2期82-83,172,共3页
Control & Automation
