基于NI EVS和PXI的机器视觉驾驶辅助系统开发技术研究

Study on development technologies of driver assistance systems based on machine vision using NI EVS and PXI

开发基于机器视觉的驾驶辅助系统时,需考虑多种外界因素,导致算法复杂、代码繁多。而开发初期,由于安全因素,又难以直接进行实车试验,无法得到有效验证。针对以上问题,本文提出一种基于NI EVS和PXI的机器视觉驾驶辅助系统的快速开发及测试方案。EVS嵌入式视觉系统适合快速检测及大型图像处理,能连接多架相机,并实现同步检测。借助LabVIEW编程语言及VDM编程模块可快速实现要求的识别功能,使机器视觉产品的开发更加高效。基于PXI的实时仿真平台,结合虚拟现实软件CarMaker,通过VeriStand集成整个测试系统,构建出真实有效的虚拟测试场景,实现机器视觉功能的全面、有效和可重复的测试。

When developing driver assistance systems based on machine vision,many environmental factors should be considered.This leads to complex algorithms and codes.Moreover,in early developing stage,the system can hardly be validated effectively,since it is impractical to do real vehicle tests because of security reasons.To solve this problem,a rapid development and test platform for driver assistance systems based on machine vision is designed by using NI EVS and PXI.EVS is suitable for fast developing of image processing and detection algorithms.It can connect multiple cameras and realize synchronous acquisition.With LabVIEW and VDM module,the developing process of detection algorithm and its product is greatly simplified and improved.The real time platform based on PXI and the virtual reality software CarMaker are integrated by VeriStand.Thereby the developing platform is set up,which can simulate effective virtual test scenarios for comprehensive,effective and repeatable tests of machine vision systems.