APP下载

基于改进HOG算法的AGV小车避障研究

2020-05-28万伟刘子龙

软件导刊 2020年3期

万伟 刘子龙

摘 要:为解决激光传感器等避障装置难以解决爆炸性危险环境下的防爆问题,提出一种基于机器视觉的AGV小车避障解决方案。磁导航AGV小车在工作中路径相对固定,主要针对来回走动的工人进行检测,进而实现减速或制动。为实现高精确度下实时检测,采用改进HOG算法,结合线性支持向量机实现更快和更可靠的分类。实验证明,该方法识别率达到92.84%,漏检率4%,其准确度和实时性基本满足危险环境下的行人检测要求。

关键词:AGV; 行人检测; HOG; 线性支持向量机; 危险环境

DOI:10. 11907/rjdk. 191617

中图分类号:TP311.5   文献标识码:A                文章编号:1672-7800(2020)003-0006-04

Research on AGV Vehicle Obstacle Avoidance Based on Improved HOG Algorithm

WAN Wei, LIU Zi-long

(School of Optical Electrical and Computer Engineering,University of Shanghai for Science and Technology, Shanghai 200093, China)

Abstract: In order to solve the problem of explosion-proof in the explosion-hazard environment, a machine vision-based AGV trolley obstacle avoidance solution is proposed. The magnetic navigation AGV trolley has a relatively fixed path during work, and is mainly used for detecting and moving around the working path to realize deceleration or braking. To ensure real-time detection with high accuracy, the improved directional gradient histogram (HOG) algorithm and linear support vector machine (LSVM) are used to achieve faster and more reliable classification. The experiment proves that the recognition rate of the method reaches 92.84%, and the missed detection rate is 4%. Its accuracy and real-time performance can basically meet the pedestrian detection in industrial hazardous environment.

Key Words: AGV; pedestrian detection; HOG; Linear support vector machine; dangerous environment

0 引言

AGV(Automated Guided Vehicle)是装备有电磁或光学等自动导引装置,能够沿规定的导引路径行驶,具有安全保护和移载功能的运输车[1]。AGV小车由于其灵活性和大负载能力越来越多地应用于工业场所。通用型AGV小车避障手段有超声波传感器、红外传感器和激光雷达等,这些传感器多属于光辐射和电磁波射频源[2]。2017年5月,《自动导引车(AGV)在危险生产环境应用的安全规范》标准启动会在中国科学院沈阳自动化研究所召开,意味着应用于爆炸性危险环境下的特种AGV将逐步实现标准化、规范化。

AGV小车主要由车载控制器、驱动系统、导航系统和避障系统等几大部分组成[3]。控制、驱动等几个部分可以通过隔爆型、增安型、本质安全型等防爆方式解决防爆问题。文献[4-5]通过超声波传感器实现AGV小车的避障模块设计,Zoltan Rozsa等[6]采用的激光雷达避障系统是通用型AGV小车最常见的避障方式。……

登录APP查看全文