基于FPGA与SVM实时双目视觉巡检机器人避障研究  

Research on Obstacle Avoidance of Real-Time Binocular Vision Inspection Robot Based on FPGA and SVM

作　　者：樊民革[1] 吴珈漉 FAN Min-ge;WU Jia-lu(School of Automation and Software,Shanxi University,Tiayuan 030031,China)

机构地区：[1]山西大学自动化与软件学院,山西太原030031

出　　处：《机械工程与自动化》2024年第6期56-57,61,共3页Mechanical Engineering & Automation

摘　　要：设计了一种基于FPGA和SVM二分类的实时双目视觉检测系统。探讨了双目立体视觉理论及FPGA加速算法,并通过500次避障仿真实验验证了在不同障碍物前的高效避障能力。实验结果显示,视觉检测系统显著提高了巡检效率。A real-time binocular vision detection system based on FPGA and SVM binary classification has been designed.Explored the theory of binocular stereo vision and FPGA acceleration algorithm,and verified the efficient obstacle avoidance ability in front of different obstacles through 500 obstacle avoidance simulation experiments.The experimental results show that the visual inspection system significantly improves the inspection efficiency.

关 键 词：智能巡检机器人 双目立体视觉 FPGA SVM二分类 避障实验 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]