基于单目视觉的AGV间防撞系统设计

Design of AGV anti-collision system based on monocular vision

为了解决智慧工厂中AGV行进过程中面对其他车辆无法自动躲避的问题,本文设计了一种基于单目视觉测距的AGV防撞系统。该系统在AGV顶部安装摄像头,用于采集AGV行驶过程中的前方图像,车体内安装微处理器对图像进行分析,并将分析结果转换为控制信号控制AGV,从而达到防撞的目的,其中图像分析首先采用三帧差分法将图像中的AGV目标检测出来,然后基于逆投影变换测距原理计算出AGV间的距离。实验结果表明,该系统在模拟实验环境下测量精度基本控制在10%以内,4 m内最大绝对误差为51 mm,能够满足物流工厂AGV的设计要求。

In order to solve the problem that AGV can't avoid other vehicles automatically in the process of moving in intelligent factory,this paper designs an AGV anti-collision system based on monocular vision ranging.The system installs a camera on the top of AGV to collect the front image of AGV during driving.A microprocessor is installed in the car body to analyze the image,and the analysis results are converted into control signals to control the AGV,so as to achieve the purpose of anti-collision.In the image analysis,the AGV target in the image is detected by three frame difference method,and then the distance measurement is calculated based on the principle of back projection transformation distance from AGV.The experimental results show that the measurement accuracy of the system is basically controlled within 10%and the maximum absolute error within 4 m is 51 mm,which can meet the design requirements of AGV in logistics factory.