为避免船桥发生碰撞事故，设计基于机器视觉技术的船桥防碰撞自动控制系统。利用双目视觉传感器采集船舶航行环境图像，由A/D转换芯片预处理环境图像，再利用双目视觉测距技术结合预处理的环境图像计算船舶与桥墩间的距离；通过全球定位系统采集船舶位置信息，利用改进差分进化PID控制方法得到舵角自动控制结果，再利用异步收发传输器根据控制结果生成自动控制指令，作用于船舵，完成船桥防碰撞自动控制。实验证明：该系统可有效采集船舶航行环境图像，计算船舶三维坐标；在不同航行环境时，该系统均可完成船桥防碰撞自动控制，且自动控制的稳定性较优。
In order to avoid collision accidents between ships and bridges, an automatic anti-collision control system for ships and bridges based on machine vision technology is designed. The ship navigation environment image is collected by binocular vision sensor, the environment image is preprocessed by A/D conversion chip, and the distance between the ship and the pier is calculated by binocular vision ranging technology combined with the preprocessed environment image. The ship position information is collected by the global positioning system, the rudder angle automatic control result is obtained by the improved differential evolution PID control method, and the automatic control command is generated by the asynchronous transmitter and receiver according to the control result, which acts on the ship rudder and completes the automatic control of bridge anti-collision. The experiment shows that the system can effectively collect the ship navigation environment image and calculate the ship three-dimensional coordinates. In different navigation environments, the system can complete the automatic control of bridge anti-collision, and the stability of the automatic control is better.
2023,45(5): 144-147 收稿日期：2022-10-24
DOI：10.3404/j.issn.1672-7649.2023.05.027
分类号：TP391
基金项目：辽宁省教育厅科学研究经费项目（LJKX202015）
作者简介：杨明(1981-),女,硕士,副教授,主要研究方向为机器视觉、智能技术、嵌入式系统开发与设计及模糊算法等
