基于全局粒子群算法的车轮定位参数视觉检测系统尺寸优化.docVIP

  基于全局粒子群算法的车轮定位参数视觉 检测系统尺寸优化# 卢雪，徐观，苏建，张立斌，孙丽娜** 5 10 15 20 25 30 35 40 （吉林大学交通学院,长春 130022） 摘要：针对现有汽车车轮定位参数检测系统尺寸较大、结构复杂的问题，提出了基于双目立 体视觉的进行车轮定位参数检测方法。研究了小结构尺寸、大基线距的车轮定位参数双目视 觉测量系统模型，构建了以系统水平面积最小为目标的最优目标函数，提出了采用粒子群算 法对该目标函数求解的方法。通过仿真分析得出了以系统尺寸为约束条件的系统结构的最优 解，仿真结果表明：当虚拟基线距为 1400 mm 时，系统的最优结构参数为基线距等于 600 mm， 两小镜子之间的距离等于 120 mm，小镜子到摄像机的距离等于 600 mm，大镜子到摄像机 的距离等于 700 mm，小镜子与基准面的夹角等于 15°，大镜子与基准面的夹角等于 30°，大 镜子的长度等于 500 mm，系统面积为 1.03 m2。 关键词：车轮定位；双目视觉；全局粒子群算法 中图分类号：U467.5 Size optimization for vision-based test system of vehicle wheel alignment parameters based on global particle swarm method LU Xue, XU Guan, SU Jian, ZHANG Libin, SUN Lina (College of Transportation,Jilin University,Changchun 130022) Abstract: Focus on the problem of large size and complex structure of current detection system for automotive wheel alignment parameters ,a method to detect wheel alignment parameters based on binocular stereo vision is proposed. The binocular vision measurement system model with small size and large baseline for wheel alignment parameters is studied, an optimal objective function to minimize the system horizontal area is constructed, a method using particle swarm algorithm for solving the objective function is proposed. The optimal solution for system structure with system size as constraint condition is calculated through simulation analysis. Simulation results show that when the virtual baseline is 1400 mm, the optimal parameters of the system structure are as follows, the baseline distance is 600 mm, the distance between the two smaller mirrors is 120 mm, the distance from a smaller mirror to the camera is 600 mm, the distance from a larger mirror to the camera is 700 mm, the angle between the smaller mirror and the benchmark plane is 15 °, the angle between the large mirror and the benchmark plane is 30 °, larger mirror is 500 mm long, the system area is 1.03 m2. Key words: wheel alignment; binocular vision; global particle swarm 0 引言 汽车在长期使用过程中，由于悬架系统磨损、路面冲击、交通事故等，车轮