视觉检测系统LED光源照度优化方法

针对当前视觉检测系统LED光源照度优化研究中存在的照度效果评价因素单一、照度优化方法通用性不足等问题，以芯片封装质量视觉检测为例，提出一种基于改进樽海鞘算法的LED光源照度优化方法。该方法在单个LED光源照度数学模型基础上，建立标准条形LED阵列光源照度数学模型，获取条形LED阵列在任意空间位姿与被测面的照度值；基于照度均匀度、照度梯度变化与对中度、平均照度、目标与背景区分度等因素建立平面照度效果评价函数；提出改进樽海鞘算法，通过改进算法收敛系数、速度、领导者与追随者位置等更新策略，增强区域搜索的多样性；应用改进樽海鞘算法对平面照度效果评价函数进行优化求解，获取具有最优照度效果的空间位姿参数。实验结果表明:考察优化区域的相对照度分布，文中提出的LED光源照度优化方法所得照度分布与实际测量所得照度分布结果基本一致，目标区域理论照度均匀度在98.78%以上，误差在5.57%以内。因此文中提出方法优化目标合理，可用于视觉检测系统具有最优照度效果时光源位姿信息参数的获取。

Illumination optimization method of LED light source for visual inspection system

To address the issues that the illumination effect evaluation elements were limited and the illumination optimization methods were lack of universality in current research of illumination optimization method of LED light source for visual inspection system, an illumination optimization method of LED light source for visual inspection system based on improved Salp swarm algorithm, taking visual inspection on chip package quality as example in this paper. A mathematical illumination model of single LED light source was established, based on which the mathematical illumination model of standard strip shape LED array was proposed to obtain the illumination values of the LED array in arbitrary spatial position and measured plane. The plane illumination effect evaluation function was constructed in comprehensive consideration of the illumination uniformity, the average illumination, the illumination gradient variation, the threshold reaching standard, the distinguishability of target and background area. An improved salp swarm algorithm (ISSA) was proposed to optimize the computing procedure of plane illumination effect evaluation function and obtain the spatial position parameters of optimal illumination effect. The update strategies of the convergence coefficients, the Salp speed, the leader and its followers’ position were improved, to enhance the diversity of area searching in ISSA. The experimental results show that based on the observation of the relative illumination distribution of the optimize area, the illumination distribution obtained by the propose method is basically consistent with the illumination distribution by actual measurement, the illumination uniformity of the target area is above 98.78%, and the maximum difference is within 5.57%. Therefore, the proposed method had a reasonable optimization objective and could be used to obtain the spatial position parameters of optimal illumination effect for visual inspection system.