摘要
针对原棉破籽类杂质视觉检测中OTSU方法最佳阈值难以确定的问题,提出基于粒子群优化最佳阈值的杂质检测方法。首先对图像进行灰度变换,去除图像中冗余信息与噪声数据,然后将类间方差设置为目标函数模型,通过动态调整惯性权重因子与自适应动态调整学习因子进行优化,迭代整个群体的全局最优位置和粒子自身的历史最优位置,进而求解最大类间方差,获取最优图像分割阈值。实验结果表明:与OTSU方法相比,改进OTSU方法的杂质检测相对误差平均降低2.74%,检测到的杂质边缘信息能够有效表征杂质像素属性,杂质边缘描述清晰明确,适用于原棉破籽类杂质快速检测。
In view of the difficulty in determining the optimal threshold of OTSU method in visual detection of raw cotton seed broken impurities,an impurity detection method based on particle swarm optimization was proposed.Firstly,the image is transformed into gray scale to remove redundant information and noise data.Then,the inter-class variance was set as the objective function model,and the optimization was carried out by dynamically adjusting the inertia weight factor and the adaptive dynamic adjustment learning factor.The global optimal position of the whole population and the historical optimal position of the particle itself were iterated,and the maximum inter-class variance was solved to obtain the optimal image segmentation threshold.The experimental results show that compared with OTSU method,the relative error of the improved OTSU method is reduced by 2.74%on average.The detected impurity edge information can effectively characterize the impurity pixel attributes,and the description of impurity edge is clear,which is suitable for the rapid detection of raw cotton broken seed impurities.
作者
齐英兰
QI Yinglan(Department of Information Engineering,Henan Vocational College of Light Industry,Zhengzhou,Henan 450002,China)
出处
《毛纺科技》
CAS
北大核心
2022年第6期90-94,共5页
Wool Textile Journal
基金
国家重点研发计划项目(2016YFF0201902)。
关键词
原棉
杂质
破籽
OTSU
视觉检测
raw cotton
impurities
broken seeds
OTSU
vision inspection
作者简介
齐英兰,副教授,硕士,主要研究方向为智能信息处理和计算机网络,E-mail:gdxkln00@126.com。
