摘要
风机叶片是风力发电系统的核心部件,在受到气候条件、工作负荷等因素的影响后,容易出现各类缺陷,如裂纹、磨损、腐蚀等;如果不能及时发现和解决这些缺陷,将导致风机性能下降、损坏甚至引发安全事故;为此,研究一种基于无人机的风机叶片表面缺陷自动检测方法;利用无人机搭载摄像机,拍摄空中运行的叶片图像;对叶片图像实施灰度化、去噪以及照度均衡化处理,提升图像质量;提取叶片图像中的几何特征和纹理特征,利用差异演化算法改进概率神经网络平滑参数,以优化后的概率神经网络为基础构建分类识别模型,将几何特征和纹理特征作为输入,计算每种类别的输出概率,将最大值响应原则将概率数值最大的类别作为判定的缺陷类别,以此实现风机叶片表面缺陷自动检测;结果表明:所研究技术应用下,杰卡德系数可以达到0.9823,说明该方法的检测结果更为准确;所花费时间低于15.69 s,说明该方法的检测效率更高,可以更快地完成检测任务。
Wind turbine blades are the core components of wind power generation systems,after the influences of climate conditions,working loads,they are prone to various defects such as cracks,wear,corrosion,etc.If these defects can not be detected and resolved in a timely manner,it will lead to a decrease in fan performance,damage,and even safety accidents.To this end,an automatic detection method for surface defects of fan blades based on drones is studied.the drones equipped with cameras are used to capture the images of blades running in the air.The images of blades are implemented grayscale,denoising,and illumination equalization processing to improve image quality and extract geometric and texture features from blade images,differential evolution algorithms are used to improve the smoothing parameters of probabilistic neural networks,present the classification recognition model based on the optimized probabilistic neural network,and calculate the output probability of each category by taking geometric and texture features as inputs.According to the principle of maximum value response principle,the category with the highest probability value is used to determine the defect category,and achieve automatic detection of surface defects on fan blades.The results show that under the application of the studied technology,the Jaccard coefficient can reach up to 0.9823,indicating that this method has more accurate detection results;The running time is less than 15.69 s,indicating that this method has a higher detection efficiency and can effectively achieve detection tasks.
作者
闫浩伟
YAN Haowei(Electrical Engineering and Automation of Shanxi Datang Kelan Wind Power Co.,Ltd.,Xinzhou 034000,China)
出处
《计算机测量与控制》
2024年第11期72-79,共8页
Computer Measurement &Control
关键词
无人机
风机叶片
特征提取
改进概率神经网络
缺陷自动检测技术
drones
fan blades
feature extraction
improved probabilistic neural networks
automatic defect detection technology
作者简介
闫浩伟(1975-),男,大学本科,高级工程师。
