摘要
针对大型风电机组叶轮裂纹焊接缺陷检测精度低、效率低等问题,提出了一种基于粒子小波变异算法的图像检测技术。首先,对原始焊接缺陷图像进行灰度化、滤波和视觉增强处理,有效消除噪声并增强焊接缺陷特征。然后,采用ROI(Region of Interest,ROI)波形分析法提取图像中焊接缺陷的一阶差分序列内的波形特征,进而分析焊接缺陷的特征。将提取的特征作为输入,利用机器学习算法建立图像检测模型。采用粒子小波变异算法对建立的模型进行训练和优化,实现叶轮裂纹焊接缺陷图像的精准检测。试验结果表明,所提方法能够有效消除噪声,显著提高检测焊接缺陷图像的清晰度,并实现高精度、高效率的焊接缺陷检测,为风电机组的安全运行提供技术保障。
In response to the low accuracy and efficiency of defect detection in the welding of large wind turbine impeller cracks,an image detection technology based on the particle wavelet mutation algorithm is proposed.Firstly,the original welding defect images are subjected to graying,filtering,and visual enhancement processing to effectively eliminate noise and enhance the features of welding defects.Then,the ROI(Region of Interest)waveform analysis method is used to extract the waveform features within the first-order difference sequence of the welding defects in the image,and further analyze the characteristics of the welding defects.The extracted features are used as inputs to establish an image detection model using machine learning algorithms.The particle wavelet mutation algorithm is used to train and optimize the established model to achieve precise detection of impeller crack welding defect images.Experimental results show that the proposed method can effectively eliminate noise,significantly improve the clarity of welding defect image detection,and achieve high-precision and efficient welding defect detection,providing technical support for the safe operation of wind turbine units.
作者
商少波
姚博
周云涛
刘田红
SHANG Shaobo;YAO Bo;ZHOU Yuntao;LIU Tianhong(State Grid Jiangsu UHV Company,Nanjing 224300,China;Jiangsu Electric Power Information Technology Co.,Ltd.,Nanjing 210000,China)
出处
《电焊机》
2025年第6期118-124,共7页
Electric Welding Machine
基金
安徽省高校省级自然科学研究项目重点项目(KJ2019A1229)。
关键词
大型风电机组
叶轮裂纹
焊接缺陷
图像检测
粒子小波变异算法
large wind turbine
impeller cracks
welding defects
image inspection
particle wavelet mutation algorithm
