灰色关联融合霍夫变换的碳纤维板材边缘检测

Detection of carbon fiber plate edge based on grey relational fusion Hough transform

风力发电是能源结构中的重要组成部分,碳纤维板材是制造高性能风力发电叶片的关键复合材料。由于碳纤维板材边缘没有脱模布包覆,容易出现缺陷。为保证碳纤维板材的质量,需要在生产检测环节严格执行质量检测标准。机器视觉在质量检测方面具有实时且高效的特点,能够满足生产线实时检测的需要。碳纤维板材边缘检测算法是以工业控制计算机为算力平台,对碳纤维板材边缘实现机器视觉检测。利用灰色关联融合霍夫变换对板材边缘贴合距离与脱模布贴合精确率进行检测。在进行碳纤维板材边缘检测时,首先预处理图像,将降噪算法与模糊边缘关联度融合在灰色关联中降低处理数据量;其次,利用灰色关联自适应滤波区分边缘图层与无效图层,对边缘图层的有效数据进行筛选;最后,将有效边缘图层灰色关联度数据融合霍夫变换进行边缘参数空间运算,实现碳纤维板材边缘检测。试验数据验证表明,在每秒检测帧数、平均误差与检测精确率上符合碳纤维板材质量检测要求。

Wind power generation is an important part of energy structure,and carbon fiber sheet is the key composite material for manufacturing high-performance wind power blades.The edge of carbon fiber sheet is easy to appear defects because it is not covered with release cloth.In order to ensure the quality of carbon fiber sheet,it is necessary to strictly implement quality inspection standards in production inspection.Machine vision has the characteristics of real-time and high efficiency in quality inspection,which can meet the needs of real-time inspection in production line.The edge detection algorithm of carbon fiber sheet is based on industrial control computer,which realizes machine vision detection of carbon fiber sheet edge.Grey relational fusion Hough transform was used to detect the bonding distance between the edge of the plate and the bonding accuracy of the release cloth.In the process of edge detection of carbon fiber sheet,the image was preprocessed firstly,and the noise reduction algorithm and fuzzy edge correlation degree were combined in gray correlation to reduce the amount of data to be processed.Secondly,the edge layer and invalid layer were distinguished by gray correlation adaptive filtering,and the effective data of the edge layer were filtered.Finally,the effective edge layer gray correlation degree data was combined with Hough transform for edge parameter space operation to realize edge detection of carbon fiber sheet.The experimental data show that the number of frames per second,average error and detection accuracy meet the requirements of carbon fiber plate quality inspection.