CFRP超声检测半波高度四邻域缺陷量化方法  

Half-wave height four-neighborhood defect quantification algorithm based on CFRP ultrasonic detection

作　　者：王海军 王涛[2] 俞慈君[2] WANG Haijun;WANG Tao;YU Cijun(College of Engineers,Zhejiang University,Hangzhou 310058,China;School of Mechanical Engineering,Zhejiang University,Hangzhou 310058,China)

机构地区：[1]浙江大学工程师学院,杭州310058 [2]浙江大学机械工程学院,杭州310058

出　　处：《振动与冲击》2024年第19期319-325,共7页Journal of Vibration and Shock

基　　金：自然科学基金重点项目资助(91748204);国家自然科学基金创新研究群体科学基金项(51821093);浙江省重点研发计划(2020C01039)。

摘　　要：复合材料缺陷面积准确量化是缺陷自动识别、自动分析的重要环节,对研究复材制造工艺性能具有指导作用。该研究基于碳纤维增强塑料(carbon fiber reinforced plastic,CFRP)的超声相控阵无损检测数据开展研究,分析比较了迭代阈值法、大津法两种图像分割方法对缺陷面积量化结果,结合超声A扫描信号与C扫描图像,提出一种半波高度四邻域缺陷量化算法,定位连通缺陷区域的超声信号最大幅值点,以最大幅值点处一半幅值为临界判断,扩展四邻域缺陷边界搜索,确定缺陷边界。与图像分割方法相比,精度有较大提高,针对较小缺陷,量化的缺陷面积与实际缺陷面积的偏差为13.2%,其余缺陷面积量化偏差均小于2%。Accurately quantifying defect area of composite materials is an important link of defect automatic identification and automatic analysis,and it plays a guiding role in studying manufacturing process performance of composite materials.Here,based on ultrasonic phased array non-destructive testing data of carbon fiber reinforced plastic(CFRP),2 image segmentation methods of iterative threshold method and Otsu method on quantification results of defect area were analyzed contrastively.Combining ultrasonic A-scan signals and C-scan images,a half-wave height four-neighborhood defect quantification algorithm was proposed to locate the maximum amplitude point of ultrasonic signals in connected defect area.Half of amplitude of the maximum amplitude point was taken as the critical judgment,four-neighborhood defect boundary search was extended to determine defect boundary.It was shown that compared with image segmentation methods,the accuracy of the proposed method is more largely improved;for smaller defects,the deviation between quantified defect area and actual defect area is 13.2%,while quantization deviations of rest defect areas all are less than 2%.

关 键 词：复合材料 缺陷量化 图像分割 半波高度法 

分 类 号：TH212[机械工程—机械制造及自动化] TH213.3