【摘要】 提出了一种采用模糊C均值聚类的方法来识别焊缝超声检测缺陷类型的方法.以夹渣、气孔、未焊透、未熔合4种焊缝常见缺陷为对象,对25个实际焊缝样本进行识别计算,并与射线检测结果进行比对.结果表明:该方法正确识别率达88%,具有较好的识别效果.更多还原

【Abstract】 This paper presents a fuzzy C-Mean clustering algorithm to recognize the type of a weld defect pattern in ultrasonic test.By the method,twenty-five weld samples with four types of welding flaws were recognized and the results were compared with those of the X-Ray test.Experimental results show that the method is feasible and effective.Its recognized accuracy can reach 88%.更多还原