CAJ | 학술논문

图像分割是实现热波检测缺陷定量识别的关键。根据有缺陷区域与无缺陷区域图像灰度值不连续变化的特点，提出一种基于尖点突变理论的分割方法，在计算图像幅值和相位的基础上，通过建立尖点突变模型及相应的坐标变换，将系统变量进行归一化处理，根据处理后的尖点突变模型的判决式标记热波图像中像素值突变的点，提取突变点的骨架，对其膨胀和填充运算实现缺陷的分割。最后通过试验验证了该方法的可行性。
도상분할시실현열파검측결함정량식별적관건。근거유결함구역여무결함구역도상회도치불련속변화적특점，제출일충기우첨점돌변이론적분할방법，재계산도상폭치화상위적기출상，통과건립첨점돌변모형급상응적좌표변환，장계통변량진행귀일화처리，근거처리후적첨점돌변모형적판결식표기열파도상중상소치돌변적점，제취돌변점적골가，대기팽창화전충운산실현결함적분할。최후통과시험험증료해방법적가행성。
Image segmentation is the key technique to achieve the quantitative identification in the thermal waving inspection. An image segmentation method was advanced based on the cusp catastrophic theory according to the characteristics of discontinuous change between the region with defect and without defect in thermal images. The amplitude and phase of the image were computed at first, then the model of the cusp catastrophic was established and the corresponding transform of coordinate was studied, the variable of the system was dealt with unitarity. The salutatory dots were signed based on the judicative formula of the cusp catastrophic model after the unitariness, the framework of the salutatory dots was picked up to realize the segmentation of the thermal waving images through the operation of the expansion and fill. At last the experiments were carried out to prove the feasibility of this method.