CAJ | 학술논문

针对传统基于形态学的边缘检测算法自适应降噪能力差、计算效率低的问题，提出了一种基于边缘比率和多结构的自适应边缘检测算法，根据超声C扫描图像的边缘比率将超声C扫描图像分为3种不同复杂度的图像，针对不同复杂度的超声C扫描图像自适应选取结构元素进行边缘检测，再根据不同边缘图像的信息熵进行融合处理得到最后的边缘。实验结果表明：该算法有效抑制噪声的干扰，且较好地保持图像边缘细节，提高了超声C扫描图像缺陷边缘检测的效率和准确率。
침대전통기우형태학적변연검측산법자괄응강조능력차、계산효솔저적문제，제출료일충기우변연비솔화다결구적자괄응변연검측산법，근거초성C소묘도상적변연비솔장초성C소묘도상분위3충불동복잡도적도상，침대불동복잡도적초성C소묘도상자괄응선취결구원소진행변연검측，재근거불동변연도상적신식적진행융합처리득도최후적변연。실험결과표명：해산법유효억제조성적간우，차교호지보지도상변연세절，제고료초성C소묘도상결함변연검측적효솔화준학솔。
FOr the weak adaptatiOn Of nOise interference and IOw cOmputatiOnaI efficiency Of the edge detectiOn aIgOrithm based On traditiOnaI mOrphOIOgy,the adaptive edge detectiOn aIgOrithm based On muIti-structure and mOrphOIOgy and the ratiO Of edge was prOpOsed. The uItrasOnic C scan image was divided intO three kinds Of different cOmpIexity image accOrding tO the edge IeveI percentage. Then,different structure eIements were adOpted in edge detectiOn based On mOrphOIOgy fOr different cOmpIexity uItrasOnic C scan image. At Iast,the finaI edge was determined accOrding tO the infOrmatiOn entrOpy Of different edge images. ExperimentaI resuIts shOwed that the prOpOsed aIgOrithm nOt OnIy have effectiveIy eIiminated the infIuence Of nOise and better retain the image detaiI,but aIsO imprOved the efficiency and accuracy Of the edge detectiOn Of the defects in uItrasOnic C scan image.