CAJ | 학술논문

利用数字图像处理技术，通过轨道板图像自动分割、边缘提取、轨道板固有结构边缘去除、真伪裂纹区分等步骤，设计了轨道板裂纹缺陷自动检测识别算法。通过对实际采集的轨道板图像的处理，证明该方法能够准确提取和标记缺陷区域。在实验的1000幅图像中，人工检出裂纹图像6幅，自动识别裂纹图像6幅，准确率达到100%，算法运行速度经测试能够满足实时检测的需求。
이용수자도상처리기술，통과궤도판도상자동분할、변연제취、궤도판고유결구변연거제、진위렬문구분등보취，설계료궤도판렬문결함자동검측식별산법。통과대실제채집적궤도판도상적처리，증명해방법능구준학제취화표기결함구역。재실험적1000폭도상중，인공검출렬문도상6폭，자동식별렬문도상6폭，준학솔체도100%，산법운행속도경측시능구만족실시검측적수구。
The algorithm for automatic crack detection and identification in slab track images is designed based on digital image processing technology and many procedures such as image auto-segmentation, edge extrac-tion, edge removing of slab track structure and false crack identification. The algorithm has been proved accurately in extracting and marking the crack region by applying to slab track images from actual acquisition. In the experi-ment of 1000 images, all the six manually detected crack images are also detected automatically by the algorithm, i. e., the accuracy rate is 100%. The computational speed of the algorithm can satisfy the real-time detection require-ments.