CAJ | 학술논문

枪弹外观缺陷图像分割是枪弹质量检测的关键环节,为了实现枪弹外观缺陷自动准确分割,结合Hough变换和双峰法,提出了一种新的分割方法.该方法针对枪弹外观图像中间偏亮两侧偏暗的灰度分布特性,提取光照不均特征表现明显的G颜色分量图,利用改进的快速Hough直线检测法确定预分割缺陷区域;针对预分割缺陷区域的直方图特性,改进双峰法,分割缺陷,对分割结果利用数学形态学进行优化.实验表明,该方法能够实现快速准确的枪弹外观缺陷提取.
창탄외관결함도상분할시창탄질량검측적관건배절,위료실현창탄외관결함자동준학분할,결합Hough변환화쌍봉법,제출료일충신적분할방법.해방법침대창탄외관도상중간편량량측편암적회도분포특성,제취광조불균특정표현명현적G안색분량도,이용개진적쾌속Hough직선검측법학정예분할결함구역;침대예분할결함구역적직방도특성,개진쌍봉법,분할결함,대분할결과이용수학형태학진행우화.실험표명,해방법능구실현쾌속준학적창탄외관결함제취.
The image segmentation of bullet surface defect is an important step in checking the quality of bullet. In order to make sure the bullet surface detect is segmented accurately and automatically,a new method of image segmentation is proposed based on Hough transformation and two-peak algorithm. Because of the image of bullet surface is brighter in the center and darker at both sides,the gradation would be distributed asymmetrically,especially in the image of G component. According to the properties of bullet image,an improved HT algorithm is proposed to make certain the region of interest,which contains the property of the detect. In order to segment the detect and optimize the result,the two-peak algorithm based on the unique gradation histogram is improved,and the mathematicl morphology is exerted. The experiment result shows that this method makes it possible to extract the detect quickly and effectively.