作　　者： ; ;

机构地区： 南京航空航天大学计算机科学与技术学院计算机科学与技术系

出　　处： 《南京航空航天大学学报》 2000年第6期691-695,共5页

摘　　要： 提出了一种通过图像处理技术设计和实现的自动报靶系统。针对靶图像背景颜色对比度大的特点 ,采用一种基于灰度期望值的图像二值化算法 ,通过逐步分离的方法 ,将靶环从背景中分割出来 ;针对影响靶数精度的主要因素 (图像的非线性畸变 ) ,将靶数分解为整数与小数两部分。整数部分根据靶图像的物理意义确定 ,对靶数精度不会造成任何损失 ;小数部分根据弹孔最内侧点在两条靶环之间的相对位置确定。试验结果表明 ,用该系统确定的靶数可以满足一般射击比赛的精度要求。 A method based on image processing to design and realize the auto counting system in shooting competition is presented. Because of the distinct difference in the background color, we use a binary algorithm based on gray mean value and a two step segmentation method to segment the image. In order to reduce the influence of unlinear change on the image, we divide the shooting score into two parts: integral and floating. The integral part is calculated by the physical meaning of image, so its accuracy is high. The floating part is calculated by the relative location between two circles, but there is some inaccuracy inevitably because of the unlinear change on image. According to the experimental results, the accuracy of the system can satisfy the requirement of the normal shooting competition. And it will be used in the real application.

关 键 词： 图像处理 图像分割 图像二值化

领　　域： [自动化与计算机技术] [自动化与计算机技术]