【摘要】 为了提高图像分析法测量金刚石颗粒粒径的精度,简化测量过程中金刚石图像采集的步骤,进行了相连金刚石颗粒图像分割的研究。以最小外接圆法评定了金刚石颗粒投影的圆度。在对采集到的金刚石颗粒图像二值化的基础上,首先提取金刚石颗粒图像的轮廓,获取轮廓坐标序列并采用五点平均法对轮廓坐标序列进行了平滑;基于金刚石投影的凸性,沿轮廓逐点搜索出分割点(凹点)。结合金刚石颗粒投影圆度系数的特征,进行了分割算法的设计,解决了分割点的配对问题。算法考虑了金刚石枝状相连、环状相连及枝状-环状相连的情况,提高了算法的实用性。分割实例表明,算法能有效分割相连的金刚石颗粒图像,准确性较高。更多还原

【Abstract】 In order to increase the measurement accuracy of size of diamond grits and to simplify the acquiring procedure of diamond images,partition of image of attached diamond grits is presented.The roundness of grit’ s projection was assessed by minimum circumcircle method.Based on the binarization of acquired image of diamond grits,the contour of projective region was extracted and afterwards smoothed by averaging the coordinate values of 5 neighboring points.Considering the convex of diamond grit’s projection,partition points (concave points)were detected along the contour.Taking use of the characteristic of the roundness of diamond projection,partitioning algorithm was designed to solve the matching problem of partition points,which takes branch-shaped attachment,ring-shaped attachment and branch-ring-shaped attachment into account. Experimental results show that this algorithm can partition the image of attached diamond grits effectively.更多还原