基于图像处理的玻璃瓶口裂纹检测系统设计与实现

Design and Implementation of Glass Bottle CrackDetection System Based on Image Processing

摘要:

为了快速、准确地对玻璃瓶口裂纹进行无接触检测，基于图像处理技术设计与开发了一个玻璃瓶口裂纹检测系统。首先对获取的玻璃瓶口图像进行降噪与分割处理，并利用重心法获取玻璃瓶口圆心的粗定位，然后利用边界跟踪算法对玻璃瓶口进行边界跟踪，并通过 最小二乘拟合获取玻璃瓶口圆心的精确位置，最后利用深度优先遍历算法对裂纹点集进行聚群，从而实现对玻璃瓶口裂纹的检测。实验结果表明，该系统操作简单、效率高、抗干扰能力强，能满足实际使用的要求。

Abstract:

In order to quickly and accurately realize the non-contact detection of glass bottle crack, a glass bottle crack detection system based on image processing technology is designed and implemented. Firstly, the glass bottle mouth image is denoised and segmented, and then the centroid method is used to obtain the rough location of the center of the glass bottle. Secondly, the least squares fitting method is used to get the precise location of the center of the glass bottle through tracking the boundary of the glass bottle based on the boundary tracking algorithm. Finally, the depth first traversal algorithm is used to crack point set of clusters, so as to realize the detection of glass bottle crack.