拉链齿数的数字图像处理研究

传统拉链齿数检测采用人工来完成,存在工作量大、效率低、准确率低的缺点。提出了一种新的拉链齿数检测方法。该方法采用中值滤波去除图像噪声,通过最大类间方差法获取阈值,然后进行图像二值化,并对处理后的图像采用质心间距法计算每一个相邻拉链齿的距离,并与标准间距比较判断,依据判断结果来完成拉链齿数的检测。实验结果表明,这种方法能快速准确地检测拉链的多齿、少齿缺陷,具有很强的适应性。

Study of Zipper Teeth Detection Using Image Processing Technology

The traditional detected method of the number of zipper teeth is implemented by man,and is heavy-worked,inefficient and low-accurate.This paper puts forward a new kind of detected method of zipper teeth.In this scheme,it adopt median filtering to remove image noise,through the Otsu to obtain the threshold value,then the binary image.After the processed image,the centroid distance method is adopted to identify and judge.In MatlabR2007a simulation experiments,the results show that the proposed scheme can accurately detect tooth-missing drawback in the production process,and have a strong adaptability.