Research On Defect Detection Of Ceramic Substrate Image Based On Machine Vision

With the thermal printing technology being applied to more and more fields, the demand for thermal printing head has been increasing. The surface quality detection of ceramic substrate which is the core component of the thermal printing head is completely depended on manual detection methods in the production process. The manual detection methods are not only inefficient, but also cannot guarantee the accuracy of detection. In this case, developing an automatic detection system based on machine vision becomes an inevitable choice to improve the current technology of detecting surface defect. In this paper, we discuss automatic defect detection based on machine vision for conductive pattern on ceramic substrate.According to the characteristics and detection status of conductive pattern on ceramic substrate, we first collect images based on linear array CCD camera to ensure the image acquisition being efficient and clear. And then we present the preprocessing methods, template matching methods and image cutting methods etc. for conductive pattern on ceramic substrate. Based on this, we discuss the types of defect detection for conductive pattern on ceramic substrate. Our work in this paper includes:(1) we design a number of image processing methods through analyzing the images of conductive pattern collected by camera. According to the experimental results, we determine to use the median filter and Laplace operator for filtering image and sharpening image, respectively. (2) For the problem of images from linear array CCD camera being too big to process, we design an image cutting method based on template matching, and discuss the basic principles and practical applications of square difference matching method and correlation matching method. According to the experimental results, we determine to use the square difference matching method for template matching of conductive pattern on ceramic substrate. (3) The defects are analyzed based on specific characteristics of conductive pattern on ceramic substrate and common defect types. According to the analysis results, we determine to use the defect detection method based on template comparison. The difference images are filtered by means of the open operators of morphology to get more accurate results in defect detection. Finally, we determine and mark defects by using contour extracting method, and output the detection results.In this paper, we discuss automatic defect detection methods for conductive pattern on ceramic substrate based on digital image processing techniques. The experimental results show that our work can effectively accomplish automatic defect detection for ceramic substrate image. The algorithm is simple and feasible and can meet the requirements of automatic surface detection. This work has significance in theory and in practice for the development of automatic surface detection technology of ceramic substrate image.