基于改进Canny算子的光纤面板暗影检测结果边缘优化方法

摘要

光纤面板是光电成像器件中不可缺少的元件,面板中暗影缺陷的自动检测对改进生产工艺、提高产品合格率具有重要的指导意义.在实际开发过程中,主要利用高速高分辨率面阵CCD传感器以及VC++可视化程序组成自动检测系统,优选的Canny算子检测边缘还存在缺陷,通过小波域去噪、自适应阈值的实现等改进了Canny算子,对改进的Canny算子检测的边缘进行了边缘连接、闭合曲线的获取、小区域连通、边缘细化等优化处理后,得到了清晰完整的暗影图像,并且降低了暗影缺陷标识的复杂度,该优化方法确实可行.%Optical fiber panel is an indispensable element for electro-optical imaging. It has important guidance significance for improving production process and product quality to detect the shadow automatically. In the actual development process, the automatic detection system is composed by high-speed and high-resolution plane array CCD sensors and VC + + visual program. But the detected edge still has some defects used preferred Canny operator, and the Canny operator is improved by wavelet denoising, implementation of adaptive threshold. Then the edge detected by improved Canny operator is processed by edge connection, closed curve obtainment, small regional connectivity and edge refinement etc. and got more clearly complete shadow image. Furthermore, complexity to mark shadow defects is reduced greatly, and the method is feasible.