论文部分内容阅读
针对印刷电路板(PCB)光电图像在获取过程中含有噪声和较模糊的边缘等各种原因,提出了一种基于图像融合的含噪声且较模糊PCB光电图像边缘检测方法。首先,讨论了基于小波变换及Canny边缘检测算子的图像边缘检测法,分析了其基本原理。然后,结合两种方法的优点,提出了基于图像融合检测方法的PCB光电图像边缘检测基本原理及中值滤波、增强去模糊、小波分解、图像融合、图像去噪等步骤。最后,将由CCD成像系统及显微镜获取的主要含有高斯噪声且较模糊的PCB光电图像用三种检测方法进行了主观实验对比,采用本文融合方法得到的边缘图像效果最好,在抑制噪声的同时得到了连续的边缘;为了客观地评价PCB光电图像边缘检测的效果,用峰值信噪比及图像边缘信息熵作为评价指标,采用本文方法的峰值信噪比及信息熵的实验结果值都是最大的。
Aiming at various reasons, such as noise and fuzzy edges in printed circuit board (PCB) electro-optical image, a method of edge detection based on image fusion is proposed. Firstly, the image edge detection based on wavelet transform and Canny edge detection operator is discussed, and its basic principle is analyzed. Then, based on the merits of the two methods, this paper proposes the basic principle of PCB photoelectric image edge detection and the steps of median filtering, de-blurring, wavelet decomposition, image fusion and image denoising based on the image fusion detection method. Finally, the subjective experimental comparison of the three kinds of PCB images obtained by CCD imaging system and microscopy, which are mainly Gaussian noise and blurring, is carried out. The edge image obtained by this fusion method is the best, while the noise is suppressed In order to objectively evaluate the effect of PCB edge detection, using the peak signal-to-noise ratio and the edge information entropy as the evaluation index, the experimental results of the peak SNR and the information entropy using this method are the largest .