应用于PCB缺陷检测系统的图像质量评价法

Image Quality Assessment Method Applied to PCB Defect Detection System

基于机器视觉的PCB缺陷检测系统易于因图像模糊等失真问题,影响后续缺陷检测的准确性。针对此问题建立了相关模糊图像数据集,并提出一种空频结合的无参考图像质量实时评价方法。方法提取图像的最大局部变化信息作为空域信息。随后对空域信息进行一级小波分解,求出高频分量中水平方向的小波系数,随后利用处理后的小波系数的最大值,得到客观评价值。同时,建立PCB模糊图像数据集(PBID)用于算法的验证。大量实验结果表明,与其它无参考图像质量评价方法相比,该方法与主观评价值具有较高的一致性,在PBID数据集上的皮尔逊线性相关系数(PLCC)达到了0.9835,且运行速度快,仅为每帧0.1302 s,适合对实时性要求较高的应用场合。

The accuracy of PCB defect detection system based on machine vision is easily affected by image blur and other distortion problems.Aiming at this problem,this paper proposes a no-reference image quality assessment algorithm based on time-frequency combination.The image information in spatial domain is extracted by maximum local variation.Afterwards,the horizontal wavelet coefficients in high-frequency components are obtained by one-scale wavelet.Meanwhile,the objective evaluation value is computed by the maximum value of processed wavelet coefficients.We also established a PCB blurred image database(PBID).A large number of experimental results show that,compared with other no-reference image quality assessment methods,the proposed algorithm has higher correlation with subjective evaluation value.The Pearson linear correlation coefficient(PLCC)on PBID database is 0.9835.The run-time of proposed method is only 0.1302 s per frame.The proposed method has the advantage of fast running speed,so it is suitable for applications with high real-time requirements.