基于图像饱和度分量Zernike矩的元件封装识别

针对电子元件的自动检测识别问题,给出一种图像饱和度分量的Zernike矩识别电子元件的新方法。根据电子元件的图像颜色特征,分层提取原始图像的饱和度分量,通过遗传算法对饱和度图像分量进行二值化分割并计算改进的Zernike矩作为元件图像的识别特征,利用切比雪夫与曼哈顿距离加权线性来取代传统的欧氏距离判断相似性,降低相似性判定时的运算时间,从而达到识别元器件外观的目的。实验表明,该方法能够准确地识别待测的元器件封装,可应用于电子元件生产和检测中。

Component Packaging Recognition Based on Zernike Moment of Image Saturation

With regard to automatic detection and recognition of electronic components, a new Zernike moment recognition method for electronic components based on the saturation of image is presented in this paper. According to the image color characteristics of electronic components, the saturation components of the original image are extracted hierarchically. The saturation image components are binarized by genetic algorithm and the improved Zernike moments are computed as the recognition features of the component image. Then, Chebyshev and Manhattan distance weighted linearity are used to replace the traditional Euclidean distance to judge similarity and reduce the operation time of similarity determination, so as to achieve the purpose of identifying the appearance of components. Experiments show that this method can accurately identify the components package to be tested, and can be applied to the production and detection of electronic components.