基于小波能量熵与二维Zernike矩的IC芯片表面微观不平度图像检测技术研究

Study of micro-plainness detecting technology based on IC chip surface wavelet energy entropy and two-scale zernike moment

在IC芯片制造检测过程中,需要完整而准确地对芯片表面形貌进行检测。针对IC芯片表面形貌检测的要求,提出了基于小波能量熵与二维Zernike矩的芯片表面微观不平度图像检测技术,并运用于高精度要求的计算机芯片图像处理工作中。通过对获得的芯片显微图像进行小波变换并求取信息熵图像的基础上,建立二维Zernike矩与采用BP神经网络进行芯片表面不平度的识别与评定。编程实验证明,采用本方法可准确对IC芯片表面微观不平度进行等级评定。

In IC chip accurate optical inspecting, it is necessary to gain detect chip＇s micro-surface feature completely and accurately. For this purpose ,a new micro-plainness detecting technology based on IC chip surface wavelet energy entropy and two-scale Zernike moment is investigated to detect IC surface＇ s plainness. By wavelet transformation and energy entropy of image topography image ,two-scale Zernike moment is built up and BP neural network is used to calibrate surface＇s micro-plainness. Experimental result indicates that accurate scaling result of IC chip＇s surface can be detected by this method.