Abstract
Since machine vision systems (MVS) lead to a wide usage of monitoring systems for industrial applications, the research on the statistical process control (SPC) of image data has been promoted as an automated method for early detection and prevention of unusual conditions in manufacturing processes. In this paper, we propose a non-parametric SPC approach based on the 2D wavelet spectrum (WS-SPC) to extract the feature that contains the spatial and directional information of each subspace in an image. Using the 2D discrete wavelet transform and spectrum analysis, the representative statistic, the Hurst index, is calculated, and a single matrix space that consists of estimated statistics is reconstructed into a spatial control area for SPC. When a control limit is determined by the density of statistics, real-time monitoring based on WS-SPC is available for time releasing images. In the application, an analysis of wafer bin maps (WBMs) is conducted at a semiconductor company in Korea in order to evaluate the performance of the suggested approach. The results show that the proposed method is effective in terms of its fast computation speed and spectral monitoring.
Highlights
Statistical process control (SPC) is one of the monitoring schemes that can be applied to fault detection in manufacturing processes
The basic SPC procedure consists of two phases: the calculation of the statistics to represent the objects’ status from multivariate variables and the determination of the control limits that define an abnormal status of the objects
After wavelet coefficients are obtained by 2D discrete wavelet transform (DWT), the Hurst matrices from three directions are generated from spectrum analysis, where the distance matrix is calculated from the Hurst features to represent the status of processes as control statistics
Summary
Statistical process control (SPC) is one of the monitoring schemes that can be applied to fault detection in manufacturing processes. Though image based multivariate SPC approaches are widely applied to distinguish if the will have a defective status or not, using a representative value, there are few literature works providing process will have a defective status or not, using a representative value, there are few literature works the spatial information of faulty regions; e.g., the defect clustering area in the wafer bin map (WBM). Providing the spatial information of faulty regions; e.g., the defect clustering area in the wafer bin resulting from the semiconductor manufacturing process. The majority of processing methods in SPC to improve the detection accuracy of abnormal status from (spatial) image studies assume normality or a pre-specified distribution a priori to calculate the statistics designating data. The proposed method is applied to real WBM image data to prove its potential in fault detection or process monitoring.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have