Image processing for a line-scan camera system

A PC-based system equipped with a high-speed line scan camera is used to inspect and time moving objects. Due to uncertain and changing lighting conditions of the images captured by the system, the resulting digital image can have very poor contrast, making human interpretation difficult. Because the line scan images are captured one column at a time, we propose a 1-D homomorphic column filter which enhances the image while it is being generated. We also implement this algorithm in conjunction with an automated ends-in search and piecewise linear mapping, Wiener filtering, and sharpening. A memory technique called EMS-mapping is used. The results show our method can improve line scan images within an acceptable time on the currently used computing system, thus improving the system performance.