Content based threshold adaptation for image processing in industrial application

In this paper the use of feedback information from a higher level of image processing to adapt the parameter of algorithm at a lover level of processing is considered. The idea behind is to achieve the robust and reliable image processing. The automatic determination of optimal threshold value in two different industrial applications: welding process and character recognition for the applications such as car manufacturing, is presented. The optimal threshold value at image segmentation level is estimated according to the desired result of the overall image processing system.