The adaptive threshold for image segmentation in shape metering of hot-rolling strip steel

An image processing system in laser shape-meter for hot-rolling strip steel is introduced. During production, a grey's variation of image background caused by the strip's temperature fluctuation has an adverse effect on the steady shape measurement online. For this reason, a method to seek an adaptive threshold for image segmentation is presented. According to the adaptive principle, with the help of the online analysis on histogram of the regional images, the adaptive threshold is obtained.