基于焊缝视觉跟踪的图像阈值化研究

在焊缝视觉跟踪的图像处理过程中，信息量大，干扰因素多、图像处理速度慢，浪费计算机时。为提高处理速度，增强焊缝跟踪的实时性，结合焊缝图像的特点，提取主要信息，进行阈值化处理，并通过实验分析了人工选取阈值和自动选取阈值的优缺点，即人工选取阈值的方法适合于单幅图像，自动选取阈值的方法适合于多幅图像连续处理。

Study on making threshold on the image processing of vision tracing weld line

On the image processing of vision tracing weld line, there are large quantity of data information with lots of interfere factors,the speed of image processing is very slow,which waste lots of computer running time.In order to improve the speed and enhance the timely accuracy of seam tracing, we pick up the main information and making threshold based on the characteristic of weld line image.On the base of experiment, the merit and demerit of artificial and automatic threshold are analyzed.it is resulted that the artificial threshold is fit for single image processing and the automatic threshold for muti image processing.