基于模糊推理的V形焊缝图像阈值的自动选取

Automatic selection for the threshold of arc-v-shape weld based on fuzzy reasoning

介绍了机器人焊接V形焊缝图像处理中基于模糊推理自动获取局部阈值的算法。采用9种不同形状的窗口模板,以列扫描方式求取这些窗口所覆盖区域内的最大平均灰度值,然后以获取的最大平均灰度值以及相邻区域最大平均灰度值的改变量作为模糊推理器的两维输入,经过推理获得这些窗口所覆盖区域的阈值。通过变换最大平均灰度值这一论域中各个语言变量灰度值的取值范围,对以不同的激光器作为辅助光源而获取的焊缝图像进行了处理试验。

An arithmetic of threshold selected automatically based on fuzzy reasoning for image processing of arc-v-shape weld by robot is introduced.Designing nine kinds of window templates,and getting the maximum average gray level value in the region where these templates can cover through scanning by rows,using the maximum average gray level value and it＇s change quantity as the two inputs of fuzzy controller,and getting the local threshold where these templates can cover through reasoning.At last,the phrase of the maximum average gray level value includes seven language variables,and every variable accords with different gray level value range,it can obtain thresholds of different kind of weld images that use different laser as assistant lamp-source,and making experiment on these images adopting the way above.