基于多曝光的焊接图像融合方法  

Welding Image Fusion Method Based on Multi-exposure

作　　者：罗祥[1] 王宗义[1] 

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001

出　　处：《热加工工艺》2016年第7期232-237,共6页Hot Working Technology

摘　　要：在明弧焊接中,由于弧光的强烈干扰,常用的视觉系统难以获得清晰的图像,通过采集多曝光图像然后融合的方法可以解决这个问题。但是现有的多曝光图像融合方法没有针对焊接图像做相应的处理,致使图像中弧光干扰无法得到有效抑制。本文针对明弧焊接光强分布的特点,将图像划分为弧光区域和非弧光区域,使用图像熵评价和阈值处理的方法分别获得两个区域的权重图像,并使用高斯模糊权重图像的方法消除了加权融合时的拼接痕迹。非弧光区域的权重值是固定的从而无需重复计算,只需要对弧光区域进行高斯模糊,这样的处理方式能有效降低计算量。试验结果表明,该方法具有运算速度快的特点,能够获得细节丰富、视觉观感好的焊接融合图像。In the open arc welding process, the common vision system is difficulty to achieve clear images due to the arc interference. To solve this problem, multi-exposure image acquisition and fusion method is proposed. Existing multi-exposure image fusion methods do not make the appropriate process for the welding images, so that the arc interference cannot be suppressed effectively. Aiming at the characteristics of the light intensity distribution of arc welding, the images were divided into arc region and non-arc region. The weighted images of two regions were obtained through entropy evaluation and threshold processing. The synthetic trace was eliminated completely by Gaussian blur method during weighted fusion. Because the weight of the non-arc region was fixed without having to repeat calculation and the Gaussian blur processing only needs to operate with the pixels of the arc region, computation was reduced greatly. The experimental results show that the fusion method has low computational complexity, and can get rich detailed and looked comfortable images.

关 键 词：图像融合 权重图像 多曝光图像 焊接视觉系统 

分 类 号：TG409[金属学及工艺—焊接]