一种基于频域变换的熔池运动模糊图像恢复算法  

A weld pool blur image restoration algorithm based on frequency domain transform motion

作　　者：汪衍广 马国红[1] 姜芙蓉[1] 叶佳[2] 

机构地区：[1]南昌大学机电工程学院,330031 [2]肯塔基大学电子显微镜中心,美国列克星顿市40506

出　　处：《焊接》2015年第11期27-29,68,共4页Welding & Joining

基　　金：国家自然科学基金项目(61165008);教育部回国基金项目(13006199)

摘　　要：针对熔池运动模糊图像的特点,提出了一种基于频域变换的图像恢复算法。将图像变换到频域下,分别使用Radon变换和微分自相关估算熔池图像的模糊角度和模糊长度,利用该结果再使用维纳滤波对熔池图像进行恢复处理。通过焊接图像采集系统获取到熔池运动模糊图像,对其进行恢复处理试验。图像处理结果及客观评价参数分析表明,该算法能够有效地恢复熔池图像,获得清晰的边缘信息。According to the characteristics of the weld pool motion blurred images,a image restoration algorithm was proposed based on frequency domain transform.The image was transformed to the frequency domain,the fuzzy angle and length of weld pool images were estimated by using the Radon transformation and differential autocorrelation,respectively.Based on the results,the wiener filtering was used to recover the weld pool images,which was verified by experiment.Image processing result and objective evaluation parameter analyses show that the algorithm can effectively restore the weld pool image and get a clear edge information.

关 键 词：熔池图像 运动模糊恢复 PSF 维纳滤波 

分 类 号：TG407[金属学及工艺—焊接]