摘要
在视觉检测中,对焊缝图像的识别通常需要使用激光作为辅助工具,一束线型的激光中往往包含着大量的焊缝特征信息。由于CMOS图像传感器采集到的图像数字信息存在较大的噪声,提出一种基于线性滤波器的补偿滤波方法,利用图像传感器的红色通道对激光的强敏感性,用红色像素R的像素值去补偿其周边的像素值,以加强激光在图像中的特征。利用现场可编程逻辑门阵列(Field programmable gate array,FPGA)并行处理的优点,设计了一种FPGA驱动CMOS图像传感器,并采集其输出的图像数字信息的方法。图像数字信息以流水线的形式被FPGA处理,并通过焊缝特征图像计算出偏移信息。该方法减少了数据存取操作,从而减小了处理延时,具有很高的实时性。
In visual inspection, the laser is usually used as an auxiliary tool for the identification of weld images, and a bunch of linear laser often contains a large amount of weld characteristic information. Because the image digital information collected by CMOS image sensor has big noise, a compensation filtering method based on linear filter was proposed. By using the strong sensitivity of the red channel of the image sensor to the laser, the pixel value of the red pixel R was used to compensate its peripheral pixel value to enhance the characteristics of the laser in the image. By using the advantages of parallel processing of field programmable gate array （FPGA）, a method of FPGA driving CMOS image sensor and collecting the image digital information was designed. The image digital information was processed by FPGA in the form of pipelining, and the offset information was calculated through welding seam characteristic image. The method reduces data access operation, thus reducing processing delay and it has high real-time performance.
出处
《热加工工艺》
CSCD
北大核心
2017年第23期216-220,224,共6页
Hot Working Technology
基金
河北省高等学校创新团队领军人才培训计划资助项目(LJRC013)
河北省自然科学基金面上项目(F2016 203249)
国家冷轧板带装备及工艺工程技术研究中心开放课题资助项目(2012006)
