论文部分内容阅读
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 the visual inspection, the laser welding is usually used to identify the image of the weld. The laser beam usually contains a lot of weld characteristic information. Because of the large noise of the image digital information collected by the CMOS image sensor, a compensation and filtering method based on the linear filter is proposed. The red channel of the image sensor is strongly sensitive to laser light, and is compensated by the pixel value of the red pixel R Its peripheral pixel values enhance the characteristics of the laser in the image. Based on the advantages of field programmable gate array (FPGA) parallel processing, an FPGA-driven CMOS image sensor was designed and its digital image information was collected. The image digital information is processed by FPGA in the form of pipeline, and the offset information is calculated through the weld seam feature image. The method reduces the data access operation, thus reducing the processing delay, with high real-time.