摘要
针对LED显示屏以模组为最小单位进行拼接而出现模组发生空间畸变的问题,提出了基于CCD相机采集并利用图像处理技术定位检测LED显示屏模组畸变度的方法。首先,根据LED像素在采集图像上的光型分布,利用canny边缘检测方法结合闭合填充算法将LED点阵像素分割;然后,利用被分割出的空间区域信息和灰度信息,采用加权质心法对每个LED像素进行亚像素级的加权定位。接着,以模组拼接的LED显示屏为例,构建了模组畸变模型,并根据畸变模组和理想模组的线性关系利用最小二乘法进行拟合;最后,利用梯度下降法求得每个模组的畸变参数。实验结果表明:采用这种方法测量多种模组LED点间距的精度为0.5mm,模组畸变特征参数的测量过程中,角度测量精度为1.3°,位移测量精度为2.2mm,基本满足自动影像测量的稳定可靠、精度高、抗干扰能力强等要求,并可在户外机械检测无法测量的情况下应用。
In order to measure the position distortion arouse from using the LED display modules as the smallest units to splice together, a method based on CCD capture with image processing has been proposed. First, apply canny edge detection operators to segment each LED pixels from the background picture according to their facular grayscale distribution captured in the image. Second, use the weighted centroid location method to compute each LED pixel＇s sub-pixel level position according to their region been split out from the position information and the gray information of the image. Then, taking a LED display module which is spliced together for example, the module position distortion model has been constructed. After that, comparing the distortion model with the ideal model and applying their linear relationship, a calculation based on the method of least square fitting has been given out. Finally, use gradient descent algorithm to obtain each modules distortion parameters. Experimental results show that using this method to measure the position distortion parameters of LED modules, LED pixels＇ spacing accuracy is 0.5mm;In the process to measure each modules distortion characteristic parameters, the angle measurement accuracy is 1.3~ and the displacement measurement accuracy is 2.2mm. It can satisfy the system requirements of non-contact, higher precision and rapid speed, as well as strong anti-jamming and stabilization. It can be used in case of mechanical testing can not be achieved in the outdoor.
出处
《现代显示》
2012年第9期233-239,共7页
Advanced Display
关键词
畸变
边缘检测
质心
最小二乘
distortion
edge detection
centroid
least squares
