Cortex-A9平台钢筋焊缝光条中心提取实现

Extraction of Cortex-A9 platform reinforced welded light bar center

在钢筋焊缝三维形貌的现场检测中,由于传统检测仪不便移动与笨重,则利用Cortex-A9处理器构成一种便携式检测系统进行三维视觉检测。其中提取焊缝的光条中心非常重要,因此将实现在嵌入式系统下光条中心的提取,首先由处理器控制cmos摄像头实时采集获得的图像进行平滑滤波并用大津法得到动态阈值进行图像分割,然后利用灰度平方加权阈值重心法得到焊缝光条纹的亚像素精度中心。实验结果表明该算法抗干扰能力强,在开发板平台上提取原始图像的光条中心用时0. 15 s,可实现焊缝光条中心的精确提取。

In the on-site inspection of the three-dimensional shape of the steel weld,the traditional detector is inconvenient to move and cumbersome,and the Cortex-A9 processor is used to construct a portable detection system for three-dimensional visual inspection. The center of the strip of the weld is very important,so the extraction of the center of the strip under the embedded system will be realized. Firstly,the processor controls the image acquired by the CMOS camera in real time to smooth the filter and obtain the dynamic threshold for image segmentation by the Otsu method. Then,the sub-pixel precision center of the weld light stripe is obtained by the gray squared weighted threshold centroid method. The experimental results show that the proposed algorithm has strong anti-interference ability.The center of the strip of the original image is extracted by 0. 15 s on the development board platform,which can accurately extract the center of the weld strip.