基于BP算法的焊缝X射线成像校正算法研究

摘要

In order to solve the X-ray imaging distortion problem of long-distance pipeline welding,according to the characteristics of long distance pipeline,this paper fixed pictures based on BP neural network and used the method of clustering and adjust the weights and threshold.This paper improved algorithm on the basis of analyzing algorithm to achieve the image distortion correction of long distance pipeline.The algorithm can significantly inhibit the amount of calculation and improve the accuracy of correction.It's the first time completed the corrected image of X-ray in the industry.%为解决长输管道焊缝X射线成像时畸变问题,文中根据长输管道的特点基于BP神经网络方法进行图像校正,采用聚类与调节权值、阈值的方式应用于图像校正,在分析算法的特点的基础上,优化并改进了该算法,从而对长输管道成像畸变完成了校正.能够明显地抑制计算量、提高校正的精确度.首次在工业上完成了对X射线的图像校正.