论文部分内容阅读
为快速提取金属熔化焊接缺陷位置,提出了一种基于灰度分布曲线特征的算法。首先,通过平板探测器获取焊缝的X射线透射数字图像;第二,针对数字图像随机噪声大、动态范围低的问题,采用时间域序列图像帧积分方法,计算最优的积分时间,降低图像噪声,强化焊缝缺陷特征细节;第三,平行于焊缝方向,计算并形成多组图像灰度值曲线,基于其二阶导数曲线分布规律提取焊缝图像;最后,在焊缝图像曲线中进行极值点搜索,若在曲线存在明显极大值点,则可以确定缺陷的大小和位置。试验结果表明,本算法具有较高的检测精度。
In order to quickly extract the location of defects in metal fusion welding, an algorithm based on the gray-scale curve is proposed. Firstly, the X-ray transmission digital image of the weld is obtained by the flat panel detector. Secondly, aiming at the problem of large random noise and low dynamic range of the digital image, the time-domain image frame integration method is used to calculate the optimal integration time and reduce the image Noise and strengthen the weld defects; Third, parallel to the weld direction, calculate and form multiple sets of image gray value curve, based on the second derivative curve of the distribution of the weld image; Finally, the weld image curve In the extreme point search, if there is a significant point in the curve of the curve, you can determine the size and location of defects. Experimental results show that this algorithm has higher detection accuracy.