基于多数据融合和小波分解的油气管道缺陷检测方法研究

针对油气管道检测过程中泄漏区域及缺陷区域难以准确定位的问题,提出了一种基于多数据融合的油气管道缺陷自动检测方法。首先利用无人机和红外热成像技术实现油田管道红外图像的分段采集,并通过移动终端实时上传到处理系统;采用改进的加权中值滤波算法对红外图像做去噪处理,并根据综合特征来自动识别管道区域以提升检测效率;然后采用Sobel算子和Canny算子对管道区域进行图像处理,并加权平均;最后,采用Haar小波做分解得到高频分量,高频分量和红外图像中温度异常区域重合的部位即为油气管道缺陷区域。实验结果表明:该方法可准确地定位油气管道的缺陷区域,其准确率最高可达97.8%,检测效率也有所提升。

Research on Oil and Gas Pipeline Defect Detection Based on Multi-data Fusion and Wavelet Decomposition

Aiming at the difficulty of locating the leakage area and defect area accurately in the process of oil and gas pipeline inspection, an automatic detection method for oil and gas pipeline defects based on multi-data fusion is presented. First, UAV and infrared thermal imaging technology are used to collect the infrared image of oilfield pipeline in segments and upload it to the processing system through mobile terminal in real time. Then, an improved weighted median filter algorithm is used to denoise the infrared image, and the pipeline area is automatically identified according to the comprehensive features to improve the detection efficiency. Sobel operator and Canny operator process image of pipeline area, and weighted average the two images. Finally, Haar wavelet is used to decompose the high-frequency component, and the location where the high-frequency component coincides with the temperature anomaly area in infrared image is the defect area of oil and gas pipeline. The experimental results show the method can accurately locate the defective areas of pipeline, with the highest accuracy of 97.8%, and the detection efficiency has also been improved.