基于钻采管材断口图像特征的图像预处理方法研究

Image Preprocessing Method Based on Fracture Image Characteristics of Drilling Tubular Goods

分析钻采管材断口宏观图像可以为管材材料的性能及服役行为等许多方面的研究提供重要信息,准确分析钻采管材断口宏观形貌需要对断口宏观图像进行剪切、灰度化、直方图均衡化、去噪、滤波和边缘检测等预处理。图像分块后,通过子块的灰度均值、方差以及方向图识别噪声子块,利用噪声子块邻近的纹理子块沿纹理方向滑动对噪声子块进行修复,实现图像去噪。对去噪后的图像进行Gabor滤波处理、边缘检测等操作。实验结果表明,通过上述预处理方法可以获得较为理想的断口宏观纹理图像,为后续特征提取、断口类型识别等奠定基础。

The analysis of the macroscopic image of the fracture of the drilling tubular goods can provide important information for the study of many aspects such as the performance and service behavior of the tubular goods.Accurate analysis of the macroscopic morphology of the fracture requires pre-processing such as shearing,grayscale,histogram equalization,denoising,filtering and edge detection,etc.After the image is divided into blocks,the noise sub-blocks are identified by the gray mean,variance and pattern of the sub-blocks.The noise sub-blocks are repaired by sliding the texture sub-blocks adjacent to the noise sub-blocks along the texture direction to achieve image denoising.Gabor filtering and edge detection are performed on the denoised image.The experimental results show that the ideal macroscopic texture image of fracture can be obtained by the above pretreatment method,which lays groundwork for subsequent image feature extraction and fracture type identification.