套损图像特征模式匹配识别技术及应用

Casing Damage Image Technology of Feature Pattern Recognition and Matching of Casing Damage Image and Its Application

和志明 伍东

HE Zhiming WU Dong

中国石油集团长城钻探工程有限公司国际测井公司, 北京 朝阳 100101

International Logging Company of Greatwall Drilling Company, CNPC, Chaoyang, Beijing 100101, China

针对套管损伤腐蚀检测识别问题，给出了4种损伤腐蚀类型的定义，开展了损伤腐蚀特征模式匹配识别研究，对传统图像轮廓跟踪与提取做了优化改进，将套损图像变换到三维空间，采用三维空间图像特征模式识别技术，通过对套损数据集中点间距测度阈值和点数值阈值的约束，搜索得到损伤腐蚀闭合三维体，将其划分为表示不同损伤腐蚀形状的特征子集，并提取特征子集的轴向和径向特征，对提取的特征和标准腐蚀模式特征进行匹配，可准确评判套管损伤腐蚀类型。实际应用表明，该技术能够准确地检测环状腐蚀、线状腐蚀、片状腐蚀和斑点状腐蚀，解决了目前无法识别套管损伤腐蚀类型的难题。

Aiming at the problem of casing damage and corrosion detection and recognition, four kinds of damage and corrosion types are defined, and the pattern matching recognition of damage and corrosion characteristics is studied. The traditional image contour tracking and extraction are optimized and improved. The casing damage image is transformed into three-dimensional one, and the feature pattern recognition technology of three-dimensional image is used to search a three-dimensional damage-corrosion closure volume, by constraining on distance thresholds between data points and numerical thresholds in casing damage data sets. Different three-dimensional damage-corrosion closure volume represent the different characteristic subsets of damage and corrosion shapes. The axial and radial features of the feature subset are extracted, and the extracted features are matched with the standard corrosion pattern features to accurately evaluate the damage-corrosion type of casing. Practical application shows that it can accurately detect ring corrosion, line corrosion, general corrosion and isolated corrosion, and solve the problem of identifying the corrosion type of casing damage at present.

10.11885/j.issn.1674-5086.2019.03.29.01