【摘要】 裂缝识别与评价是非常规油气储集层评价的重要任务之一，而成像测井是目前可信度最高的裂缝评价手段.针对现有成像图中裂缝识别方法的智能化程度、识别精度和效率不足等问题，本研究提出一种基于点面匹配算法的裂缝智能识别方法，其主要包括图像预处理、裂缝智能拾取、裂缝参数计算三大部分.首先，采用直方图均衡化、自适应滤波、阈值分割以及形态学操作对成像数据进行预处理，提高裂缝响应的信噪比和对比度，同时抑制非裂缝信息的干扰；其次，采用关键的点面匹配技术，将二维成像图中的数据映射到三维立体空间坐标系后，根据二维裂缝曲线在三维空间上呈现的空间曲线形态特征设立一个识别平面，利用映射后的三维成像数据与识别平面之间的拟合关系对裂缝进行识别，提高裂缝识别的准确度.计算拟合关系时，采用自适应窗口识别技术以节省计算资源，加快裂缝识别速率；最后，利用密度聚类算法对已识别的裂缝面进行分析归纳，提取出倾向、倾角等裂缝参数.为检验所提出算法对测井成像图的裂缝智能识别效果，本文一方面利用所构建出的不同裂缝模型图像对算法的裂缝识别效果进行探究，另一方面将该算法应用于油田实际成像测井图中的裂缝识别.结果表明，该算法能够准确识别成像图中的裂缝，并进一步有效计算出倾向、倾角、宽度以及长度等裂缝属性参数，从而验证了所提出算法的有效性与普适性.点面匹配算法能实现对复杂交叉裂缝参数的精确计算，有利于推动成像测井图中裂缝智能评价的进一步发展.更多还原

【Abstract】 Fracture identification and evaluation are critical components in the characterization of unconventional oil and gas reservoirs. Among the available techniques, micro-resistivity imager logging is currently regarded as the most reliable method for fracture assessment. To address the limitations in the intelligence level, accuracy, and efficiency of existing fracture identification algorithms in imaging logging, this paper proposes an intelligent fracture identification method based on a point–surface matching algorithm. The proposed method consists of three main components: image preprocessing, intelligent fracture picking, and fracture parameter calculation. First, histogram equalization, adaptive filtering, threshold segmentation, and morphological operations are applied to preprocess the imaging logging data, aiming to enhance the signal-tonoise ratio and contrast ratio of fracture responses while suppressing the interference from non-fracture features;Second, a key point–surface matching algorithm is applied by mapping the 2D imaging logging data into a 3D spatial coordinate system. An identification plane is established based on the spatial curve characteristics of the2D fracture in 3D space. The fitting relationship between the mapped 3D imaging data and the identification plane is then utilized for fracture identification, enhancing identification accuracy. During the calculation of the fitting relationship, an auto-adaptive window recognition technique is employed to optimize computational efficiency and accelerate fracture identification; Finally, a density clustering algorithm is used to analyze the identified fracture surfaces and extract parameters such as dip direction and dip angle. To evaluate the effectiveness of the proposed point-surface matching algorithm for intelligent fracture identification in imaging logs, this study examines its performance using both synthetic fracture model images and real imaging logging data from an oil field. The results demonstrate that the proposed algorithm can accurately identify fractures in imaging logs and effectively compute fracture attributes, including dip direction, dip angle, width, and length,thereby confirming the algorithm’s effectiveness and applicability. The point-surface matching algorithm enables precise calculation of complex cross-fracture parameters, promoting further advancements in intelligent fracture evaluation in imaging logging.更多还原