基于阈值分割与特征提取的支撑剂砂床形态分析方法

Morphological Analysis of Proppant Sand Beds Based on Threshold Segmentation and Feature Extraction

摘要: 裂缝内砂床形态预测是水力压裂领域研究的重点和难点。目前支撑剂运移实验数据处理仍以人工方式为主，数据数量有限且效率低下，砂床形态的形成和演化亟需深入研究。本文提出了一套用于砂床形态识别和特征参数提取的方法。该方法首先利用阈值分割算法进行精准的砂床轮廓提取。随后，通过计算机视觉方法识别和量化砂床的形态特征，进而应用特征参数提取技术获得六个关键的砂床形态特征参数。这些方法能够在支撑剂运移连续时刻下实现自动化的特征提取和定量分析，无需依赖人工测量。该方法能够准确且完整地捕捉支撑剂图像中的砂床形态特征及其变化规律，为水力裂缝内支撑剂砂床形态的研究提供了新的方法和思路。通过自动化的特征提取，可以更深入地理解砂床形态的演化过程，从而推动水力压裂技术的进步和优化。

Abstract: The prediction of proppant bed morphology within fractures is a key and challenging topic in the field of hydraulic fracturing. At present, the processing of proppant transport experimental data still relies primarily on manual methods, which are inefficient and limited in data volume. The formation and evolution of the proppant bed require further in-depth investigation. This study proposes a method for identifying proppant bed morphology and extracting characteristic parameters. The method first employs a threshold segmentation algorithm to accurately extract the bed contours. Then, computer vision techniques are used to identify and quantify morphological features of the bed, followed by the application of feature extraction algorithms to obtain six key morphological parameters. These methods enable automated feature extraction and quantitative analysis at continuous time steps of proppant transport, eliminating the need for manual measurement. The proposed approach can accurately and comprehensively capture the morphological features and dynamic changes of the proppant bed in images, offering new tools and perspectives for studying bed morphology within hydraulic fractures. Through automated feature extraction, the evolution of bed morphology can be better understood, thereby promoting the advancement and optimization of hydraulic fracturing technology.