CAJ | 학술논문

为了从FMI资料中定量提取参数,一个重要的步骤是从实际FMI资料中分离出反映溶孔、溶洞、裂缝的子图像.本文给出的方法,考虑图像像元邻域的特征,应用二维小波变换求出目标与背景边缘的点集,按这个边缘点集的坐标点所对应的原图像像素灰度值的平均值作为分割阈值进行图像分割.实际资料处理表明,应用这种方法可以从实际的FMI资料中准确地分割出孔洞、裂缝的子图像并且可以按深度段连续自动处理,为后续定量计算参数奠定了良好基础.
위료종FMI자료중정량제취삼수,일개중요적보취시종실제FMI자료중분리출반영용공、용동、렬봉적자도상.본문급출적방법,고필도상상원린역적특정,응용이유소파변환구출목표여배경변연적점집,안저개변연점집적좌표점소대응적원도상상소회도치적평균치작위분할역치진행도상분할.실제자료처리표명,응용저충방법가이종실제적FMI자료중준학지분할출공동、렬봉적자도상병차가이안심도단련속자동처리,위후속정량계산삼수전정료량호기출.
A key aspect in extracting quantitative information from FMI logs is to segment the FMI image to get images of pores, vugs and fractures. A segmentation method based on the dyadic wavelet transform in 2-D is introduced in this paper. The first step is to find all the edge pixels of the FMI image using the 2-D wavelet transform. The second step is to calculate a segmentation threshold based on the average value of the edge pixels. Field data processing examples show that sub-images of vugs and fractures can be correctly separated from original FMI data continuously and automatically along the depth axis. The image segmentation lays the foundation for in-situ parameter calculation.