论文部分内容阅读
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。
In order to quantitatively extract the parameters from the FMI data, an important step is to separate the sub-images that reflect the dissolved pores, karst holes and fractures from the actual FMI data. The method given in this paper considers the features of neighborhoods of image pixels and applies two-dimensional wavelet transform to find the point set of target and background edge. According to the average value of the original image pixel gray value corresponding to the coordinate point of this edge point set The image is divided as a segmentation threshold. The actual data processing shows that the sub-images of holes and fractures can be accurately segmented from the actual FMI data by this method and can be processed automatically and continuously according to the depth, which lays a good foundation for subsequent quantitative calculation of parameters.