基于图像识别的原始地质编录矢量化三维建模方法 |
投稿时间:2022-11-02 修订日期:2023-03-09 点此下载全文 |
引用本文: |
摘要点击次数: 39 |
全文下载次数: 0 |
|
|
中文摘要:针对传统原始地质编录建模未考虑图件污染、矢量化效率低、模型精确度不高等问题,本文提出了一种基于图像识别的原始地质编录矢量化三维建模方法。该方法主要包括以下几个步骤:(1)读取并裁剪原始地质编录图纸;(2)采用Rosenfeld算法降噪细化图形;(3)采用Douglas-Peucker算法提取并抽稀分离地质素描界线;(4)三维化地质编录界线;(5)采用Coons曲面精准构建三维地质体模型。本文所提出的方法被成功应用于云锡个旧矿区高松矿田的三维地质建模中。数值模拟实验结果表明,相比于传统原始地质编录三维建模,使用本方法的建模结果更能真实反映矿体形态变化,建模速度从20min缩短至30s内,极大地提高了地质编录建模效率,进而降低了井下生产勘探和采掘作业工作的风险。 |
中文关键词:图像识别 图像处理 原始地质编录 地质建模 曲面重建 |
|
3D Modeling Method Based on the Pattern Recognition of Vectorized Original Geological Logging Data |
|
|
Abstract:At present, the vectorization efficiency of traditional original geological logging data is low, and the accuracy of 3D model is not high, regardless of the contamination of the drawings. An 3D modeling method based on the pattern recognition of vectorized original geological logging data is proposed. The method mainly consists of the following several steps:(1) reading and cropping the original geological logging drawings;(2)reducing noise of graphic and refining graphic based on Rosenfeld algorithm;(3) extracting and separating the geological boundary lines based on Douglas-Peucker algorithm;(4) making three dimensions of the geological logging boundary lines;(5) Constructing 3D geological models based on Coons surface. This method was successfully applied to the GaoSong field in the GeJiu mining area of Yunxi. The results of the numerical simulation experiments show that the modeling results using this method can reflect the morphological changes of the ore body more realistically than the traditional 3D modeling of the original geological logging data. The modeling speed is reduced from 20 min to less than 30s. It greatly improves the efficiency of geological logging modeling and reduces the risk of underground productive exploration and work of extraction and development. |
keywords:Pattern recognition Image processing Original geological logging Geological modeling Surface reconstruction |
查看/发表评论 下载PDF阅读器 |