| 来俊海,白向举,蔺华锋,李会双,孙世勇.基于改进 FCM聚类算法的隐伏地质构造三维图像分割方法[J].矿产勘查,2024,15(12):2337-2344 |
| 基于改进 FCM聚类算法的隐伏地质构造三维图像分割方法 |
| 3D image segmentation method of hidden geological structure based on improved FCM clustering algorithm |
| 投稿时间：2023-03-28 |
| DOI：10.20008/j.kckc.202412016 |
| 中文关键词: 三维图像 隐伏地质构造 图像分割 改进 FCM聚类算法 曲面去噪 遥感图像 |
| 英文关键词: 3D image hidden geological structure image segmentation improved FCM clustering algorithm surface denoising remote sensing images |
| 基金项目: |
|
| 摘要点击次数: 140 |
| 全文下载次数: 49 |
| 中文摘要: |
| 隐伏地质构造中蕴含着较多的矿产资源，是现今地质研究领域的关键所在。由于隐伏地质构造位置较 深，再加之信号传输衰减、环境、光线等因素的影响，致使隐伏地质构造三维图像复杂度较高，无法获取精 准的隐伏地质构造信息。本文提出基于改进 FCM聚类算法的隐伏地质构造三维图像分割方法研究，归一 化处理三维图像，应用拉普拉斯算子去除三维图像曲面噪声信息，提升三维图像信噪比与清晰度，基于分 形维数理论定位目标隐伏地质构造，采用自适应邻域加权改进 FCM聚类算法，基于改进 FCM聚类算法制 定三维图像分割程序，执行制定程序即可实现隐伏地质构造三维图像的分割。实验数据显示:应用提出 方法获得的三维图像分割精度最大值达到了 96.32%，充分证实了提出方法三维图像分割效果更好。 |
| 英文摘要: |
| Hidden geological structure contains more mineral resources, which is the key in the current geo.logical research field. Due to the deep location of the hidden geological structure and the influence of signal trans.mission attenuation, environment, light and other factors, the 3D image of the hidden geological structure has a highcomplexity, which makes it impossible to obtain accurate information of the hidden geological structure. A study on the segmentation method of the 3D image of the hidden geological structure based on the improved FCM clusteringalgorithm is proposed. Normalize the 3D image, apply Laplace operator to remove the surface noise information ofthe 3D image, improve the signal-to-noise ratio and clarity of the 3D image, locate the hidden geological structureof the target based on the fractal dimension theory, use adaptive neighborhood weighting to improve FCM clusteringalgorithm, develop a 3D image segmentation program based on the improved FCM clustering algorithm, and imple.ment the developed program to achieve the segmentation of the 3D image of the hidden geological structure. The ex.perimental data show that the maximum accuracy of 3D image segmentation obtained by the proposed methodreaches 96.32%, which fully proves that the proposed method has better 3D image segmentation effect. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |