基于SegNet网络和概率图模型的工区障碍物提取  

Obstacle Extraction for Exploration Areas Based on SegNet and Probabilistic Graphical Model

作　　者：胡敏 陈楠 毕进娜 HU Min;CHEN Nan;BI Jinna(Sinopec Geophysical Research Institute,Nanjing 211100,China)

机构地区：[1]中石化石油物探技术研究院有限公司,南京211100

出　　处：《科技和产业》2023年第17期266-272,共7页Science Technology and Industry

基　　金：中国石油化工股份有限公司科技部项目(P21062-1)。

摘　　要：为适应5G(第5代移动通信技术)智能节点后续投入生产后的高效地震采集工作,利用少量自制训练数据创新性地将传统算法与深度学习相结合,提出障碍物自动提取方法。搭建SegNet网络,得到粗糙的语义分割结果,利用条件随机场、特征与空间概率融合等概率图模型依次做了边界平滑和噪声消除的优化处理,最终的语义分割结果在各类障碍物上的准确率较高。与单一使用深度网络需要数万级数据来提升泛化能力相比,提出在少量训练集条件下具有较强泛化能力的语义分割方法,从而能够低成本、灵活高效地运用到各种特定的语义分割场景中。In order to adapt to efficient seismic acquisition after 5G(fifth-generation mobile communication technology)intelligent node being put into production,the scheme of automatically extracting obstacles has been proposed by using a little self-made training data and innovatively combining traditional algorithms with deep learning.A SegNet network has built,obtaining a rough segmentation result.Optimization processing including boundary smoothing and noise elimination,which achieved by utilizing probabilistic graphical model-conditional random fields and feature-space probability fusion respectively.Finally,the segmentation has attained higher accuracy rates on each class of obstacles.Compared with the deep network single using which needs tens of thousands of levels of data to improve the generalization ability,a semantic segmentation method has proposed with strong generalization ability under the condition of a little training data,which can be low-costly,flexibly and efficiently used in various specific semantic segmentation scenarios.

关 键 词：深度学习 条件随机场 语义分割 障碍物 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]