Scholarlink

<p>Interface of geological formations (IGF) refers to the interface between adjacent strata of different types, presenting significant challenges in underground structure construction. This study addresses two key issues regarding to IGF, including the IGF inversion and IGF similarity characterization. The IGF inversion algorithm was developed by integrating the Bayesian-based conditional random field with trend functions. The proposed algorithm not only considers the mean uncertainty to mitigate the effect of human factors on determination of the prior information, but also simultaneously handles the sparsity and trend characteristics inherent in the measurements. IGF similarity characterization method was also developed by combing this novel IGF inversion algorithm with image structural similarity (SSIM) theory. The performance of the IGF inversion algorithm was demonstrated using a set of real rockhead data. The effectiveness of IGF similarity characterization method was also validated through an advance drilling project. The results indicate that the proposed IGF similarity characterization method provides a rational assessment of IGF similarity among different sites, yielding an absolute and monotonic similarity indicator within the range of [0, 1]. Moreover, the uncertainty associated with the similarity characterization results can be quantified.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Inversion and similarity characterization on the interface of geological formations using conditional random field and image structural similarity

  • 期刊论文

  • 发表日期：

    2026年2月12日

查看全文
  • Liang Han,
  • Haijun Liu,
  • Wengang Zhang,
  • Yu Wang,
  • Hanting Diao

摘要

Interface of geological formations (IGF) refers to the interface between adjacent strata of different types, presenting significant challenges in underground structure construction. This study addresses two key issues regarding to IGF, including the IGF inversion and IGF similarity characterization. The IGF inversion algorithm was developed by integrating the Bayesian-based conditional random field with trend functions. The proposed algorithm not only considers the mean uncertainty to mitigate the effect of human factors on determination of the prior information, but also simultaneously handles the sparsity and trend characteristics inherent in the measurements. IGF similarity characterization method was also developed by combing this novel IGF inversion algorithm with image structural similarity (SSIM) theory. The performance of the IGF inversion algorithm was demonstrated using a set of real rockhead data. The effectiveness of IGF similarity characterization method was also validated through an advance drilling project. The results indicate that the proposed IGF similarity characterization method provides a rational assessment of IGF similarity among different sites, yielding an absolute and monotonic similarity indicator within the range of [0, 1]. Moreover, the uncertainty associated with the similarity characterization results can be quantified.