Scholarlink

<p>Rock core logging is critical for evaluating rock mass quality, yet traditional manual measurement of rock quality designation (RQD) indicator from boreholes remains highly labor-intensive. To address the limitations of insufficient detection accuracy and generalizability in mainstream deep learning methods, we propose MG-FSAM4Seg, a vision foundation model adapted for rock core instance segmentation, combined with a fine-grained RQD analytics method. MG-FSAM4Seg adopts an enhanced multi-scale feature learning design that seamlessly integrates the Segment Anything Model (SAM) encoder with additional decoders, where the encoder is further fine-tuned via low-rank adaptation. We evaluate our model on two core imagery datasets, including the self-constructed Core1000 and the independent cross-project Core-G47, which exhibit considerable lithological and resolution differences, benchmarked against 15 advanced segmentation methods. On the Core1000 test set, MG-FSAM4Seg achieves competitive performance in both qualitative and quantitative evaluations, reaching an average precision of 88.07%. In a zero-shot transfer evaluation on Core-G47, the model maintains strong generalization, achieving an average precision of 86.46%. Moreover, RQD values estimated from MG-FSAM4Seg segmentation results, combined with our proposed core length calculation approach, show high agreement with manual measurements, yielding an <i>R</i><sup>2</sup> of over 0.95. These results confirm that MG-FSAM4Seg provides an effective solution for vision-based RQD analytics, demonstrating its potential to advance automated geological logging.</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

Fine-Tuned SAM Adaptation with Multi-scale Guidance for Automated Detection Toward Image-Based Core Length and RQD Measurement

  • 期刊论文

  • 发表日期：

    2026年2月26日

查看全文
  • Kunpeng Shi,
  • Qiubing Ren,
  • Mingchao Li,
  • Yantao Yu,
  • Zuguang Zhang,
  • Huihui Jia

摘要

Rock core logging is critical for evaluating rock mass quality, yet traditional manual measurement of rock quality designation (RQD) indicator from boreholes remains highly labor-intensive. To address the limitations of insufficient detection accuracy and generalizability in mainstream deep learning methods, we propose MG-FSAM4Seg, a vision foundation model adapted for rock core instance segmentation, combined with a fine-grained RQD analytics method. MG-FSAM4Seg adopts an enhanced multi-scale feature learning design that seamlessly integrates the Segment Anything Model (SAM) encoder with additional decoders, where the encoder is further fine-tuned via low-rank adaptation. We evaluate our model on two core imagery datasets, including the self-constructed Core1000 and the independent cross-project Core-G47, which exhibit considerable lithological and resolution differences, benchmarked against 15 advanced segmentation methods. On the Core1000 test set, MG-FSAM4Seg achieves competitive performance in both qualitative and quantitative evaluations, reaching an average precision of 88.07%. In a zero-shot transfer evaluation on Core-G47, the model maintains strong generalization, achieving an average precision of 86.46%. Moreover, RQD values estimated from MG-FSAM4Seg segmentation results, combined with our proposed core length calculation approach, show high agreement with manual measurements, yielding an R2 of over 0.95. These results confirm that MG-FSAM4Seg provides an effective solution for vision-based RQD analytics, demonstrating its potential to advance automated geological logging.