基于SwinV2-EfficientNetV2的銅礦石品位分類(lèi)方法研究
DOI:
作者:
作者單位:

東華理工大學(xué)

作者簡(jiǎn)介:

通訊作者:

中圖分類(lèi)號:

TD92????????????? ?????????????

基金項目:

國家自然科學(xué)基金資助項目(U2067202);江西省主要學(xué)科學(xué)術(shù)和技術(shù)帶頭人培養計劃(No.20225BCJ22004);江西省重點(diǎn)研發(fā)計劃(20203BBG73069)


Research on copper ore grade classification method based on SwinV2-EfficientNetV2
Author:
Affiliation:

East China University of Technology

Fund Project:

  • 摘要
  • |
  • 圖/表
  • |
  • 訪(fǎng)問(wèn)統計
  • |
  • 參考文獻
  • |
  • 相似文獻
  • |
  • 引證文獻
  • |
  • 資源附件
  • |
  • 文章評論
    摘要:

    針對現有銅礦石品位分類(lèi)中所應用的卷積神經(jīng)網(wǎng)絡(luò )缺乏構建與歸納長(cháng)距離特征關(guān)系不足的問(wèn)題,作者提出了一種結合SwinTransformer-EfficientNet集成模型的銅礦石品位分類(lèi)方法。該方法充分利用了SwinTransformer V2-t架構對長(cháng)距離特征關(guān)系的歸納能力,以及EfficientNet V2-s捕捉細微局部特征上的優(yōu)勢,通過(guò)增設線(xiàn)性層以整合兩模型的輸出結果,并根據單個(gè)模型自身的輸出動(dòng)態(tài)調整線(xiàn)性層的權重,以?xún)?yōu)化映射關(guān)系,進(jìn)而顯著(zhù)提升分類(lèi)性能时项。實(shí)驗驗證表明,此融合模型在分類(lèi)任務(wù)上的準確率達到92.891%,精確率達到93.095%,召回率達到92.654%兑碧。相較于未集成前的單一模型,集成后的綜合模型在分類(lèi)準確率上提升了1.30%,精確率分別提升了1.9%和2.186%,召回率則分別提高了0.474%和0.237%,效果明顯。

    Abstract:

    In response to the inadequacy of constructing and encapsulating long-distance feature relationships in convolutional neural networks currently utilized for copper ore grade classification, the author proposes a method that combines a SwinTransformer-EfficientNet ensemble model. This methodology fully exploits the SwinTransformer V2-t architecture's capability in summarizing long-range feature associations, as well as the EfficientNet V2-s's strength in discerning subtle local characteristics. By incorporating a linear layer to amalgamate the outputs of both models and adaptively tuning the weights of this linear layer according to the individual model's output, the mapping relationship is optimized, leading to a substantial enhancement in classification performance. Empirical validation indicates that this fused model attains an accuracy of 92.891%, precision of 93.095%, and recall of 92.654% in classification tasks. Relative to the standalone, non-integrated models, the integrated composite model exhibits an increase of 1.30% in accuracy, 1.9% and 2.186% in precision, and 0.474% and 0.237% in recall, respectively, manifesting considerable advancements.

    參考文獻
    相似文獻
    引證文獻
引用本文
分享
文章指標
  • 點(diǎn)擊次數:
  • 下載次數:
  • HTML閱讀次數:
  • 引用次數:
歷史
  • 收稿日期:2024-04-10
  • 最后修改日期:2024-06-28
  • 錄用日期:2024-06-28
  • 在線(xiàn)發(fā)布日期:
  • 出版日期: