基于深度学习算法的杉木人工林单木冠幅回归模型

打开文本图片集

中图分类号：S757.2 文献标志码：A 文章编号：1000-2006（2026）01-0231-10

Abstract：【Objective】Thisstudyconstructedadeepleamning-basedcrownwidth model forCunninghamia lanceolata plantations intheJiangleregion，aiming toanalyzethecontributionsoftreesize，sitequality，standstructure，and competition tocrown width prediction.【Method】Six deepneural network（DNN）models weredeveloped，and the shapleyadditiveexplanations （SHAP）interpreter wasemployed to analyze the feature importance of each variable. Additionally，a generalizedcrown width model was builtandcompared with the DNNmodelthatutilized the same setof optimallyselectedvariables basedoncorrelationcoefficients，inorder tovalidatethereliabilityof theDNNapproach. 【Result]The constructed DNN models exhibited no overfiting，with stable root mean square error（RMSE）and mean absolute error（MAE）valuesin10-fold cross-validation.The DNN modelincorporating all13variables（DNN5） achieved the highest performance，with an R2 ofO.6O.However，models with six input variables（DNN3 and DNN4） yielded R2 values of O.58 and O.57，respectively，which beter meet practical application needs.The DNN6 model， constructed using variables selected based on their correlation coefcients，achieved an R2 of 0.54 ，outperforming the generalized crown width model（ R2=0.46 ）that used the same variables.The SHAP contribution value of DBH was the highest.Additionally，inthe featureranking，basalarea（BA），stand densityindex（SDI），McIntosh evenessof diameterclass（Emi），andGinicoefficient（GC）held important positions.The inclusionof both typesofvariables significantlyimproved prediction theacuracyof thecrown widthmodel.【Conclusion】Theresults demonstrate that the constructed DNN model effectively predicted the crown width of c . lanceolata in the study area，indicating that deep learning holds significant potential for crown width modeling.

Keywords：crown width model; Cunninghamia lanceolata （Chinese fir））;dep learning;deep neural netword （DNN）; SHAP interpretability analysis

杉木（Cunninghamialanceolata）是亚热带地区优质速生针叶树种，分布广阔，也是我国人工林面积最大的造林树种和南方地区经营历史最长的用材树种，在我国林业产业中占有重要地位[。（剩余16300字）