Method for predicting mechanical properties of wood based on fractal dimension

By processing multi-resolution pore structure image data and using machine learning models, the problem of the difficulty in characterizing the heterogeneity of pore structure across scales in wood has been solved, achieving high-precision and highly interpretable prediction of wood mechanical properties and improving the effectiveness of non-destructive testing and quality grading of wood.

CN122244056APending Publication Date: 2026-06-19SOUTHWEST FORESTRY UNIVERSITY
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SOUTHWEST FORESTRY UNIVERSITY
Filing Date
2026-05-25
Publication Date
2026-06-19

AI Technical Summary

⚠Technical Problem

Existing technologies are insufficient to effectively characterize the heterogeneity of wood's cross-scale pore structure and its impact on mechanical properties, resulting in limited accuracy and insufficient interpretability in the prediction of wood's mechanical properties.

⚗Method used

By acquiring multi-resolution pore structure image data, spatial alignment, adaptive segmentation, and information fusion are performed to generate a multi-scale fused binary model of wood pore phase. Scale spectrum analysis is conducted and feature scale intervals are adaptively divided. A local fractal dimension algorithm with scale decoupling is applied to calculate the multi-scale features of the pore structure. An enhanced multi-scale pore fractal feature vector is constructed, and a machine learning model with physical information embedding and multi-task attention mechanism is used for prediction.

🎯Benefits of technology

It achieves high-precision and highly interpretable prediction of wood mechanical properties, improves the prediction accuracy and reliability of the model, and provides an effective technical approach for non-destructive testing and quality grading of wood.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
Patent Text Reader

Abstract

This invention relates to the field of material performance prediction technology, and particularly to a method for predicting the fractal dimension mechanical properties of wood. The method includes: acquiring multi-resolution pore structure image data of a wood sample to be tested; performing spatial alignment, adaptive segmentation, and information fusion processing to generate a multi-scale fused binary model of the wood pore phase; performing scale spectrum analysis and adaptively dividing multiple feature scale intervals; applying a scale-decoupled local fractal dimension algorithm to each interval to calculate the corresponding scale-decoupled local fractal dimension value; combining and deriving calculations of multiple dimension values ​​to construct an enhanced multi-scale pore fractal feature vector; inputting the enhanced multi-scale pore fractal feature vector into a wood mechanical property prediction model obtained by joint optimization training using physical information embedding and a multi-task attention mechanism; and outputting predicted mechanical property indicators and corresponding prediction uncertainty metrics. This invention achieves high-precision and highly interpretable prediction of wood mechanical properties.
Need to check novelty before this filing date? Find Prior Art