Performance prediction method based on Liang-Kleeman information flow and wood tracheid morphology

Abstract

In order to improve the production efficiency of the wood industry and achieve intelligent feedback control in production processes, non-destructive online prediction of performance is crucial. With the increasing maturity of nanotechnology, the use of scanning electron microscopy (SEM) is becoming increasingly frequent. The Liang Kleeman (L-K) information flow theory is a quantitative causal analysis method based on time series data, which has the advantage of fast calculation. In order to predict the mechanical properties of wood in real-time, fast, and non-destructive manner, this article obtains feature data from SEM photos of wood tracheids, A performance prediction method based on L-K information flow and wood tracheid morphology was proposed. This method first extracts the aspect ratio, wall cavity ratio, double thick wall, and cavity diameter ratio characteristics of wood tracheids from SEM photos, and then calculates the L-K information flow of the tracheid characteristics on mechanical properties (compressive strength, flexural strength, flexural modulus of elasticity, tensile strength). The causal influence strength between them is quantitatively characterized by the coefficient of variation of the L-K information flow, Then, a linear function is constructed with tracheid characteristics as the independent variable, mechanical properties as the dependent variable, and causal influence strength as a coefficient to achieve non-destructive prediction of wood properties. The experimental results show that the highest prediction accuracy of this method reaches 92%, effectively achieving online performance prediction based on causal relationships and image data, providing data reference for screening high-quality performance wood.