Page 33 - FDMAsia Sep/Oct 2026

TECHNOLOGY 31
www.fdmasia.com | FDM ASIA SEP/OCT 2026

Traditional defect inspection in wood primarily relies on
manual methods, which are subjective and labour-intensive, Francis Eatherington
failing to meet the demands of modern automated manufacturing.
Moreover, the reliability of manual inspection rarely exceeds
70 percent, as it is affected by human factors such as visual
fatigue and lapses in concentration.

Consequently, alternative inspection techniques have been
developed, including acoustic, X-ray, and terahertz methods.
Among these, vision-based systems integrated with machine
learning have emerged as a promising solution for wood surface
defect inspection due to their cost-effectiveness, efficiency,
and adaptability to field conditions. Main Wood Surface Defects
Extensive research on this approach can be found in the The categories of wood defects considered vary across
literature. Although the types of input images vary across industries, depending on their specific objectives. In the
studies, most approaches follow two key steps: feature construction industry, defect inspection primarily focuses on

extraction and classification. assessing strength-related characteristics, with emphasis on
For feature extraction, techniques such as Gray Level insect damage, decay, and splits.
Co-occurrence Matrix, Gabor filters, and Local Binary Pattern In the furniture industry, inspections aim to eliminate both
analysis have been employed to capture the distinctive colour aesthetic imperfections and structural weaknesses, paying
and texture characteristics of wood surface defects. particular attention to knots, holes, and cracks.
Subsequently, classifiers including Artificial Neural Networks
and Support Vector Machines analyse these features and Table 1. The Types of Wood Surface Defects Concerned

produce the final decision.
Alternative methods such as clustering and compressed
sensing have also been explored. However, both feature
extraction and classifier training are manually designed, making
them dependent on human expertise.
Furthermore, the performance of these methods degrades
with changes in wood species, surface conditions, or lighting Live knots, dead knots, and checks are consistently identified
environments. Therefore, developing reliable techniques for as critical defects in the literature due to their substantial
wood surface defect inspection remains an ongoing challenge impact on the quality of wood products.

for the wood industry. A high density of live knots complicates the wood grain
Deep learning, an advanced subfield of machine learning, pattern and reduces its ornamental value. In contrast, a dead
autonomously learns feature hierarchies from image datasets. Its knot results from a deceased branch. Its fibre structure is often
capacity to extract discriminative features with minimal human partially or completely detached from the surrounding wood
intervention or specialised domain knowledge has profoundly tissue, significantly compromising the mechanical properties
influenced wood surface defect inspection, positioning deep of solid wood panels.
learning as a leading research trend. Checks are fissures caused by the separation of wood