Page 51 - FDMAsia JulAug 2024

TECHNOLOGY 49
www.fdmasia.com | FDM ASIA JUL/AUG 2024

Lin, including forest inventory. This approach could also be a empirically measured variables to create the most effective
promising area of research in the evaluation of wood quality. prediction model possible.
Overall, further developments are needed before achieving On the other hand, process-based models reconstruct
their widespread application, not only for assessing tree size the tree’s architecture. Obviously, the development of wood

but also for evaluating qualitative attributes. One aspect quality modelling relies on data collected in the field, making
that certainly holds great potential for progress is software them vital for improving model validity and their applicability
development. on a large scale.
Faster and more effective algorithmic advancements Finally, particular attention should be directed towards studies
for automatic recognition and subsequent measurement of that reconstruct the development of branches internally to the
pertinent characteristics are critical for enhancing tool usability. stem (branch geometry) starting from externally detectable
For instance, computer vision and deep learning methods parameters, such as branch diameter, insertion angle, position
appear very promising, so they enable real-time evaluations. in the stem, tree height, and stem diameter.
Apart from further enhancements to improve the usability Accurate modelling in this regard could significantly improve

of survey techniques, as discussed previously, it is undeniable the qualitative evaluation by simulating knottiness in the sawn
that they hold great potential for evaluating the visible product obtained from log processing.
characteristics of standing trees. Returning to the definition reported in the first paragraphs,
The next question, therefore, concerns the effectiveness of wood quality is inextricably linked to the requirements of
the metrics detected in predicting the quality of the product the final product. To better analyse which attributes of the
obtained from cutting the tree and if there are better evaluation standing tree best allow it to satisfy the quality requirements
criteria for wood quality that could be included in the survey. of the sawn product, it is necessary to connect the information

Addressing this involves investigating how current metrics collected in the forest with the data collected in the sawmill.
correlate with end-product quality and identifying additional To achieve this, the traceability of tree-log-sawn timber
parameters that provide a more accurate assessment. information is a key aspect of linking the data to be used in
Research in this area is crucial to develop more reliable the learning of prediction algorithms.
and predictive models for wood quality, ultimately enhancing Implementing traceability systems would track the
the precision and utility of forestry assessments. resource through all the processing phases, optimising both
Better estimations were reported when linking branch management and cultivation decisions, as well as harvesting
scars on the stem surface with the knots and the clear wood and processing, with the conscious aim of achieving the
amount measured with X-ray computer tomography on logs. highest possible product quality.

Thus, further studies are certainly necessary to explore the Several traceability solutions have been developed and
most effective metrics for describing product quality. tested. Some are further behind, mainly due to high application
Additionally, the development of models estimating internal
tree characteristics from externally visible ones appears very Minwax Blog
promising.
A detailed review of studies dedicated to modelling wood
quality is provided by Drew et al. In this work, the models

were divided into two main types: empirical and process-
based models. The former established relationships between