Automatic Timber Property Assessment for Variable Fibre Feedstocks
Project Description
This project aims to develop a framework and necessary digital tools for adaptively assessing mechanical properties of timbre components in timbre mill. Expected outcomes of this project include an enhanced capacity to identify and communicate:
• Review and identify sources of fibre (timber waste/byproducts, recycled timber, under- utilised timber resources).
• Vision-based defect detection and material property evaluation (such as MOE/MOR)
• Connect vision-based detection/evaluation with automatic design decision-making, allowing real-time material processing that reduces storage and informs human handling.
• Evaluation of recovery material value and sustainability metrics for possible applications and stakeholder communication.
• Assessment of application with industry systems.
Objectives/Deliverables
General Comments
• The below scope is limited to Australia, sawn softwood timber and derivatives.
• Data on improved fibre utilisation/resource recovery, design and fabrication systems from
other Node 6/7 projects can be integrated.
Objective A – Review of Current Mechanical Property Assessment Methods in Timber Mills and State-of-the-Art
This objective aims to systematically review the state-of-the-art in vision-based mechanical property testing for timber mills, with a focus on current applications in Australia.
• Review and identify sources of timber products for evaluation, including wood type, dimensions, and grading, in large-scale milling production (data provided by AKD).
• Identify the common mechanical grading systems used in Australian timber mills, assessing their data types, accuracy, scope and limitations.
• Examine state-of-the-art automatic mechanical property testing methods, conducting a comparative analysis between research-based and industry systems to identify limitations
and potential advancements.
Objective B- Vision-Based Detection and Evaluation
The goal is to develop a model capable of assessing the Modulus of Rupture (MOR) of timber products using image-based analysis. This initiative utilizes data from AKD, including mill data and on-site testing results. Key steps include:
• Collect and process image data from various mechanical grading systems used in timber mills, ensuring data consistency across different sources.
• Process the MOE, MOR, and potentially acoustic data collected from the milling facilities for the identified timber products provided by AKD. This includes creating a training dataset with corresponding images of timber boards.
• Train the vision detection model using the dataset to ensure accurate and stable predictions.
• Explore diverse approached and model structures to optimize the efficiency and accuracy of mechanical property prediction.
Objective C- Industry integration and assessment
To evaluate the model’s performance in real-world industrial settings and explore its feasibility for seamless integration into existing grading and assessment systems.
• Validate the model’s effectiveness by comparing it against existing benchmarks and assessing its commercial viability.
• Explore the integration of the model with the existing system such as MiCROTEC system and evaluate the feasibility of on-site deployment.
• Assess the model’s applicability to different feedstocks and industrial contexts, considering variations in species, product dimensions, data structures, and grading systems.
Targeted industry focused outcomes
This project aims to transform timber processing by integrating digital tools and advanced vision-based detection techniques to enhance efficiency, sustainability, and economic viability. Through a structured approach to mechanical property assessment, automated detection, and industry integration, the project will drive significant improvements in timber milling operations:
- A comprehensive review of mechanical property testing methods will provide valuable insights into the strengths and weaknesses of current industry practices. This will aid in identifying opportunities for technological advancements and standardization.
- The development of a vision-based detection model for defect identification and mechanical property evaluation will enhance grading accuracy, improving quality control and minimizing material waste.
- Seamless integration of automated detection with real-time processing will reduce reliance on manual handling, leading to increased operational efficiency and cost savings.
- The project will facilitate industry-wide improvements in material utilization by assessing the potential applications of recovered and underutilized timber resources, supporting sustainability goals.
- By aligning with industry needs and evaluating commercial feasibility, this project ensures that research outcomes translate into practical applications with measurable benefits.
Objectives/Deliverables
- Literature and Technical review: A review report on the current progress in mechanical property assessment for timber mills, and common systems used in Australia timber mill.
- Prototype and Exhibitions: The computational tool, robotic assembly system and structure prototypes will be exhibited at a minimum 1 domestic venue, such as Melbourne Design Week or National Gallery Victoria.
- Computer vision model: a computer vision model for performance evaluation and defect detection of specific types of fibre source and timber products.
- Publications: Peer-reviewed journal articles published in leading journals in the relevant areas (e.g., Automation in Construction; Construction and Building Materials, Additive Manufacturing).
- Conferences: Present research findings with industry partners at key national and international conferences, both academic and industrial.
- Final Report: A final report will be generated that includes all technical aspects of the digital toolbox, workflow, prototypes, and an evaluation of the system's commercial application based on quality assessment.
- A quarterly progress report to monitor and communicate ongoing developments.
- An annual report that provides a formal overview of the year's achievements and findings.
Project Leader/s
Dan Luo
Node Leader - Value-Chain Innovation; Project Leader
The University of Queensland
Project Staff
HDR Candidate
Higher Degrees by Research Scholarship Opportunity
Project Investigators
Dan Luo
Node Leader - Value-Chain Innovation; Project Leader
The University of Queensland
Joe Gattas
Theme Leader - Innovative Solutions; Node Leader - Manufacturing Innovation & Value-Chain Innovation
The University of Queensland
Mateo Gutierrez
Partner Investigator; Executive Board Member
AKD Softwoods
Lead Project Partner Organisation
Project Partners
