Walailak University researchers developed a revolutionary new system for detecting wood cores in rubberwood logs. With 100 percent accuracy, the system not only enhances processing efficiency but also cuts costs and the need for imported machinery. The research has been published in an International Journal listed in Scopus Q1 database and the researchers plan to apply their findings to the medical field for detecting strabismus in humans.
Associate Professor Dr. Wattanapong Kurdthongmee and Assistant Professor Korakot Suwannarat, researchers from Walailak University’s School of Engineering and Technology Program in Computer Engineering and Artificial Intelligence, have developed a system that uses deep neural network techniques to detect wood cores in rubberwood log cross-section. This system, which is connected to systems in rubberwood processing plants, significantly reduces the loss of rubberwood during the baking and cutting process.
In the Thai rubberwood industry, the processing of rubberwood currently requires skilled workers to control the core of the wood during the cutting process into sheet wood. However, due to a shortage of skilled workers, the crooked appearance of the wood core often results in damaged sheet wood during the baking process, leading to unsellable logs and a waste of valuable wood worth over 500,000 baht per month, or 6 million baht per year.
Assoc. Prof. Dr. Wattanapong, the head of the research project, revealed that the invention of the wood core detection system for rubberwood log cross-section has two components: a log rotation subsystem and a log drive system that integrates with the saw system. The system was trained using AI cameras processed by deep neural networks on 400+ images of rubberwood logs to capture cross-sections in real time as the log rotated. This leads to the detection of the wood core with up to 100% accuracy and error not exceeding 2mm. The saw cuts the core area into the thinnest sheet, but covers the entire core in the same sheet, increasing the yield of high-quality wood and reducing wood losses and labor to only one person.
“These cameras are a new development in Thailand's rubberwood processing and agro-industries because of the use of artificial intelligence. One key characteristic is that it cannot be substituted by foreign software due to the unique nature of rubberwood which often has a crooked core deviating from the center of the log, is contaminated with latex, has saw marks, and has mold on the cross-section. These issues pose a specific challenge in rubberwood processing in Thailand and cannot be solved with foreign machinery designed for cutting oak or pine with straight core,” said Assoc. Prof. Dr. Wattanapong.
The use of artificial intelligence has improved the efficiency of rubberwood processing. It minimizes waste wood, reduces costs, decreases the need for foreign machinery and computer systems, and strengthens the competitiveness of Thai rubberwood processing plants on a global scale. Additionally, this technology has the potential for further development as a system for detecting eye diseases like strabismus and diagnosing pupil diseases.
Article by Chonthicha Limpiti
Translated by Settaboot Onphakdee
Division of Corporate Communication