Harvesting trees that contain internal defects such as knots and cracks are neither financially nor environmentally sustainable. In hardwood plantations, it is impossible to produce sawlogs from knotty or cracked timber. The challenge is to identify internal defects in a timely and cost-effective manner prior to harvesting. The aim of this paper is to investigate non-destructive testing (NDT) methods to rapidly detect the presence of internal defects in standing live trees in plantation plots. The study highlights that whilst several methods exist, few have been actively applied in-field harvesting operations to optimise log handling and to increase transportation efficiencies. Key constraints are portability of the NDT equipment for use in-field, speed versus accuracy of measurements undertaken and the usability of different evaluation approaches for decision-support. In this paper, the field assessment involved using two non-destructive techniques, ground penetrating radar (GPR) and ultrasonics that use electromagnetic and ultrasonic sound waves respectively to penetrate the internal structure of standing trees. These assessment techniques can assist forest growers to more accurately evaluate the quality of growing stems in the field. They also open the opportunity to investigate differences across a wide selection of growing conditions and forest types to generate data that may support the generation of a software algorithm for predictive imputation of likely internal defect rates within particular forests and under particular growing conditions. The plan being to integrate this predictive imputation software into existing geographical information systems owned by industry partners to enable accurate mapping of land areas where high ratios of defects are likely to be detected to further optimise infield harvesting.
Unmanned Aerial Systems (UAS) are a cost-effective means of collecting forest data conventionally used above the forest canopy. Where forest canopies are dense, limited information about stem structures can be extracted directly due to obscuration by foliage. In these circumstances, complementary ground-based methods including manual measurement and terrestrial laser scanning are deployed, but these techniques are often limited in terms of the scope and scale of data collected by factors including time, field cost and site accessibility. This paper describes the application of a UAS flown below the forest canopy as an efficient and effective approach for stem measurement in areas where the canopy is difficult to penetrate, and as a potential solution to measuring trees in areas of dense undergrowth. The study sites were scanned with a helicopter-mounted VUX-1LR LiDAR sensor and the resulting point clouds were used as a comparison dataset. The measurements extracted from these point-clouds were compared with ground-based measurements of diameter at breast height and relative positions. The below-canopy UAS and the VUX-1LR at 30m had the lowest root-mean-squarederror (RMSE) of 4.1cm, followed by the VUX-1LR at 90m with a RMSE of 4.4cm. The VUX-1LR 60m flight was the most consistent with the highest coefficient of determination, however due to a positive bias, there was an RMSE of 4.5cm. The photogrammetry-based, below-canopy UAS was found to be an efficient and accurate method of extracting DBH and relative position of stems in forests.
Tasmania eucalyptus nitens is one of the most important plantation hardwood species used for paper production. Forest growers and wood processing companies have recently considered it for the production of high quality sawlog. The high quality sawlog, however, can be produced from pruned plantation eucalyptus niten as the unpruned one contains several knots and cracks which lessen the quality of the log. Thus, it is vital for forest growers to deliver pruned log to wood processing companies. The pruned log, however, could not be discriminated from unpruned stems by harvester within the plantation plot due to self (natural) pruning process of unpruned tree. This leads to the delivery of the pruned log to the processors challenging. Although wood processors use large x-ray image machines during processing to optimise wood recovery, high costs are incurred from transporting poor quality, knotty timber following the harvest. In this paper, a 17 year old eucalyptus nitens has been considered for non-destructive evaluation. The aim is to investigate the effects of the defects including knots and cracks on the ultrasonic wave. 12 samples from different parts of trunk have been selected and conditioned at the forest moisture content of 120% (70% water content). The samples were scanned by ultrasonic waves at every 10 cm distance in longitudinal direction and at every 45 degree spacing in circumferential direction along the samples. Results show that there is a significant difference between recorded ultrasonic waveforms propagated through unpruned billets and pruned ones. The unpruned billets had a larger effect on ultrasonic waves while the waves are relatively steady when pruned billets are tested.