Non-destructive testing of tree roots has the vast application potential to the root system evaluation of fruit and ancient trees for the better management of plant health. Among them, ground-penetrating radar (GPR) can be expected to non-destructively detect the tree roots, due to the portable and low-cost. However, the interpretation of GPR data is dependent mainly on the manual analysis, leading to the reduced automation and accuracy. Moreover, the clutters and noise that caused by soil heterogeneity can severely affect the accuracy of root detection and identification. In this study, a detection method was proposed for the clutter suppression and root parameter prediction of ground penetrating radar. The data collection was divided into two parts: the dataset and field test. The datasets consisted of simulated and synthesized real data. The roots with the radius ranging from 5 to 30 mm were randomly distributed within 300 mm underground. The field experiment was conducted in three sand pits. The soil heterogeneity was simulated to create the different moisture levels, when adding water into different parts in two of the dry sand pits. The rest sandpit was a condition of higher water content after heavy rainfall. At the same time, the roots were excavated to verify the model under real soil conditions. Four test pits contained the roots of different tree species, the different soil environments, as well as the different radii and depths. The attention mechanism was integrated into the U-net model, in order to enhance the capability in the clutter suppression and root target reflection restoration. A clutter suppression network was constructed to test the effectiveness of the model. The encoding-decoding structure of U-net was separated the target root hyperbolic reflection from the overall background. The attention mechanism enabled the network to better adaptively focus on the hyperbolic reflection of the target root system without identifying the clutter as the target reflection. The clutter suppression was realized for the original B-scan image to remove the adverse effects, due to the soil heterogeneity and radar antenna coupling. Furthermore, the network of root parameter prediction was constructed using residual blocks and inception. The multi-scale receptive field of inception was used to extract the global and local features, and simultaneously to predict the root radius and depth. The predicted results were obtained by parallel inputting the clutter suppressed image and the original B-scan image into the network. The effectiveness of the model was evaluated using datasets and field experiments. The test results of clutter suppression were evaluated using peak signal to noise ratio (PSNR) and structural similarity (SSIM), which were 39.42 dB and 0.991, respectively, better than before. The test results on datasets showed the mean absolute error (MAE) of 1.7 mm, and the coefficient of determination R² value of 0.914 for the root radius prediction, while the mean absolute error (MAE) of 6.3 mm and the R² value of 0.989 for the root depth prediction. The clutter suppression was also achieved in the better performance than the rest on the field data. The maximum prediction errors for the root radius and depth were 1.85, and 13.6 mm, respectively, where the total average relative error was 6.55%. Influencing factors were determined for the prediction effectiveness of ground-penetrating radar. The future research directions were pointed out as well. The findings can be applied to the actual prediction of root depth and radius of tree roots, particularly for the health administration of fruit trees and the protection of ancient trees.protection of ancient trees.protection of ancient trees.

Internal defects in trees, such as cavities and decay, have threatened the safety of heritage trees, due mainly to the mechanical stability and economic value of trees. However, the conventional detection of the internal defects in trees can rely mainly on destructive sampling or low-resolution imaging techniques. Both high accuracy and non-invasiveness are often required in the forestry resource protection and health assessment of heritage trees. Meanwhile, the existing stress wave tomography cannot consider the spatial variability in the velocity information, resulting in some discrepancies between the reconstructed images and the actual defect distribution. In this study, a high-precision tomographic imaging algorithm was proposed to detect the internal defects in trees. The CTIA (confidence tomographic imaging algorithm) integrated the elliptical spatial interpolation into the stress wave propagation paths as elliptical influence zones. Grid velocities were calculated to weight the average ray data within elliptical influence zones. Among them, the confidence metrics were derived from elliptical coverage density, local velocity properties, and geometric configurations. Spatial consistency between adjacent grids was enforced to estimate the iterative neighborhood optimization refined velocity. A weighting scheme was also balanced to combine the current grid confidence with neighborhood velocity standard deviation. To enhance interpretability, A trichromatic visualization scheme was adopted to distinguish healthy tissues, transitional decay areas, and cavities, using normalized velocity thresholds as the classification criteria. In addition, a tradeoff between computational complexity and imaging accuracy was achieved for the stable convergence and reliable reconstruction under varying experimental conditions. Experimental validation was conducted on five log samples, including Ginkgo biloba, Sapium sebiferum, Cinnamomum camphora, Cedrus, and Carya cathayensis, with the moisture contents ranging from 10.5% to 18.3% and defect area ratios from 2.97% to 26.84%. Four simulation samples contained circular defects with the area ratios from 3.08% to 16.21%. Quantitative analysis revealed that there was the CTIA's superiority over EBSI (ellipse-based spatial interpolation) and Intersection IFDD (intersection fitting-based defect detection in woods), with 77.24% for average recall, 81.72% for precision, 79.04% for F1 score, and 70.49% for IOU. Compared with the EBSI, the CTIA improved the average F1 score by 8.69 percentage points and the average IOU by 14.00 percentage points, respectively. Compared with the IFDD, the CTIA was improved by approximately 7.20 percentage points in the precision, indicating the high accuracy and robustness. The absolute relative error in the predicted defect area proportion was only 1.87% among all tests, compared with 2.65% for IFDD and 6.07% for EBSI, demonstrating the enhanced accuracy of CTIA in defect area estimation. A more stable and reliable reconstruction was obtained over different tree species and defect morphologies, indicating its generalization. The irregular defects were also reconstructed in the multi-defect scenarios. Nevertheless, its detection performance declined significantly as the boundary adhesion occurred. Consequently, the internal structures of trees were non-destructively evaluated to dynamically balance the velocity reliability and spatial consistency. A practical tool can offer for the forestry and heritage tree preservation, as well as the ecological and economic challenges in sustainable resource utilization. Adaptive confidence allocation or deep-learning approaches can be incorporated to further improve the performance in complex and multi-defect conditions.

Wood defects have significantly jeopardized the health of trees in the utility value of timber, leading to an immeasurable impact on ancient and venerable trees. Therefore, non-destructive testing (NDT) technology can play a crucial role in the efficient utilization of timber and the conservation of heritage woods. Among them, stress wave testing has been widely applied in recent years, due to its safety, portability, and adaptability to complex environments. However, the conventional stress wave NDT can often fail to consider the influence of wood anisotropy on wave propagation, thereby limiting the imaging accuracy and effectiveness in the detection of internal defects in tree trunks. Furthermore, stress wave data can be collected to uniformly distribute several sensors along the cross-section of the wood under test, in order to accurately map actual cavities. There is also some mismatch between the initial stress wave data visualization and the actual situation in the tree cross-section, due to the uneven speed of stress wave propagation that is caused by wood anisotropy. In this study, an approach was proposed to correct the collected stress wave velocity, and then normalize the corrected velocity using deviation rates. The imaging area was subdivided into multiple grid cells. The speed function of each stress wave ray was refitted, according to the intersection speeds of the rays. A stress wave propagation ray diagram was then drawn using the fitted ray speeds. The speeds of each grid cell were calculated in the imaging area. The nearest neighbor interpolation was used to realize some cells without data. Defect status was determined using the grid cell speeds. The internal defect images of the tree were reconstructed using image processing techniques. Five log samples were tested with the proportion of the defect area after image segmentation and the degree of overlap with the defect shape as evaluation criteria. The results indicate that the imaging algorithm achieved an overall average relative error, accuracy, precision, and recall rate of 8.25%, 93.19%, 80.37%, and 82.30%, respectively. The positions and sizes of defect areas were more consistent with the actual situation. The wave speed model improved the data processing of traditional ones. The robustness against crack interference was also enhanced to identify the crack regions. The findings can greatly contribute to the efficient utilization of timber. A theoretical basis can also provide for the conservation of heritage woods. However, some limitations still remained to detect the micro-cracks and decay on the tomography imaging using intersection fitting, indicating some improvement and optimization. Future research efforts can focus primarily on the robustness and imaging accuracy of tomography imaging using intersection fitting. A solid foundation can also be laid to develop three-dimensional imaging technology.

Edge computing enabled Intelligent Road Network (EC-IRN) provides powerful and convenient computing services for vehicles and roadside sensing devices. The continuous emergence of transportation applications has caused a huge burden on roadside units (RSUs) equipped with edge servers in the Intelligent Road Network (IRN). Collaborative task scheduling among RSUs is an effective way to solve this problem. However, it is challenging to achieve collaborative scheduling among different RSUs in a completely decentralized environment. In this paper, we first model the interactions involved in task scheduling among distributed RSUs as a Markov game. Given that multi-agent deep reinforcement learning (MADRL) is a promising approach for the Markov game in decision optimization, we propose a collaborative task scheduling algorithm based on MADRL for EC-IRN, named CA-DTS, aiming to minimize the long-term average delay of tasks. To reduce the training costs caused by trial-and-error, CA-DTS specially designs a reward function and utilizes the distributed deployment and collective training architecture of counterfactual multi-agent policy gradient (COMA). To improve the stability of performance in large-scale environments, CA-DTS takes advantage of the action semantics network (ASN) to facilitate cooperation among multiple RSUs. The evaluation results of both the testbed and simulation demonstrate the effectiveness of our proposed algorithm. Compared with the baselines, CA-DTS can achieve convergence about 35% faster, and obtain average task delay that is lower by approximately 9.4%, 9.8%, and 6.7%, in different scenarios with varying numbers of RSUs, service types, and task arrival rates, respectively.