Reliability of Noninvasive Sonic Tomography for the Detection of Internal Defects in Old, Large Trees of Abies holophylla Maxim
3.1. Cross-Comparison of SoT and RM Measurements
The results of RM and SoT measurements are presented in Table 1. To verify the reliability of the results, the RM was performed on the same 54 MPs as SoT, and the collected data were analyzed by independent t-test and regression. Figure 1 illustrates the SoT and RM results. The red line indicates the depth of the internal defect (e.g., decay or cavity), and green indicates sound wood in RM results. The defective section (blue) on the tomogram of each of the 54 MPs and the defect lengths measured by RM showed a deviation of approximately 5.6 cm on average. Based on the t-test, the defective area (blue) on the tomogram was 13.2 cm on average, and the defect measured by RM was 11.9 cm on average, with no significant intergroup difference (t = 0.437, p = 0.663) (Table 1).
Figure 1. Sonic tomograms superimposed with the RESISTOGRAPH profile for selected drilling measurement points (MPs). In the sonic tomograms and resistance profiles graphs, green lines indicate the depth of the drilling path and sound region, red lines represent the depth of the decay region, and brown lines represent bark. In the sonic tomograms, brown: healthy woods, red and blue: defects, green: initial decays.
Table 1. Results of t-test on lengths of defects measured with sonic tomography (SoT) and resistance microdrilling (N = 54). MP: measuring point; SoT blue: blue and red color in the SoT tomograms.
The regression analysis for determining the correlations between the two measurement results showed a significant regression model with 67% explanatory power for the changes in RM values with R2 = 0.675 (F = 108.214, p < 0.001; revised R2 = 0.669). The defect lengths by SoT and RM were found to be significantly positively correlated, indicating that an increase in defects detected by SoT meant an increase in defects measured by RM (Figure 2). This suggested that SoT was a reliable method to detect internal wood defects in A. holophylla trees.
Figure 2. Predicted distance of compromised wood using sonic tomography versus the distance of defects using resistography. Trend line = linear regression. p < 0.001, r2 = 0.675, n = 54.
3.2. Comparison of Sonic Tomograms and Cross-Section Images
The SoT results for the five disk samples from three cut trees of A. holophylla are presented in Figure 3. In the tomogram shown in Figure 3a, the decay area is presumed to be increasing towards the pith at 6 and 7 MPs. In the disk image, likewise, the decayed section (red dotted line) is shown to be approaching the pith at 6 and 7 MPs (open cavity). A similar result is shown in Figure 3b,c, where the defect is expanding toward the center from the rings at 8 and 9 MPs in the tomogram. The area was estimated to be a decayed section based on the high percentage of water content in the ERT. As the decay can be seen at the same location on the disk, the tomogram may be taken to indicate the actual defect with similarities, but the estimated values indicated the detection of a larger area of defect. Figure 3d,e show sound wood with some cracks (10 cm in length), a knot, and discoloration. No crack was detected in the tomogram, while such a crack was detected in the RM profile. As a result of quantitatively analyzing the defective area using ImageJ, the error between the defective area by SoT and the actual defective area of the wood section was found to be ±11.3% (Table 2). Specifically in Table 2a, the defective area detected by SoT was 26.5%, whereas the actual defective area was 38.8%, indicating an error of about 12.3%. In Table 2b,c showing the cross section of tree no. 4, an error of about 21–22% was identified. For the cross section of tree no. 4, SoT overestimated the defective area. Figure 3d,e show sound wood with some cracks (10 cm in length), a knot, and discoloration. These results were almost the same as those for SoT, indicating a sound internal wood condition, with few or no internal defects. However, no crack was detected in the tomogram, while cracks were detected in the RM profile.
Figure 3. Comparison of the sonic tomograms and ERT tomograms images and the tree cross section (the red dotted lines indicate the decay area). In the sonic tomograms, brown: healthy wood, red and blue: defects, green: initial decays; in the ERT tomograms, blue: high water content, red: lower water content, green and yellow: decreasing water content.
資料來源:
https://www.mdpi.com/1999-4907/12/8/1131/htm
0 意見:
張貼留言