Embed Size (px)
Citation preview
Computers and Electronics in Agriculture 108 (2014) 123–129
Contents lists available at ScienceDirect
Computers and Electronics in Agriculture
journal homepage: www.elsevier .com/locate /compag
Application of an ultrasonic wave propagation field in the quantitativeidentification of cavity defect of log disc
http://dx.doi.org/10.1016/j.compag.2014.07.0150168-1699/� 2014 Elsevier B.V. All rights reserved.
⇑ Corresponding author. Tel./fax: +86 451 82191369.E-mail addresses: [email protected], [email protected] (S. Gao),
[email protected] (N. Wang), [email protected] (L. Wang), [email protected] (J. Han).
Shan Gao a, Na Wang a, Lihai Wang a,⇑, Jingquan Han b
a College of Engineering and Technology, Northeast Forestry University, Harbin 150040, Chinab School of Renewable Natural Resources, Louisiana State University AgCenter, Baton Rouge, LA 70803, USA
a r t i c l e i n f o a b s t r a c t
Article history:Received 10 July 2013Received in revised form 7 March 2014Accepted 28 July 2014
Keywords:Ultrasonic wavePropagation fieldCavity defectQuantitative detectionLog section
This study introduced the concept of the ultrasonic propagation field in wood and verified its applicabil-ity through experiments conducted on a log disc. A green Betula costata log section was used in the ultra-sonic propagation measurements. An ultrasonic wave detector (RSM-SY5) was used to measure theultrasonic propagation time (UPT) on the cross section of the log sample. After the ultrasonication wascompleted on the non-defective log, five man-made cavities with diameters of 4, 8, 12, 16 and 20 cmwere made sequentially in the heart of the same log and utilized in the ultrasonic test. The even UPT val-ues were then approximately outlined as isolines. Based on the UPT isolines, the ultrasonic propagationfield in the log disc was simulated using MATLAB software, and the cavity was identified by overlappingthe ultrasonic wave fields from different directions. The size of the cavity defect was proven to be effec-tively determined by overlying the maps of the ultrasonic wave propagation field through a cross sectionof the wood. The results indicated that the accuracy of detection values for cavity sizes with diameters of4, 8, 12, 16 and 20 cm were 83.2%, 84.8%, 94.0%, 95.8% and 96.2%, respectively.
� 2014 Elsevier B.V. All rights reserved.
1. Introduction
Internal wood decay is a common structural defect in trees,logs, utility poles and construction timber. Each year, wood decaycauses significant economic loss (Wiedenbeck et al., 2004). Thus,detecting decay in trees and logs can be of great benefit in assess-ing wood quality and utilizing forest resources (Habermehl, 1982).Foresters can prescribe silvicultural treatments to improve forestmanagement based on the information obtained through decaydetection, thus helping to maintain healthy forests (Wang et al.,2007a,b). The applicability of ultrasonic radiation, an emergingnon-destructive technology, has been widely demonstrated forassessing wood quality and detecting decay (Wang et al.,2009a,b; Gao et al., 2012). The measurement of ultrasonic wavepropagation in wood provides the basis for a nondestructive eval-uation of the wood elastic and viscoelastic properties. The ultra-sonic wave technique for wood quality assessment is based onwave velocity or transmission time measurements in either theradial or tangential direction. This method is typically used as
the primary indicator in examining the conditions of wood quality,grade and defects. (Gao et al., 2013; Wang, 2011).
The direct transmission technique for ultrasonic measurementin non-destructive wood testing is related to the measurement ofthe ultrasonic propagation time (UPT). Wood materials that canbe examined using ultrasonic waves include trees, logs, small clearspecimens of solid wood, wood-based composites and engineeredproducts. Direct transmission is believed to be the most appropri-ate ultrasonic testing technique for quality assessment and decaydetection in these types of wood. Ultrasonic tomography of theimaging technique, which can provide an image of any discontinu-ity, is performed with both velocity (travel time) and attenuationas the contrast-producing parameters (Bucur, 2003; Tomikawaet al., 1990). This technique led to the development of a systemfor detecting heartwood and rotted zones using transducers.
Li et al. (2012) used acoustic tomography and CT to detect inter-nal decay in trees and logs. This method allowed the users to visu-alize the velocity distribution of the acoustic waves as the wavespropagated through the cross section of a tree. Because the ultra-sonic wave propagation in wood is directly related to mechanicalanisotropy, the velocity mapping of a cross section could serve asa diagnostic image to detect internal decay in trees (Bucur,2005). Through velocity mapping via two-dimensional ultrasonictomography, severe decay can be visualized and cavities located(Wang et al., 2001, 2009a,b; Divos and Divos, 2005). Ultrasonic
Fig. 1b. Wave surface of the ultrasonic wave as it travels in the wood.
124 S. Gao et al. / Computers and Electronics in Agriculture 108 (2014) 123–129
wave propagation in wood depends on the grain direction and theexcitation and is associated with the mechanical anisotropy. Theultrasonic variability in a cross section was particularly affectedby the decay or cavities in wooden tree trunks or logs (Larssonet al., 2004).
As an acoustic nondestructive testing method, ultrasonic mea-surement has advantages in the evaluation of wood materials.Based on its applicability and efficiency for assessing wood qualityand detecting decay, this study aims to investigate the behavior ofultrasonic wave propagation in wood when the wave encounters acavity or holes. In addition, the applicability of the ultrasonic wavepropagation field to quantitatively detecting the cavity defects of alog is discussed and demonstrated. Lastly, a reasonable interpreta-tion for the application of the ultrasonic technique to two-dimen-sional decay scanning is provided.
2. Theoretical considerations
2.1. Concept of ultrasonic wave spread field
The natural orthotropic feature of wood is characterized byhighly anisotropic properties. Bulk waves are the principal wavetypes used to measure wood properties and are characterized bythe direction of propagation and the particle motion—i.e., for longi-tudinal waves, the particle trajectory is in the direction of propaga-tion; for transverse waves the particle motion is perpendicular tothe direction of propagation; for Rayleigh waves the particle trajec-tory is elliptical in the plane that is perpendicular to the tested sur-face and parallel to the direction of propagation (McSkimin andChambers, 1964). The ultrasonic velocities of bulk waves can beused to describe the anisotropy within wood in various ways. Forthe direct transmission technique of ultrasonic measurement, anultrasonic pulse is transformed from an electronic signal, travelsthrough the specimen and is received by receiver; the pulse is thentransformed back into an electronic signal, which can be visualizedon an oscilloscope. The time delay, defined as the time that elapsesbetween the emission and reception, is measured on the oscillo-scope over the path length of the ultrasonic signals. The ultrasonicmeasurement principle is simple, and the time measurement isaccurate (with an error below 1%) (Bucur and Bohnke, 1994;Bucur, 1995).
If one assumes that the wave surface of a sound source trans-mits along a specific direction and is conveniently detected bythe sensor in its propagation area, then overlaying the waves fromdifferent directions can yield a packet of ultrasonic wave points atits typical wave surface if the sound source has a limited area
Fig. 1a. Propagation direction of the wave packet sent by a sound source.
(Pellerin and Ross, 2002). Based on the Huygens theorem, the pos-sible interference path directions are shown in Fig. 1a when thesound source is taken as the focus of a series of point wave sources.The wave packet transmits along each direction, but the velocityvector coordinates only with the phase vector (Bodig, 2001). InFigs. 1a and 1b, a sound wave is introduced and spread around;each point of the wave surface in the medium can be taken as asub-wave source of the secondary spherical wave. A number ofpoint sources constitute a plane wave. Therefore, when an ultra-sonic wave transmits in timber, a wave surface will emerge fromthe sound source along the direction of the velocity vector(Fig. 1b). Layers of the wave surface spread apart like the filledcurves from the same UPT of the wave to yield ultrasonic wavespread or a propagation field.
2.2. Applicability of an ultrasonic wave field spread for detecting wooddefects
During the propagation of an ultrasonic wave, the transmissionpaths alter when they encounter different media, and the spreadfield curves with regularity and twists due to variations in theacoustic impedance. By comparing this phenomenon with thespread field for ultrasound wave propagation in solid wood, defectscan be located. By overlaying the profiles with the alteration curvesfrom different directions, the defect edge can be infinitely approx-imated and information, such as the location and size of the defect,can be determined.
3. Materials and methods
3.1. Experimental materials and device
The samples were obtained from Fangzheng Forestry Bureau,Harbin (long. 125�420–130�100E, lat. 44�040–46�400N), China. A50-cm-long Betula costata log section with a 38-cm diameter wasproduced from a standing tree. The log section, with both endscoated, was stored in condition room for two months. The relativehumidity and the temperature of the condition room were at 68%and 20 �C respectively and was then sawn into 10 cm-thick discsin good condition, i.e., with no defects, for use in the ultrasonicmeasurements. The measured density of the non-defect disc was880.99 kg/m3 before it was made into cavity disc. After the ultra-sonic measurements were completed on an intact wood disc, thedisc was immediately made into a cavity defect wood disc withman-made cavities of different diameters (4, 8, 12, 16 and 20 cm)in its heart. Though the wood disc lost moisture during the ultra-sonic measurements, the moisture level was equal to the ambienthumidity. When the whole series experiments were completed
S. Gao et al. / Computers and Electronics in Agriculture 108 (2014) 123–129 125
3 months after ultrasonic test, we oven-dried the disc sample andobtained exact MC value 58.4%. Thus, the MC of disc at the ultra-sonic test is considered in the fiber-saturated state and estimatedover 58.4% to be around 68% due to the moisture lose.
An RSM-SY5 non-metallic ultrasonic detector manufactured bythe Wuhan Institute of Rock and Soil Mechanics, Chinese Academyof Sciences, was used for the ultrasonic testing of the wood disc at20 �C. The ultrasonic measurement system included two ultrasonicprobes for wood testing (one was transmitter to launch the ultra-sonic wave pulse, and the other was receiver to receive the ultra-sonic wave). The instrument parameters for this experimentwere as follows: the sensor of the trigger mode by which the datarecording system was triggered was automatic and independent ofthe received signal; the sampling frequency was 30 kHz; the sam-pling point was 2048 with a sampling period of 1 ls, a pulse widthof 100 ls and highest and lowest values of pass filtering of 10 and1000 kHz, respectively.
Fig. 2b. Testing point distribution on a transverse section.
3.2. Testing procedure
The ultrasonic testing was performed via grid map measure-ments. The UPT tests were conducted on the cross section of thewood disc to develop a map of the cross section and verify theultrasonic propagation field in the wood. The center of the crosssection of the disc was labeled (0,0,0), and the disc was markedwith 4 � 4-cm grids on the top surface (Figs. 2a and 2b). Fourpoints, located 1 cm below the cross section around the trunkwhere the bark had been removed, were designated in evenintervals and numbered using #1, #2, #3 and #4, successively, asthe starting points to fix the transmitter for launching theultrasound wave. The coordinates of the starting points were asfollows: #1, (x,y,z) = (0,�16,�1); #2, (x,y,z) = (�16,0,�1); #3,(x,y,z) = (0,16,�1); and #4, (x,y,z) = (16,0,�1). First, we conductedtesting on the intact disc without cavity defect in it. In the firsttesting turn on the intact disc, the transmitter was fixed at point#1 position, and then tapped the receiving probe (the receiver) intoeach grid intersection and pointed toward the launch probe (thetransmitter at point #1) with a 45� inclination to the disc surface.The UPT between receiver at each grid point and transmitter atpoint #1 were measured. After completing all grid intersectionultrasonic tests matching the #1 interval on the non-defective disc,we conducted second turn testing on the intact disc. We re-fixed
Fig. 2a. Schematic diagram of the UPT tests on the thick wood disc.
the transmitter at point #2 position and tapped the receiver intoeach grid intersection and pointed toward the transmitter at #2with a 45� inclination to the disc surface. The UPT between receiverat each grid point and transmitter at point #2 were measured. Sim-ilarly, the next measurements were conducted by fixing the trans-mitter at point #3 and point #4 to finish the test on the intact discsample.
After tests on intact disc, we then made 4 cm diameter cavity onthe same disc to use for the UPT test on decay sample. Following,similar tests were performed on this cavity size defect disc. Inthe first testing turn on 4-cm cavity disc, we fixed the transmitterat point #1 position, and then tapped the receiving probe (thereceiver) into each grid intersection and pointed toward the trans-mitter with a 45� inclination to the disc surface. The UPT betweenreceiver at each grid point and transmitter at point #1 were mea-sured. After completing all grid intersection ultrasonic tests match-ing the #1 interval, we conducted next testing turns by fixing thetransmitter at point #2, #3 and #4 in series. At each testing turn,the UPT between receiver at each grid point and transmitter atwere measured on the 4 cm cavity disc.
Same testing procedures were repeated on the other decaydegree disc with the different man-made cavity diameter of 8,12, 16 and 20 cm.
4. Results and analysis
4.1. Computed ultrasonic wave propagation field in non-defectivewood
Table 1 lists the UPT data measured on a non-defective discwith the transmitter placed at point #1. Sixty-seven tests areordered by the UPT value. All tested coordinates with approxi-mately even UPT values are lined as isolines with an intervalbetween isolines of 20 ls. According to the testing data, the UPTisolines from point #1 for the non-defective sample were gener-ated using MATLAB software and are illustrated in Fig. 3a. The plateof the curves that fit the isolines (from launch point #1) was sim-ulated via interpolation and is illustrated in Fig. 3b. As previouslystated, the cures plate that yields the UPT isolines distributionwas defined as the propagation field of the ultrasonic wave.
Table 1UPT values in a non-defective log section with launch probed at point #1.
Receiving probe locationcoordinates
UPT(10�6 s)
Receiving probe locationcoordinates
UPT(10�6 s)
Receiving probe locationcoordinates
UPT(10�6 s)
Receiving Probe locationcoordinates
UPT(10�6 s)
(8,�16) 145 (�4,�8) 153 (0,0) 182 (4,8) 280(4,�16) 103 (�8,�8) 173 (�4,0) 212 (0,8) 260(0,�16) 78 (�12,�8) 201 (�8,0) 228 (�4,8) 280(�4,�16) 102 (�16,�8) 276 (�12,0) 255 (�8,8) 281(�8,�16) 150 (16,�4) 266 (�16,0) 271 (�12,8) 293(�12,�12) 189 (12,�4) 235 (16,4) 308 (�16,8) 321(�8,�12) 164 (8,�4) 196 (12,4) 286 (12,12) 339(�4,�12) 138 (4,�4) 180 (8,4) 265 (8,12) 327(0,�12) 95 (0,�4) 153 (4,4) 250 (4,12) 314(4,�12) 132 (�4,�4) 184 (0,4) 229 (0,12) 299(8,�12) 166 (�8,�4) 203 (�4,4) 244 (�4,12) 315(12,�12) 190 (�12,�4) 228 (�8,4) 246 (�8,12) 326(16,�8) 265 (�16,�4) 254 (�12,4) 269 (�12,12) 338(12,�8) 201 (16,0) 282 (�16,4) 295 (4,16) 351(8,�8) 173 (12,0) 256 (16,8) 326 (0,16) 345(4,�8) 155 (8,0) 238 (12,8) 299 (�4,16) 349(0,�8) 126 (4,0) 216 (8,8) 287
Fig. 3a. UPT isolines of the ultrasonic wave in a non-defective disc.
Fig. 3b. Simulated propagation field of an ultrasonic wave in a non-defective disc.
126 S. Gao et al. / Computers and Electronics in Agriculture 108 (2014) 123–129
As illustrated in Fig. 3a, launch point #1 is a sound source. Theultrasound wave was launched and spread apart. The overall wavepropagation was symmetrically distributed with the line x = 0(Fig. 3a) and maintained its propagation tendency. The curvatureof the plots continually increased due to the difference betweenthe wave velocities in the radial and tangential directions. Theultrasonic wave propagated more rapidly in the radial directionthan in the tangential direction, verifying the analysis of the prop-agation field provided in the theoretical portion of this study.
4.2. Impact of cavity size on the ultrasonic wave spread
Based on the testing data for the cavity defect disc, the distribu-tion of the UPT isolines generated from point #1 is illustrated inFig. 4a, b, c, d and e with corresponding cavity diameters of 4, 8,12, 16 and 20 cm, respectively. The layout of the ultrasonic wavepropagation field in the defect disc sample differed significantlyfrom that of the non-defective sample (Fig. 3a). Fig. 4 indicates thatthe size of the cavity defect had a significant impact on the ultra-sonic wave propagation path and led to a different form of propa-gation field. When the ultrasonic wave propagated along the x = 0direction at the cross section of the log, the UPT isolines demon-strated a tendency toward inward contraction close to y = 0 beforethe wave passed through cavity in the center.
Because the wave had already passed through the cavity, theisoline curves extended in the inverse direction, indicating thatthe radial velocity of the wave propagation decreased when thewave bypassed the center cavity, but the tangential velocityincreased. When observing the entire propagation field, a ‘‘laggingbehind’’ phenomenon was found to be led by the cavity. Theincreases in the UPT time were caused by changes to the propaga-tion path of the ultrasonic wave due to the encountered cavity. Asthe diameter of the cavity increased, the ‘‘lagging behind’’ phenom-enon became more and more remarkable. The degree of curvaturefor the isolines behind the cavity increased, and the value of theradial transmission time increased as the diameter increased.
4.3. Identifying the two-dimensional defects using the wavepropagation field
The defect information can be reflected by the superimposedimage in various directions to intuitively determine the position,size and shape of the defect in a simple. Therefore, a two-dimen-sional image of a cavity in a log section was identified by overlay-ing the ultrasonic wave spread field, which was composed of
(a) Cavity diameter of 4 cm (b) Cavity diameter of 8 cm
(c) Cavity diameter of 12 cm (d) Cavity diameter of 16 cm
(e) Cavity diameter of 20 cm
Fig. 4. Simulated isolines of the UPT in a log with different sizes of man-made cavities. (a) Cavity diameter of 4 cm. (b) Cavity diameter of 8 cm. (c) Cavity diameter of 12 cm.(d) Cavity diameter of 16 cm. (e) Cavity diameter of 20 cm.
S. Gao et al. / Computers and Electronics in Agriculture 108 (2014) 123–129 127
isolines generated from four different propagation directions(Fig. 5). The UPT test data were programmed into images of thewave propagation fields launched from points #1, #2, #3 and #4using MATLAB software. The two-dimensional identificationimages for the five types of hole diameters are illustrated inFig. 5a, b, c, d and e, respectively. The defect information wasobtained by overlaying the propagation fields from the differentlaunch points. In the log section with a cavity diameter of 12 cmfor example (Fig. 5c), the grid lines crisscross in the center of the
log cross section to mark the location of the defect or cavity. Thedefect regions were identified by imposing the distorted propaga-tion field that occurred when the ultrasonic wave encountered avoid, hole or cavity. Therefore, the defect boundary can be identi-fied by determining the curve boundary of the deformation fieldfor the ultrasonic wave propagation in each direction. In contrastto the wave propagation field of the non-defective log, the interme-diate point of the deformation curve for the ultrasonic propagationin a defective disc was selected as the edge of the defect area. The
(a) Cavity diameter of 4 cm (b) Cavity diameter of 8 cm
(c) Cavity diameter of 12 cm (d) Cavity diameter of 16 cm
(e) Cavity diameter of 20 cm
Fig. 5. Image created by overlapping the ultrasonic wave propagation field in a log section with cavities of different diameters. (a) Cavity diameter of 4 cm. (b) Cavitydiameter of 8 cm. (c) Cavity diameter of 12 cm. (d) Cavity diameter of 16 cm. (e) Cavity diameter of 20 cm.
128 S. Gao et al. / Computers and Electronics in Agriculture 108 (2014) 123–129
defect area of a cavity was outlined by linking the edge points foreach propagation direction. Consequently, the size and location ofthe defect were successfully identified. The ultrasonic wave prop-agation filed was theoretically explained by the series depicted inFig. 5.
The area of the cavity estimated by ultrasonic wave propagationfiled was calculated from sum of grid square area constructed bygrid lines crisscross in the center. The edge of overall grid squarearea by overlaying the ultrasonic wave spread field was circled
using simulation program of MATLAB. Thus the covered area of cir-cle was used to calculate estimated area of the cavity detected.
To quantitatively analyze the accuracy of this identificationmethod, the error of estimation was analyzed by comparing theactual size of the cavity with that detected by applying the ultra-sonic wave propagation field. The term S0 represented the esti-mated area of the cavity detected, while S1 represented the actualarea, and D represented the error of estimation for the cavity. Theformula used to calculate the error of estimation is as follows:
Table 2Comparison of the actual and predicted defect areas.
Cavitydiameterd (cm)
Predicted defectarea S0 (cm�2)
Actual defectarea S1 (cm�2)
ErrorD (%)
Degree ofaccuracy D0 (%)
4 14.69 12.57 16.8 83.28 57.89 50.24 15.2 84.8
12 106.31 113.10 6.0 94.016 192.62 201.06 4.2 95.820 326.61 314.15 3.8 96.2
S. Gao et al. / Computers and Electronics in Agriculture 108 (2014) 123–129 129
D ¼ S1 � S0
S1
����
����� 100%
The degree of accuracy in the prediction can be derived asfollows:
D0 ¼ ð1� DÞ ð%Þ
For each cavity diameter, D0 is listed in Table 2. As the size ofthe cavity defect increased, the prediction became more accurate.
The accuracy of the detect results was much greater (over 90%)for cavities with diameters over 12 cm, proving that the ultrasonicwave propagation field provided an adequate method for identify-ing the cavity defect area and providing a rough sketch, and, in asense, interpreting the law of ultrasonic wave propagation in a cav-ity defect log.
5. Conclusions
This study introduced the concept of the propagation field of anultrasonic wave in a log and verified its applicability for detectingcavity defects. The conclusions can be stated as follows:
1. The ultrasonic wave field composed of UPT isolines was simu-lated using the abovementioned programs, and the location,size and boundaries of the cavities were successfully identifiedby overlaying the ultrasonic wave fields generated from indi-vidual sound resources.
2. The UPT isolines were symmetrically distributed along theultrasonic wave propagation path. Behind the cavity defect, atendency toward inward contraction was observed, and thecurve was higher on its two ends, but lower in the middle.The entire propagation field exhibited a ‘‘lagging behind’’behavior initiated when it encountered the cavity.
3. Based on the concept of an ultrasonic propagation field, thedetection accuracy for cavities with diameters of 4, 8, 12, 16and 20 cm was 83.2%, 84.8%, 94.0%, 95.8% and 96.2%, respec-tively. The degree of accuracy for the cavity identificationincreased with increasing cavity size. Highly accurate detectionresults (above 90%) were obtained for cavities with diametersover 12 cm.
Acknowledgements
This study was supported by the China Special Research Fund-ing Program for Forestry Public Service Sector (Grant No.201104007) and supported in part by the Fundamental ResearchFunds for the Central Universities (Grant No. DL12BB31). Theauthors thank Mr. Shiquan Song for his assistance with the testing.
References
Bodig, J., 2001. The process of NDE research for wood and wood composites. WoodSci. Civil Eng. 3.
Bucur, V., 1995. Acoustics of Wood. CRC Press Inc., New York.Bucur, V., 2003. Nondestructive Characterization and Imaging of Wood. Springer,
Berlin, Heidelberg, Germany.Bucur, V., 2005. Ultrasonic techniques for nondestructive testing of standing trees.
Ultrasonics 43 (4), 237–239.Bucur, V., Bohnke, I., 1994. Factors affecting ultrasonic measurements in solid wood.
Ultrasonics 32 (5), 385–390.Divos, F., Divos, P., 2005. Resolution of stress wave based on acoustic tomography.
In: Proceedings of the 14th International Symposium on Nondestructive Testingof Wood. Eberswalde, Germany, 309–314.
Gao, S., Wang, X., Wang, L., Allison, R.B., 2012. Effect of temperature on acousticevaluation of standing trees and logs: Part 1 – laboratory investigation. WoodFiber Sci. 44 (3), 286–297.
Gao, S., Wang, X., Wang, L., Allison, R.B., 2013. Effect of temperature on acousticevaluation of standing trees and logs: Part 2 – field investigation. Wood FiberSci. 45 (1), 15–25.
Habermehl, A., 1982. A new nondetermining internal wood condition and decay inliving trees. I. Principles, method and apparatus. Arbor J. 6, 1–8.
Larsson, B., Bengtsson, B., Gustaffson, M., 2004. Nondestructive detection of decay inlivingtress. Tree Physiol. 24, 853–858.
Li, L., Wang, X., Wang, L., Allison, R.B., 2012. Acoustic tomography in relation to 2Dultrasonic velocity and hardness mappings. Wood Sci. Technol. 46, 551–561.
McSkimin, H.J., Chambers, R.P., 1964. Measuring Mechanical Properties of Plasticswith High-frequency Ultrasound. Institute Electrical and Electronic Engineers,Incorporated, pp. 11.
Pellerin, F.R., Ross, J.R., 2002. Nondestructive evaluation of wood. Forest Prod. Soc.,135–149.
Tomikawa, Y., Iwase, Y., Arita, K., Yamada, H., 1990. Nondestructive inspection ofwooden poles using ultrasonic computed tomography. IEEE Trans. UFFC 33 (4),354–358.
Wang, X., 2011. Fundamentals of acoustic measurements on trees and logs and theirimplication to field application. Pages 25–33. In: Divos, F. (Ed.), Proc 17thInternational Nondestructive Testing and Evaluation of Wood Symposium,September 14–16, 2011, Sopron, Hungary. University of West Hungary, Sopron,Hungary.
Wang, X., Ross, R.J., McClellan, M., Barbour, R.J., Erickson, J.R., Forsman, J.W.,McGinnis, G.D., 2001. Nondestructive evaluation of standing trees with a stresswave method. Wood Fiber Sci. 33 (4), 522–533.
Wang, X., Carter, P., Ross, R.J., Brashaw, B.K., 2007a. Acoustic assessment of woodquality of raw forest materials—a path to increased profitability. Forest, Prod. J.57 (5), 6–14.
Wang, X., Ross, R.J., Carter, P., 2007b. Acoustic evaluationof wood quality in standingtrees. Part I, Acoustic wave behavior. Wood Fiber Sci. 39 (1), 28–38.
Wang, L., Li, L., Qi, W., Yang, H., 2009a. Pattern recognition and size determination ofinternal wood defects based on wavelet neural networks. Comput. Electron.Agric. 69, 142–148.
Wang, X., Wiedenbeck, J., Liang, S., 2009b. Acoustic tomography for decay detectionin black cherry trees. Wood Fiber Sci. 41 (2), 127–137.
Wiedenbeck, J., Wiemann, M., Alderman, D., Baumgras, J., Luppold, W., 2004.Defining hardwood veneer log quality attribute. Gen Tech Rep NE-313. USDAForest Service, Northeastern Research Station. p. 36.
