Note: Descriptions are shown in the official language in which they were submitted.
2/27/91 ~ ~ 2 ~ 7
REFLECTIVE GRAIN DEFECT SCANNING
Field of the Invention
The present invention relates to wood product proce~sing and
particularly to a method and apparatus for detection of grain
de~ects by re~lective s~nn;ng.
Background of the Invention
Automatic detection of grain defects improves wood processing
operations. Overall production efficiency and product quality can
be increased by automatic grain defect detection and corresponding
product grading, processingl or remedial action. Unfort~mately,
existing defect sc~nn;ng techniques have been complex and,
therefoxe, not always availa~le where ~he benefit o~ automatîc
grain defect detection does not outweigh the expense of such
sc~nning techniques. For example, grain de~ect scanning of plywood
panels or solid lumber would improve production, but cannot be
accomplished by through-grain scanning techniques and must be
accomplished by relatively more complicated and expensive
2Q reflective scanning methnds. Grain defect scanning of panels and
solid wood material has not been regarded as feasible because of
the complexity and expense o~ heretofore available reflective
scanning technologies.
2 ~ & ~ 7
A simple and relatively low cost reflective grain defe~t
scanning method and apparatus would make available the benefits of
automated grain defect detection in the production of, for example,
plywood panels and hardwood stock.
summarY of the Invention
In accordance with the present invention, the surface of a
woodgrain article is characterized by computing a ratio of specular
reflection to diffuse reflection as measured from separate view
angles of a point of incidence o~ a scanning energy beam. A grain
defect is indicated by similar ensrgy reflection magnitudes while
clearwood is in~icated by a substantially greater specular
re~lection relative to the dif~u e reflection.
Grain defe~t scanning in accordance with a preferred
embod; ent of the present invention is accomplished by scanning of
an inspection point illuminated by a collimated light beam directed
toward an inspection surface at a given angle of incidence. A pair
of light detectors provide simultaneous views of the point of
incidence as represented by reflected light intensity. One
detector, the specular dete~tor, is positioned along the specular
angle of reflection as defined by the angle of incidence with
respect to the inspection sur~ace. The other detector, the diffuse
detector, lies substantially along the angle of incidence.
Specular reflection from th~ inspection point dominates when the
inspection point corresponds to clearwood. In such case, the
' 2Q~2~
ecular detector indicates a higher reflective li~ht intensity
than the diffuse detector. Diffuse reflection dominates when the
inspection poi~t corresponds to a grain defect. Both detectors
indicate similar reflective light intensity for diffuse liyht
reflection. Grain defect dis~rimination is accomplished by
calculating a ratio o~ specular dete~tor output to diffuse detector
output whereby a ratio substantially greater than unity indicates
clearwood and a ratio close t.o unity i~dica-tes a grain defect at
the inspection point.
Brief Descri~tion of the Drawinqs
FIG. 1 is a perspective view of a reflective grain defect
apparatus according to a f irst embodiment of the present invention
and a woodgrain article subject to grain defect inspection.
FIG. 2 is a side view of the apparatus and woodgrain article
taken along lines 2-2 of FIG. 1.
FIG~ 3 is a reference syste~ used to illustrate two light
re~lection models believed to account ~or light re~lection from a
woodgrain ~urface.
FI~. 4 illustrates in perspective a diffuse light reflection
model.
2S
FIG. 5 plots re~lected light intensity according to the
7:
ffuse light reflect.ion model.
FIG. 6 illustrates in perspective a specular di~fuse li~ht
re~lection model.
FIG. 7 plots reflect~d light intensity according to the
specular di~fuse light re~1ection model.
FIG. 8 ~lots reflacted light intensity ~or clearwood and for
grain defects and illustrates the ~ethod o~ distinguishing
therebetween in accordance with the present invention~
FIGS. 9 and 10 illustrate side and top views, respectively,
of a second embod.iment o~ the present invention implemented as an
across ~rain line scanning device ~or longitudinal matexial feeding
applications.
FIG. 11 is a flow chart illustrating steps associated with
operation of the d~vice of FIGS. 1 and 2.
FIG. 12 illustrates a modi~ication for more accurate detection
o~ knotwood dimensions applicable to the embodiments o~ the present
invention illustrated herein.
Detailed Description of the Invention
FIGS. 1 and 2 illustrate a ~irsk embodiment of th~ presen~
2 ~ 7
vention, a grain defect detection apparatus 20 using reflective
scanning~ In FIGS. 1 and ~, an elongate woodgrain artisle 12,
having a longitudinal axis 1~ and inspection surface 1~, is ~ubject
to grain defect scanning at an inspection point 18 of surface 16~
As used herein, the term "inspection point" ~hall refer to a point
of incidence of s~anning light directed Upon an inspection sur~ace,
not necessarily a static point on the inspection sur~ace. ~rticle
12 has wood fiber lying substantially along it~ longitu~;nal axi5
14, but may have grain defects or gross deviations from ths normal
direction o~ the wood ~iber, e.g., kn~twood. Grain defect
detection d~vice 20 identifies grain de~ects at the inspection
point 18 of article 12. As discussed more fully below, translation
of article 12 relative to apparatus 20 and along longitudinal axis
14, i.e. as in lonyitudinal material feeding applications, provides
grain defect scAnn;ng along a line 18' corresponding to a plurality
of inspection points 18 as defined by such longitudinal movement.
The embodiment of FIGS. 1 and 2 i5 essentially a single point
or pixel based device adapted for relatively narrow scanning width
applications, but illustrates the basic opera~ional principles of
the present invention. A second embodiment o~ the present
inventiQn, described below, uses an across grain linescA~ning
techniquQ and arrays o~ light detectors to implement a more
2~ practical, i.e., broader width, grain de~ect scannlng application
in accordance with the present invention.
~2~
Grain defect detection apparatus ~0 includes a pair of light
detection devices, detectors 22 and 2~, each bearing upon
inspection point 18 and lying along the line of material feed,
i.e., in the plane of incidence 25 orthogonal to the surfaca 16 and
containing the line 18'. Detectors 22 and 24 lia symmetrically
within plane of incidence 25. The lines o~ sight to inspection
point 18 ~or each detection device are symmetric about a vertical
ref~rence axis 30 which is within plane of incidence 25, normal to
surface 16, and coincident with the point 18. A light source 26
directs a collimated light beam 27, e.g. a low power laser beam,
toward point 1a and substantially, as close as possible, along the
line of cight betwesn detector 24 and inspaction point 18. Each
detector 22 and 24 produces an output signal representative of a
level of re~lected light nergy detected. A discrimination circuit
23 rPceives the output from de ectors 22 and 24 to identify.grain
defects a~ a given inspection point 18. Discrimination circuit 28
may be a general purpose computer or dedicated circuitry adapted
in conventional manner to practice the present invention as
described herein.
The present invention is b~tter understood with refe:rence to
two light reflection models helieved to represent components of
light reflected from a woodgrain surface. FIG. 3 provides a
re~erence system, similar in arrangement to that of apparatus 20,
useful in illustrating the light reflection modals. Beam 27 is
, CA 02062447 l998-0~-l3
represented by its angle of incidence 32 relative to axis 30,
and, unless otherwise stated, remains fixed. Reflective light
intensity is represented as would be detected by a movable
light detector 36 maintained within plane of incidence 25 at a
given distance from inspection point 18 and expressed as a
function of angular position 34 relative to axis 30. Plotting
the output of detector 36, for a fixed angle of incidence 32,
against a range of angular positions 34 illustrates the
character of each light reflection model.
The first light reflection model, illustrated in
FIG. 4, represents diffuse reflection where light reflects in
all directions from the point of incidence. FIG. 5 plots
reflective light intensity according to the diffuse reflection
model for detector 36 having angular positions -90 degrees
through 90 degrees. The plot function, I=Iocos~ is an evenly
contoured response, symmetric about reference axis 30, and
having maxima normal to surface 16, and coincident with axis
30. It is noted that purely diffuse reflection is independent
of the angle of incidence 32 for beam 27. Returning briefly
to FIGS. 1 and 2, according to the diffuse reflection model
detectors 22 and 24, being symmetric about axis 30, will
register substantially equal light intensity with respect to
diffuse reflection.
The second light reflection model, illustrated in
FIG. 6, iS termed "specular diffuse", meaning that some of the
reflecting light acts as though it undergoes specular or
mirror-like
71208-66
~0~2~17
,flection while the other components act as though it reflects
from a diffuse surface. In FIG. 6, diffuse reflection component
emanates from inspection point 18 according to the above
described diffuse reElection model. Specular reflection component
5 42, however, emanates from inspection point 18 along the sur~ace
of a cone 44 having a central axis collinear with line 18' and a
half angla (9~-i) relative to surface 16 where i e~uals the angle
of incidence 32. Accordingly, the specular component 42 de~ines
a semi-circular arc 45 within a plane 47 orthogonal to sur~ace 16
and to plane of incidence 25. The cone-like shape of the specular
component 42 is believed to result from surface 16 having
substantially regular sur~ace contour along a first climension
parallel to longitudinal axis 14, i.eO, along the length of ~iber
cells, and an irregular surface contour in a second orthogonal
dimension transverse to axis 14, i.e, transverse to the fiber
cells. The sp~cular diffuse component 40 of reflected light is
believed to be caused primarily by the cellulose fibers in the cell
walls, while the dif~use component i5 caused primarily by the
remaining cell structure, cavities, resin, etc.
FIG. 7 plots light intensity according to the specular'diffuse
model for a fixed angle of incidence 32 and a range of detector 36
positions from -90 degrees through 90 degrees. In FIG. 7, the
dif~use component 40 appears, in accordance with the diffuse
re~lection model, as an evenly contoured response symm~tric about
axis 30. The specular component 42 appears as a more narrow
2 O 5 ,5~ q ~
sponse cen~ered about a specular reflection angle 46 lying along
the surface of cone 4~. Note that the specular reflection angle
46 is equal in magnitude to the incident angle 36. The composite
reflection response ~8, according to the specular diffuse
reflection model~ generally follows the diffuse component 40, but
has a characteristic maximum centered about the specular re~lection
angle 46.
Experimentation indicates that light reflecting off a
woodgrain article behavss according to a combination of the diffuse
reflection model and the specular diffuse reflection model. More
importantly, experimentation has shown that specular di~~use
reflection dominates for clearwood reflections and dif~u~e
re~lection dominates for grain defect reflectionsO Grain defect
detection according to the present invention discriminates between
specular diffuse reflection and dif~use reflection in order to
distinguish clearwood and grain defects. Accordingly, grain de~ect
detection is accomplished by discriminating between the
substantially symmetric response o~ FIG. 5 and the asymmetric
response of FIG. 7.
FIG. 8 plots a typical clearwood function 50 and a typical
grain def~ct ~unction 52. More particularly, function 50
illustrates d~tector 36 output~ Por an inspection point 18
corresponding to a clearwood portion oP surface l.6 while function
5~ illustrates detector 36 output ~or an inspection point 18
2 i.~
rresponding to a grain de~ect por~ion, e.g., kno~wood, of surface
16. The difference between functions 50 and 52 provides a basis
for discrimination by apparatus 20 clearwood v~r~us grain defect
wood.
With reference to FIGS. 1, 2 and 8, detectors 22 and 24 of
apparatus 20 are placed symmetrically relative to axis 30 and light
beam 27 is directed along an incidence angle 54 substantially
collinear with the line of sight from detector 22 to inspection
point 18. The detector 22 lies on the specular re~lection angle
46, i.e., along the surface o~ cone 44 o~ the specular diffuse
re~lection model. This arrangement provides two separate angular
views of the inspection point 18~ Detector 22, the specular
detector, is positioned to detect the ~;mll~ associated with the
specular reflection component 42 of the specular dif~use re~lection
as well as di~use reflection. The diffuse detector 24 is
positioned to monitor diffuse reflection only.
W~th detector 22 lying on the angle 46, detector 22 output
indicates a light intensity I1 for cleaxwood and a light intensity
I~ ~or grain defects. Similarly, with detector ~4 lying on axis
54, detector 24 indicates a light intensity I3 for clearwood and
a light intensity I4 for grain de~ects. $he light intensity values
Iz, I3 and I4 are substantially equal while the inten~ity value I1
is relatively greater. Measuring the relative output magnitudes
of detectors 22 and 24 provides a basis for discriminating between
2~2l~7
~ earwood and yrairl defect wood. In particular, the discrimination
~unction is expressPd as the ratio of detector 22 output to
detector 24 output where a result substantially greater than unity,
e.g., I1 divided by I3, indicates clearwood, and a result
substantially near unity, e.g. I2 divided by I4, indicates grain
defect wood. Discrimination circuit 28 divides detector 22 output
magnitude by detector 24 output magnitude~ For a result
substantially greater than unity, discri.mination circuit 28
identifies the inspection point 1~ as corresponding to clearwood,
and for a result substantially near unity discrimination circuit
28 identifies the inspection point 18 as corresponding to a grain
defect. Typical clearwood yields a ratio of approximately 1.8 to
2 while knotwood yields a ratio near unity. It may be appreciated
that discrimination is based on the relative output magni.tucles of
detectors 22 and 24, not absolute output magnitudes. ~ccordingly,
variations between woodgrain patterns of individual wood articles
in~p~cted, resulting in different absolute levels of reflected
light for individual articles, are substantially masked.
As will be apparent to those skilled in the art, by suitably
indexing the position of article 12 relative to apparatus 20 and
collecting detec~or 22 and 24 data ~or each in~exed position it is
possible to gather suf~icient data to characterize surface 16 as
to grain dafects along the line 18'. In FIGS. 1 and 2, indexing
rollers 70 contact the upper sur~ace 16 and lower sur~ace oP
article 12 and roller control 72 moves article 12, by way of
2~2l~7
llers 70, in indexed fashion while providing article 12 position
data to discrimination circuit 28. Discrimination circui 28 may
then associate a physical location on surface 16 with the
inspection point 18 for each indexed position of article 12. By
such association, the location of detected grain defects may be
specified using apparatus 20. Multiple scanning passes across
dif~erent portions of surface 16 would prsvide scanning of the
entire surface 16. Discrimina~ion circuit 28 then constructs a
data representation 74 o~ surface 16 as output useful in subsequent
wood processing operatiQns to ~;r; ze use o~ article 12. For
example, article 12 may be cut into smaller dimension products
using knowledge o~ grain de~ect positions so as to maximize stress
capabilities of the resulting product as by avoiding grain defects
near product edges.
While the embodiment of FIGS. 1 and 2 provides an accurate
method of grain defect detection, this embodiment may not be
practical in most wood processing operations.
FIGS. 9 and 10 illustrate side and top views, respectively,
of a second more practical embudiment of the present invention
imp}emented as an acro~s grain linescanning device for longitudinal
material ~eeding applications. In FIGS. 9 and 10, an across grain
linascanning apparatus 100 includes detector arrays 122 and 124,
each including corresponding individual detec~ors 122a-122c and
124a-124c, respectively. In the illustrated embodiment, each of
12
20~2;~7
.ays 122 and 1~4 contains three detectors, but it should be
apparent how the present invention may be practiced using arrays
122 and 12~ with more or less than three detectors. As described
more fully below, apparatus 100 defines an across grain inspection
line 118 comprising contiguous transverse inspection sections 118a-
118c. Apparatus 100 detects grain defects at surface 116 of a
woodgrain article 112 by movement of article 112 along i-ts
longitudinal axis ~14 relative to apparatus 100 as by indexing
rollers 170 and roller con~rol 172. Roller control 172 provides
to discrimination circuit 128 the article 112 position data whereby
circuit 128 may associate a physical location on surface 116 for
line 118 to each indexed position of article 112.
Corresponding detectors of arrays 122 and 124, e.g., detectors
122a and 124a, lie within sepaxate and parallel planes o~ incidence
(not shown) each orthogonal to the surface 116. In other words,
each set of corresponding detector pairs is similar in arrangement
and operational relation to that of detectors 22 and 24 ~FIGS. 1
and 2). The output signals from corxesponding detectors of arrays
122 and 124 are compared as described above for detectors 22 and
24 to detect grain defects where each set of corresponding
detectors inspects a corresponding one of transverse inspection
sections 118a-118c of surface 116. For example, the transverse
inspection section 118a is subject to grain defect inspec:tion by
comparing the output of detectors 122a and 124a. Movement of
article 112 along its longitudinal axis 114 while performing grain
~ 2 ~
fect inspection at inspection sections ll~a-118c a~complishes
inspection along corresponding longi~udinal areas 118a'-118c'
whereby the surface 116 is inspected for grain de~ects.
Light source 126 of apparatus 100 includes a laser source 126a
directed at a deflecting mirror 125b which in turn projects light
beam 127 through a telecentric lens/mirror assembly 126c. This
arrangement may be known to those skilled in the art as a
telecentric flying spot linescanning system whereby light beam 127
remains substantially parallel to a plane orthogonal to surface 116
and containing axis 114 and maintains a given angle of incidence
154. This longitudinal orientation ~or light beam 127 is important
as it maintains a longitudinal orientation for the cone 4~ (FIG.
6) according to the specular diffuse light re~lection model. With
light beam 127 tracing across the grain of article 112 as
~escribed, difi~use reflection measured by corresponding detectors
is similar due to the symmetry of such diffuse reflection and
symmetric positioning o~ corresponding detectors within axrays 122
and 124. Accordingly, each detector of a pair of corresponding
detectors measures similar light intensity for diffuse reflection
from sur~ace 116.
Specular di~use reflection, however, is not symmetric. This
as~mmetry relates the maximum spacing to the view angle, i.e.,
relates ths maximum spacing between detectors within each array 122
and 124 to the detector view angle. According to the specular
1~
2 ~
Pfuse reflection model, as the light beam 12~ moves transversely
across the article 11~ at a given angle of incidence, the cone ~54
(FIG. 6) similarly moves transversely across article 112. The
detectors within array 122 must be spaced so as to maintain a view
angle including the surface of cone ~4 such that the specular
component 42 of reflected light, when presentl may be detecte~
More particularly, ~he view angle of each detector of array 122
includes the arc 45 (FIG. 6) so long as the light beam 127 is
incident at the corresponding transverse portion of line 118. In
a practical application, view angles on the order of +/- twelve
degr~es should be adequate. Thus, by maintaining proper spacing
between detectors 122, i.e., not too far apart, specular reflection
from beam 127 as it moves through each of transverse sections 118a-
118c is detected by a corresponding one of detectors 122a-122c,
respectiv~ly.
In operation, article 12 is moved along its longitudinal axis
114 relative to apparatus 100 as light beam 127 traces across its
grain to define the line 118. Light source 126 providas to
discrimination circuit 128 the light beam 127 position dat:a, i.e.,
its transverse position along line 118. As previously noted, index
rollex control 172 provides the article 112 position data to
discrimination circuit 128. The light beam 127 position data from
light source 126 together with the article 112 position clata fro~
roller control 172 provides to discrimination circuit 128
sufficient information to associate a physical location on surface
2 ~ 7
; corresponding to a current point of incidence for light beam
127. By suitably sampling data from arrays 122 and 124 in
coordination with knowledge of the actual point of incidence of
light beam 127, discrimination circuit 128 constructs data 174 as
5output representing surface 116. For example, as light beam 127
moves through inspection section 118b, discrimination circuit 128
collects data from detectors 122b and 124b at given rate
corresponding to a given number of points along section .118b, and
calculates a corresponding set of ratio values thereof, so as to
10identify grain defects along the corresponding portion of surface
116. Surface 116 representation data 174 i5 then made available
to subsequent wood processing steps to ~;r;ze use of article 112,
The pxesent invention has shown an inherent ability to
15discriminate not only between normal grain patterns and grain
defects~ but also between wood and certain grading marks thereon,
particularly felt tip ink marks and wax crayon marks. Ink is taken
up more readily by the diffuse rela~ed wood structures, i.e., cell
cavities, resin, etc., than by the specular diffuse related
ZOcomponents, i.e., the cellulose fibers in the cell walls,. This
suppresses the diffuse detector 24 output much more than' the
specular diffuse detector 22 output such that the ratio of detector
22 to detector 24 output rises to a very high level, higher than
that associated with clearwood. This high ratio in connection with
25an unusually low detector 2~ output yields a clear "lnk signature.l'
In knotwoods the specular diffuse components are virtually
16
2~2~7
nexistent and the ratio for knotwoods remains unchanged when
marked by ink. Accordingly, lnk marks upon knotwood may not be
detectableO As for wax crayon marks, the wax provide~ a specular
surface on the wood while the wax pigment suppresses the diffuse
reflection. Again this yields a very high detector output ratio,
but with a characteristic low specular diffuse detector 22 output.
Wax marks are generally detectable whether on Xnotwood or
clearwood.
FIG. 11 is a flow chart illustrating operation of the device
o~ FIGS. 1 and 2 for each indexed inspection point .including
di~crimination of grain defects as w811 as ink and wax ma:rkings on
wood article 12. It should be apparent to those sXilled in the art
how the ~teps illustrated in FIG. 11 may be applied to the device
of FIGS. g and 10. More particularly, the steps illustrated in
FIG. 11 correspond to each sampling of data taken from
corresponding detector points of arrays 122 and 124.
In FIG. 11, pxocessîng by discrimination circuit 28 of each
sampling of data collected from detectors 22 and 24 begins in block
200 where the output o~ the specular detector 22 is stored in the
variable D1 and diffuse detector 24 output is stored in the
variable D2. In block 20Z discrimination circuit 2a computes the
ratio D1/D2 for storage in the variable RATI0, The value of RATIO
~5 will determine branching through the flow chart and actual RATI0
values for branching decisions will vary depending on the
~ ~4~ 7
~plication. In decision block 204 the value of RATIo ls compar~d
to unity. If substantially equal ~o unity, e.g. below the value
1.3, processing branches to block 206 where cliscrimination circuit
28 identifies the portion of surface 16 corresponding to inspection
point 18 as a grain defect. A negative result in block 204 passes
processing control to decision block 207. If decision bloc~ 207
determines that the value of RATI0 is approximately equal to the
value two, e.g. in the range o~ 1~3 ~o 2.~, processing branches to
blocX 208 where discrimination circuit 28 identiPies the portion
of surface 116 corresponding to inspection point 18 as a normal
grain pattern. Processing from either of blocks 206 or 208
continues to block 210 where discrimination circuit 28 stores the
values D1 and D2 and maintains statistical data representing
typical D1 and D2 values useful in discriminating between ink and
wax marks. Continuing to block ~12, discrimination circuit 28
incorporates the identification of the inspection point 18, e.g.
grain dePect or normal grain pattern, by updating a surface 16
database structure.
If decision block 207 determines khat RATI0 is not
substan~ially equal to the value two, e.g. greatar than the value
three, processing branches to block ~14 where the portion of
surface 16 corresponding to inspection point 18 is identified as
being m~rked either by ink or wax. Should discrimination between
ink and wax markings be desired, block 216 references statistical
~ata previously stored as typical D1 and D2 values. Decision block
2 OI62L1~7
9 tests for an unusually high D1 value, inferring a wax mark in
block 220, and decision block 222 tests ~or an unusually low D~
value, inferring in ~lock 224 an ink mark. If decision block 222
determines that D2 is not unusually low, the type of marking is
identified as unknown in block ~26. In any case, processing
eventually reaches block 212 where discriminatio.n circuit 2~ wo-lld
suitahly update the surfacP 16 database structure relative to the
corresponding portion of surface 16 as being marked, either by wax,
by ink, or unknown. ~fter sc~nni ng the entixe sur~ac 16, the
surface 16 data structure may be referenced to identify the
location oE grain defects as well as interpret detected ink and wax
markings thereon.
FIG. 12 illustrates a modification applicable to both of the
illustrated embodiments of the present invention for more
accurately determining the dimension of knotwoodO The modification
will be described with reference to the embodiment of FIGS. 1 and
2. There exists an extensive region below knots, toward the butt
o~ the log from which the sample has been cut, where the grain
carrie~ a diving component, i.e. the direction of the wood fiber
has a component oriented into the surface 16. The ratio of
detector 22 output to detector 24 output, being greater than unity,
for this region will correctly indicate clearwood, but the actual
~pecular reElection peak angle deviates from the specular angle
de~ined by the sur~ace 16 o~ article 12. This reduces the ratio
o~ detector 2.2 output to detector 24 output and th.is can extend the
lg
2~2~l~7
parent dimension of knots. FIG. 8 i~ strat2s this deviation.
In FIG. 8, detector 22 output for clearwood response 50 does not
actually peak at the specular angle 46, rather at an adj acent angle
49. Angl~ 49 may be on either side of angle 46 depending on the
direction of material fe~ding, i.e. whether the butt end leads~
It may be d~sirable to detect the peak value for the specular
component of reflected light so as to ~ e the calculated
detestor 22 to detector 2~ output ratio. Should it be desirable
to minimi~e the effect due to such diving grain, it is suggested
that a pair o~ detectors, 322a and 322b in FIG. 12, replace
detector 22. Detectors 322a and 322b are positioned above and
below, respectively, the specular reflection a~gle 46 and their
outputs are appropriately combined to captur~ the peak specular
value and produce what would otherwise be detector ~2 output~
In typical applications, the angle of incidence 54 is on the
order of 45 degrees and the diametex of light beam 27 is on the
order of one to two millimeters. Light polari~ation orientation
may be optimized in some applications, but has generally been found
satisfactory at a variety of polarization orientatians.
In certain hardwood species, the "ring porous" types, there
are wide bands of large cells t"vessels"~ that appear on the face
of a sur~aced board. These vessels can be on the order of 1 mm by
7.5 4 mm in size, and form relatively large pits when the wood i5
~urfaced. Typical species are the red and white oaks, where these
: ~0~2/-l~7
~ nds of vessels give the wood surface an appearance that is
usually desirable for i~s use in furniture or trim moldings.
Unfortunately, these ~ells can also strongly scatter and absorb
light beam 2~ since the cell size is large enough to encompass the
typical collimated laser beam used in this application. Further/
in certain species a structure called ~'tyloses'l form~ in the
vessels and absorbs the optical radiation even more strongly.
To address this problem with ring porous type hardwoods, the
apparatus 20 may be adapted in three ways. The adaptation will be
described with reference to the embodiment of FIGS. 1 and 2, but
is equally applicable to the embodiment of FIGS. 9 and 10~ First,
increase the angle of incidence 54 to about fifty-five degrees.
Second, increase the laser spot si~e, i.e. the area of point 18,
in the cross scan direction, i.e. along the normal grain direction
to 4 or 5 mm to pro~ride an elongate area o~ incidence. This is
achieved with a cylindrical lens in the laser path, usually pre-
scan. This will not increase the spot size in the orthogonal
direction (across the grain). Such elongation in the scan
direction would decrease the ability o~ the apparatus 20 to
resolve the edge of the defect areas. Third, orient the laser
polarization parallel to the plane of incidence 25. These
adaptations are not required for other hardwoods (the "diffuse-
porous" ~pecies), or for so~twood species, which do not have
vessels. Tha increases i~ angle of incidence 54 and area of point
1~ axe not generally desirable, bu~ represent a trade-off in order
~1
2~2l~7
~ better detect defects in the ring porous species. It should be
apparent to those skilled in ~he art how these adaptations may be
applied to the apparatus 100 of FIGS. 9 and 10.
Thus, a method and apparatus for detection oP yrain defects
by reflective scanning has been shown and described. Th~ method
and apparatus accordiny to the present in~ention is an inherently
simple and low cost technology when applied to across grain
scanning in longitudinal material feed applications. The method
and apparatus ha~ provided excellent discrimination for }cnotwood,
both live and dead, in a wide range of solid wood products such as
softwood veneers, hardwood and softwood boardst and shingles. The
method and apparatus is useful in a wide range of surface
conditions such as characterization of dirty, stained, rough cut,
knife planed, or abrasive planed surfaces.
While a preferred embodiment of the present invention has been
shown and described, it will be apparent to those skilled in the
art that modifications relative to the illustrated embodiment are
2~ possible without departure fro~ the scope of the invention. For
example, the present invention may be practiced by a variety of
techniques beyond light energy scanning such as acoustics or any
othar ~orm of energy that undergoes "specular-diffuse reflection'l.
Xt will be appreciated, therefore, that the SCOp~! O~ the
present invention is not restricted to the particular embodiment
2~2~7
'hat has been described and illustrated, but rather includes any
modifications as fall wi~.hin the appended claims and equivalents
thereof.
23
