Note: Descriptions are shown in the official language in which they were submitted.
CA 02382372 2002-05-10
1
LIGHT-BEAM MULTIPLE REFLECTIONS RESOLUTION
BACKGROUND OF THE INVENTION
held of the Invention
The present invention relates in general to method and apparatus for
processing
multiple reflections of a light beam from a surface onto a photo-detector, and
relates in
particular to a distance measuring technology applicable to various kinds of
manufacturing
machines and evaluation apparatuses.
This application is divided from Canadian Patent Application 2,2SS,097 filed
December 4, 1998.
Description of the Related Art
Distance measuring apparatus is an integrated part of various equipment for
manufacturing, fabricating, measuring and evaluating activities, and non-
contacting distance
measuring apparatus (distance sensor) is a known example in such applications.
Non-contacting distance sensors include ultrasonic and laser range sensors,
but laser
range sensors are preferred when the application requires rapid response and
high precision.
Figure 16 shows a schematic illustration of how a conventional laser range
sensor
operates.
In the diagram, l represents a light source, 2 a light beam generated from the
light
source 1, 3 a measuring object, 7 a reflected beam from the surface of the
measuring object
3, 8 an optical member, 9 a photo-detector member 9, and 10 a light receiving
surface of the
member 9.
As shown in this diagram, laser range sensor is comprised by a laser beam
output
section for emitting a laser beam 2 generated from the light source 3 towards
the measuring
object 3, and a light input section for focusing the reflected beam 7 leaving
the surface of the
measuring object 3 at a light spot on the surface of the light receiving
surface 10 of the
photo-detector 9 through an optical member 8.
Photo-detector 9 is a member to convert the luminous energy falling on the
light spot
focused on the light receiving surface 10 to electrical signals, and may
include, for example,
a one-dimensional charge coupled device (CCD) or one-dimensional position
sensitive
device (PSD).
CA 02382372 2002-05-10
2
The optical member 8 may commonly be a mirror, prism, or lens, but in Figure
16,
the laser range sensor uses only a lens for the optical member.
Also, although not shown in the diagram, a laser range sensor is generally
provided
with a control section for controlling the output and input sections and a
computation
section for determining the distance to the object according to measured data
from the input
section.
Such laser range sensors operate by directing a light beam 2 to form a light
spot on
the surface of the measuring object 3 generated from the Light source 1, and
forming another
light spot on the receiving surface 10 created by focusing a reflected beam 7
through the
optical member (receiving lens) 8.
In this case, as illustrated in Figure 17, the focal position on the receiving
surface 10
changes depending on the distance to the measuring object 3.
The distance to the object 3 can be determined by cahbrating the correlation
between
the focal positions and distances, in terms of the known distances to the
objects (3a, 3b).
Further details of range sensors are discussed in references, such as "Use and
Problems of Optical Devices" (Suede Tetsuo, Optronics, 1995).
One of the critical parameters in determining the precision of measurement by
laser
range sensor is the precision by which the positions of the light spots on the
light receiving
surface 10 of the photo-detector 9 are determined.
When the photo-detector 9 is made of a one-dimensional CCD, the light
receiving
surface 10 is comprised by a number of pixels disposed along a straight line,
and the
luminous cnergy of the light falling on each pixel can be converted to
clectrical signals of
given magnitudes.
Therefore, by processing the electrical signals and obtaining a peak position
or a
weighted average position of luminous energy, it is possible to know which
pixel position
corresponds to the focal position of the reflection light spot.
When the photo-detector 9 is made of a one-dimensional PSD, it is possible to
know
the focal position of reflection tight spot by processing the electrical
signals output from the
PSD to give the weighted average position of luminous energy as a ratio to the
total length
of the PSD, as illustrated in Figure 18.
CA 02382372 2002-05-10
3
Accordingly, in principle, laser range sensors determine the distance by
receiving a
beam reflected only once from the object 3 (simple reflection beam 7) in the
photo-detector
9.
When the surface of the object 3 is glossy, a beam first reflected from the
object
surface may be reflected again by other surfaces (causing multiple
reflections) and then
return to the sensor, such that there are cases in which the simple reflection
beam 7 becomes
mixed with multiple reflection beam and the correct position of the simple
reflection beam 7
cannot be determined with precision. In such a case, the measurement precision
is
significantly reduced.
To understand the loss of measurement precision in more detail, it is
necessary to
explain how multiple reflections affect the precision of measurements.
When multiple reflections occur, a plurality of light points are produced on
the light
receiving surface 10 so that the method based on peak luminosity cannot
provide the correct
position of the simple reflection beam 7, because the position corresponding
to the peak
luminosity does not necessarily indicate the position of the simple reflection
beam 7.
Using the method of weighted average luminosity, weighting tends to be shifted
towards the positions of multiple reflection, and again it is not possible to
determine the
correct position of the simple reflection beam 7. This effect is illustrated
in Figure 19.
Multiple reflection is classified as either 2nd-order (reflected twice), 3rd-
order (three
times) or 4th-order multiple refection, depending on the number of times the
tight is
reflected during the interval from leaving the light source 1 to entering into
the photo-
detector 9.
An example of a multiple reflection is ~7lustrated in Figure 20 using a case
involving a
2nd-order reflection beam 15.
The 2nd-order reflection beam 15 can be avoided to some extent by orienting
and
operating the photo-detector 9 properly. This will be explained below with
reference to
Figures 21 and 22.
When the detector 9 and a scanner mirror 12 are oriented as shown in Figure
21, the
2nd-order reflection beam 15 is received in the detector 9, but when they are
oriented as
shown in Figure 22 to coincide the two beams, the 2nd-order reflection beam 15
cannot be
received by the detector 9.
CA 02382372 2002-05-10
4
Therefore, by operating the measuring equipment suitably by redirecting the
incident
beam with respect to the object 3 with the use of a scanner mirror 12, it is
possible to avoid
detrimental effects of the 2nd-order reflection beam 15.
A 3rd-order reflection beam 16 such as the one illustrated in Figure 23 is
generated
when the surface of an object I8 is glossy and another object 19 (for example,
without
glossy surface) is present nearby.
As illustrated in Figure 23, a 3rd-order reflection 16 is produced when a
light beam 2
emitted from a light source 1 is specularly reflected (as from a mirror
surface) from the
surface of a glossy object 18, which is the measuring object3, and is then
reflected diffusely
from the surface of a dull object 19, and the diffused reflected light is
reflected, for the third
time, from the surface of the measuring object 3 before reaching the dctector
9.
In other words, this pattern of reflection may be said to be a result of the
surface of
the glossy object 18 acting as a mirror to generate a mirror image 20 of the
dull object 19 so
that a light spot focused by the 3rd-order reflection represents the distance
to the mirror
image 20.
It should be noted that the 3rd-order reflection beam 16 cannot be avoided by
simply
altering the awangement of components in the optical system, because the laser
range
sensor, which is a light-based system, follows the basic principles of optics.
Although multiply reflection beams of 4th-order or higher do exist in
principle,
optical power is attenuated at each reflection so that adverse effiects of
reflection beams of
higher than 4th-order can be neglected in practice.
SUMMARY OF THE INVENTION
It is an object of the present invention to overcome the problems in
conventional
laser range sensors by providing an apparatus and a method to enable to
provide accurate
distance measurement even if 3rd-order reflections exist in the environment,
which cannot be
avoided in principle by altering the arrangement of the optical components in
the system.
The object has been achieved in a method and an apparatus to perform the
method
comprising the steps of:
CA 02382372 2002-05-10
directing a light beam generated from a light source on a measuring object;
focusing reflection beams reflected from the measuring object on a receiving
surface of a
photo-detector member through an optical device to generate a luminous energy
distribution;
performing a spot separation process; if a plurality of peaks exist in the
luminous
energy distribution curve,
selecting a necessary light spot;
determining a focal position of the selected light spot on the light-receiving
surface, thereby deriving the distance to the object.
An aspect of the invention provides an application program recorded in a
computer-readable recording medium and executing the program by a computer
system.
Accordingly, a light spot formed by a multiple reflection beam can be
separated
from a light spot formed by a simple reflection beam, thereby enabling
derivation of the
distance precisely even for an object which is susceptible to producing
multiple
reflections because of its glossy surface.
The present invention also provides methods for each of the processes of spot
separation, spot selection, luminosity distribution restoration, and focal
position
detection.
Accordingly, a method is provided for separating a plurality of light spots,
formed
simultaneously on a light-receiving surface provided in a photo-detector
member of not
less than one-dimensional response capability, producing a luminous energy
distribution
exhibiting a plurality of peaks in not less than one-dimension, wherein a
maximum value
of each peak, minimum values and inflection points surrounding the maximum
value are
computed for each peak in the luminous energy distribution, and the luminous
energy
distribution is divided into individual luminosity curves at either a minimum
value or an
inflection point whichever is closer to the maximum value.
The invention also provides a method for selecting a light spot created from a
plurality of light spots, formed simultaneously on a light-receiving surface
provided in a
photo-detector member of not less than one-dimensional response capability,
producing a
plurality of maximum values in a luminous energy distribution generated by the
photo-
detector member, by selecting a light spot whose maximum value exceeds a
threshold
intensity value and whose derived distance to a measuring object is a minimum.
CA 02382372 2002-05-10
6
The invention also provides a method for restoring luminosity to individual
luminosity curves, when the curves are partly or wholly produced on a light-
receiving
surface provided in a photo-detector member of not less than one-dimensional
response
capability, by applying a selected distribution function to a luminosity curve
so as to
restore complete luminosity to the luminosity curve.
A more specific aspect is a method for determining the focal distance of the
light
spots from the individual luminosity curves obtained by the spot separation
process.
The above-described methods can be provided in an application program
recorded in a computer-readable recording medium and executing the program by
a
computer system.
Thus, all the necessary steps, in achieving the basic object of deriving a
precise
distance to the object, have been provided through the process of separating
the light
spots, selecting a necessary light spot, restoring luminosity of the selected
light spot and
detecting the focal position of the selected light spot.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 is a schematic diagram of a first embodiment of the distance
measuring
apparatus of the present invention.
Figure 2 is a diagram to explain the operational steps of the distance
measuring
apparatus in the first embodiment.
Figure 3 is a diagram showing the luminous energy distribution of light spots
formed on the light-receiving surface of the photo-detector, by a simple
reflection beam
and a 3rd-order reflection beam.
Figure 4 is a flowchart showing the processing steps in the spot separation
section
in the first embodiment.
Figure 5 is an illustration of light paths of a 3rd-order reflection beam.
Figure 6 is a flowchart showing the processing steps in the spot selection
section.
Figure 7 is a graph showing different forms of distribution function for
restoring
luminosity of a simple reflection beam in the input energy distribution
restoration
section.
CA 02382372 2002-05-10
7
Figure 8 is a flowchart showing the steps for restoring luminosity in the
luminosity
distnbution restoration section in a second embodiment.
Figure 9 is a schematic diagram of a fourth embodiment of the distance
measuring
apparatus.
Figure 10 is a schematic diagram of a fifth embodiment of the distance
measuring
apparatus.
Figure 11 is a flowchart showing the steps for restoring luminosity in the
luminosity
distnbution restoration section in a sixth embodiment.
Figure 12 is a graph showing the results of actual processing of a 3rd-order
reflection beam in the sixth embodiment.
Figure 13 is a schematic diagram of a distance measuring apparatus in an
eighth
embodiment.
Figure 14 is a schematic diagram of a distance measuring apparatus in a ninth
embodiment.
Figure 15 is a schematic diagram of a distance measuring apparatus in a tenth
embodiment.
Figure 16 is an illustration of the principle of operation of the conventional
laser
range sensor.
Figure 17 is an illustration of the distance measurement based on the
conventional
laser range sensor.
Figure 18 is a graph showing a focal position computed from the luminous
energy
distnbution when multiple reflection is not present.
Figure 19 is a graph showing focal positions computed from the luminous energy
distnbution when multiple reflection is present.
Figure 20 is an illustration of a case of 2nd-order multiple reflection.
Figure 21 is an illustration of an example of an L-shaped object as an example
of
objects susceptible to generating multiple reflections.
Figure 22 is an illustration of an arrangement of optical components when the
object
is lrshaped.
Figure 23 is an illustration to explain the principle of generation of
multiple
reflections.
CA 02382372 2002-05-10
g
Figure 24 is an illustration of superposition of light spots, creased by a
simple
refection beam and a 3rd-order reflection beam, overlapping their respective
inflection
points.
Figure 2S is an illustration of nearly complete superposition of light spots
formed by
a simple reflection beam and a multiple reflection beam.
Figure 26 is a flowchart showing the operational steps for the distance
measuring
apparatus of the present invention.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
The following embodiments are presented to illustrate the basic principle of
the
method, and are not intended to limit the invention disclosed in the claims.
It is not essential
that all the combination of the features presented in the embodiments are
necessary in
attempting to solve various problems which may appear in actual distance
measurements.
Preferred embodiment will be presented in the following with reference to the
drawings.
In all the drawings, those components having similar functions are referred to
by the
same reference numerals and their explanations are not repeated.
Embodiment 1
Embodiment 1 will be presented with reference to Figure 2 which shows a
schematic
arrangement of the distance measuring apparatus.
In the drawing, 1 represents a light source, 2 an optical beam generated from
the
light source l, 3 a measuring object, 9 a detector member, 21 an input
section, 30 a spot
separation section, 31 a spot selection section, 32 a luminosity distr'bution
restoration
section, 33 a focal position detection section, 34 a distance computation
section, and 35 a
control section.
In the present embodiment of the distance measuring apparatus, a light beam 2
emitted from the light source 1 is radiated onto a measuring object 3, and
reflection beams
from the surface of the object 3 (a simple reflection beam 7 and a multiple
reflection beam
16 shown in Figure 23) are passed through an input section 21 comprised by a
mirror, a
CA 02382372 2002-05-10
9
prism, and a lens to focus a light spot on the light receiving surface of the
detector 9. The
control section 35 controls and operates the beam output section and the input
section 21.
Based on a luminous energy distribution generated by the detector member 9,
each
light spot is separated individually in the spot separation section 30, and
the position of the
light spot focused by the simple reflection beam 7 is selected by the spot
selection section
31.
Further, the luminous energy distribution of the simple reflection beam
selected by
the spot selection section 31 is restored in the luminosity distribution
restoration section 32,
and focal position of the simple reflection beam 7 is obtained in the focal
position detection
section 33, and the distance to the object 3 is computed in the distance
computation section
34.
This configuration of the present apparatus enables to determine the distance
precisely, even in a multiple refection environment.
Also, the control section 35 controls all the devices in the processing
sections
(30--34).
Figure 1 shows a schematic arrangement of the apparatus of Embodiment 1.
As shown in this diagram, the present apparatus direct a beam 2 emitted from
the
light source 1 to an object 18 having glossy surface, and a reflection beam
reflected from the
surface of the glossy object 18 is passed through an input section 21
comprised by a mirror,
a prism and a lens and the h7ce, and is focused on the light receiving surface
of the detector
member 9 to form a light spot.
Also, the beam 2 emitted from the light source 1 undergoes a simple reflection
on the
glossy object 18 first, and then undergoes a multiple reflection on a
neighboring object 19,
having a dull surface, and a multiple reflected beam is again reflected on the
surface of the
object 18 to cause a 3rd-order reflection beam 16 which is also focused as a
light spot on the
light receiving surface of the detector member 9.
Figure 3 shows luminous energy distributions of the light spots focused on the
light
receiving surface of the detector member 9 by the simple reflection beam 7
(right peak) and
the 3rd-order reflection beam 16 (left peak). Figure 3 relates to a case in
which both beams
7,16 are received by the detector member 9.
CA 02382372 2002-05-10
It should be noted in Figure 3 that A2 relates to a luminous energy
distn'bution curve
created by the simple reflection beam 7 only, the A3 relates to the same
created by the 3rd-
order reflection beam 16 only, and the A1 represents the luminous energy
distribution curve
created by both beams 7 and 16.
Next, the operation of the distance measuring apparatus when the detector
member 9
detects a luminous energy distn'bution curve A1 shown in Figure 3 will be
explained with
reference to Figure 26.
When a luminous energy distribution curve exhibiting a plurality of peaks is
obtained
(step 201), the spot separation section 30 performs separation of light spots
using the
maximum, minimum and inflection points of the distn'bution curves (step 202).
Spot selection section 31 selects a light spot for the simple reflection from
the
separated light spots, using the maximum value of the distribution curves
(step 203).
Luminosity distn'bution restoration section 32 restores luminosity for the
light spot
focused by the simple reflection beam by entering the values of the maximum
and inflection
points of the light spot selected in step 203 (step 204).
Focal position detection section 33 determines the focal position of the
restored
curve on the light receiving surface of the detector member using the restored
energy
distn'bution curve restored in step 204 (step 205).
Next, the distance computation section 34 computes the distance from the focal
position detected in step 205 to the measuring object using the triangulation
technique (step
206), and outputs a value for the distance (step 207).
In the following, a method of calculating the distance from the luminous
energy
distn'bution graph shown in Figure 3 will be explained by presenting the
successive steps
performed in the computational process.
When the luminous energy distribution data obtained by the detector member 9
are
inputted into the spot separation section 30, the following operation of the
section 30 is
going to start.
[Spot separation section 30)
CA 02382372 2002-05-10
II
Figure 4 shows a flowchart of the processing steps performed in the spot
separation
section 30.
When the luminous energy distribution data obtained by the detector member 9
are
entered into the spot separation section 30 (step 100), first and second
differentiation steps
are carried out on the energy distribution data (step 101), and the extreme
values and the
points of inflection of the distribution data are computed (step 102).
Next, the minimum values of the tight spots on the detector member 9 are used
to
separate each spot of the various light spots (step I03), so as to separate
the luminosity
curve for each light spot (step 104),
Specifically, the energy distribution data obtained by the detector member 9
are
subjected to first differentiation to obtain the coordinate points p~~ (x",~,,
y",~1) and
pMnxz (xMn~c~, yMnxz) for maximum values.
Similarly, the coordinates for the minimum values are pM~,l (xMINI, yMm), pMnn
(xMIN2~ yMlN2)i and pM~r3 (xMna3~ yMnaa)~
Next, by performing a second differentiation, the inflection points per, per,
per, p~4
of the curves are obtained.
First derivatives relate to extreme values at zero slope to give maximum and
minimum values of the curves, and zero values in the second derivatives relate
to the
inflection points.
Next, each light spot on the distribution curve is separated into
corresponding
luminosity curves at their minimum values pM~ (x~, yM~).
The energy distribution data for each light spot obtained in the spot
separation
section 30 are input into the spot selection section 31.
(Spot selection section 31]
Spot selection section 31 examines all the maximum values in the distribution
data,
and selects one maximum value, which exceeds a threshold value and represents
the shortest
distance to the object, as the correct light spot focused by the simple
reflection beam 7.
Also, the luminous energy distribution data are divided into individual
luminosity
values due to the simple reflection beam 7 and the 3rd-order reflection beam
16 using the
left and right inflection points closest to the respective peak values as the
boundary.
CA 02382372 2002-05-10
12
Figure 5 shows an example of the light path of a 3rd-order reflection beam 16.
In
Figure 5, a simple reflection beam 7 reflects from a point Pl on a glossy
object 18, and a
part of the beam then focuses, through a scanner mirror 12 and a lens 8, at
point Ql, on the
detector member 9.
Here, because the point Pl is on a glossy object 18, a part of the beam
reflected from
point Pl arrives at a point on a nearby matte object 19.
Because the object 19 has a matte surface, the laser beam arriving on point P2
undergoes diffused reflection, and a part of this beam undergoes a third
reflection on the
object 18 at a point P3 which is slightly removed from the point P1, and
focuses at a point
Q2 on the detector member 9 indicating that it is located further away.
In effect, the 3rd-order reflection beam 16 appears as though it is reflected
from a
point P'2 located on a fictitious plane 20 in such a way that the distance Pl
to P2' is the
same as the distance Pl to P2, and naturally, the light spot on the detector
member 9
indicatcs that P'2 is further away by a distance equal to Pl to P'2.
In general, therefore, when a 3rd-order reflection is generated, the center of
gravity
of the luminous energy seen by the detector member 9 is biased in the
direction of longer
distance. In other words, the center of brightness appears falsely further.
To remove the influence of 3rd-order multiple reflection, it is necessary to
obtain the
focal position of the simple refection beam 7.
If a remedial step is taken for the 3rd-order reflection (beam 15 in Figure
20),
because the influence from multiple reflections higher than 4th-order is
negligible, the spot
focused by the simple reflection beam 7 will always be closer than the light
spot focused by
the 3rd-order reflection, so that a light spot corresponding to the shortest
distance to the
object can be the correct light spot.
Figure 6 is a flowchart of the processing steps performed in the spot
selection
section 31.
When the energy distr'bution data from each light spot obtained in the spot
separation section 30 is input into the spot selection section 31 (step 110),
a threshold value
to be used as a criterion for the maximum values (step 111) is determined, and
the maximum
values obtained in the spot separation section 30 are selected as candidate
values for the
CA 02382372 2002-05-10
13
simple reflection beam 7 (step 112), and these values are examined for the
requirement to
exceed the threshold value (step 113).
In step 113, if a maximum value is greater than the threshold value, the
position of
that maximum value is examined for another requirement to be the shortest
distance to the
object (right-side on the detector 9) (step 114), and the maximum value
indicating the
shortest distance (farthest on the right-side of the detector 9) is selected
as the correct
maximum for the simple reflection beam 7 (step 115).
In more detail, when a plurality of maximum values are input into the spot
selection
section 31, it selects one maximum value (step 113) if it is higher than the
threshold value
(selected in step 110 to remove noise) and corresponds to the shortest
distance to the object
(step 114). The selected value is designated as the maximum luminosity for the
simple
reflection beam 7.
In the~present example shown in Figure 3, coordinatcs for the light spot due
to the
simple reflection beam 7 are given by p(xMnxz, yes).
The inflection points per, p~4 of the distribution curve are assumed to be the
dividing
points (positions separating each light spot) and the distribution data are
divided at the
inflection points to delete all the distn'bution data other than those lying
inside the range
given by the inflection points p~ ~ p~4. This process reduces the uncertainty
in choosing
the correct light spot, caused by the influence of luminosity of the multiple
reflection beam in
agecting the position of peak luminosity for the correct light spot due to the
simple
reflection beam 7.
On the other hand, if the luminosity peaks are close together, the minimum
value
pMn~r~ (xMU~rz, yMn~x), rather than the inflection point per, become closer to
the peak values
p~ (x~, yes). In such a case, the energy distnbution data are divided using
the
minimum values pM~2 (xMn~r~, yMnrz) so as to attribute respective luminosity
to individual
light spots.
This will be explained in more detail with reference to Figures 3, 24 and ZS.
If the
superposition of the multiple reflection curve and the simple reflection curve
is only partial
as ~lustrated in Figure 3, the spot selection section 31 designates the region
bounded by the
CA 02382372 2002-05-10
14
inflection point p~ (on the left of p"~) and the inflection point pea (on the
right of per)
to be the luminosity value for the simple reflection beam 7. On the other
hand, if the two
curves are superimposed such as to hide the respective inflection points, as
illustrated in
Figure 24, then the spot selection section 31 selects the luminosity value for
the simple
reflection beam 7 to be the region bounded by the minimum value pM~ (on the
left of per)
and the inflection point pø4 (on the right of pMnxz). Lastly, when the two
curves are almost
totally superimposed upon each other, as illustrated in Figure 25, then the
spot selection
section 31 selects the energy field for the simple reflection beam 7 to be the
region bounded
by the inflection point p~ (on the left of p",~r~) and the inflection point
p~4 (on the right of
1)~
In other words, luminosity of each light spot is attn'buted in such a way that
those
extreme values and inflection points which are closest to the selected maximum
value define
the boundary points for separating into individual light spots.
It should be noted that the reason for deciding a threshold value for the
maximum
values in step 111 is that, unless there is a certain minimum luminosity for
each light spot,
precision measurements cannot be obtained so that, after adjusting the
apparatus sensitivity,
it is necessary to repeat the measurements regardless of the effects of
multiple reflections.
The maximum value together with the left and right inflection points of the
simple
reflection curve 7 are input into the luminosity distnbution restoration
section 32.
[I~minosity distn'bution restoration section 32]
,Although the spot selection section 31 defines a proper range of luminosity
for the
simple reflection curve, but the deletion of extraneous parts of the curve
created by the 3rd-
order reflection beam 16 also removes a part of the legitimate luminosity
belonging to the
simple reflection beam 7.
Therefore, the luminosity distribution restoration section 32 restores the
proper
luminosity for the simple reflection beam 7, which has been lost by deleting
the energy
contn'buted by the 3rd-order reflection beam 16. This computational step is
performed by
using some type of single-peak distribution function, for example a least-
square curve, to
approximate the lost portions of the curve.
CA 02382372 2002-05-10
is
In this example, the curve shape is approximated using a cosine distnbution
function
shown in Figure 7 (A4 in Figure 7). The maximum value of the luminosity curve
for beam 7
is substituted in the cosine distn'bution function to restore the proper shape
of the
distn'bution curve.
Using a similar technique, curve shape for all other light spots can be
restored using
some type of single-peak distribution function.
In this case, the total luminosity is the sum of the original luminosity
bounded by the
left and right inflection points (p~~p~4) and the energy contributions given
by p~ to the left
tail-off portion and p~4 to the right tail-off portion of the restored curve.
Luminosity value may also be taken to be the total area under a curve given by
a
single-peak distribution function chosen to represent the luminosity curve.
By following such methods, a correct luminosity curve for the simple
reflection beam
7, having negligible interference from the 3rd-order refection beam 16, can be
obtained.
Luminosity data of simple reflection beam 7 obtained by the luminosity
distn'bution
restoration section 32 are input into the focal position detection section 33.
[Focal position detection section 33J
Focal position detection section 33 computes a weighted average position for
the
simple reflection beam 7 according to the luminosity data obtained.
T'he focal position for beam 7 obtained by the focal position detection
section 33 is
input into the distance computation section 34.
[Distance computation section 34J
Distance computation section 34 determines the distance by using the focal
position
of beam 7 using the triangulation principle.
Summarizing the mcthod presented in Embodiment 1, a precise distance to an
object
susceptible to generating multiple rcflections, because of their shapes or
glossy surface, can
be derived by analytically separating the 3rd-order reflection beam from the
simple reflection
beam.
Embodiment 2
CA 02382372 2002-05-10
16
The distance measuring apparatus in Embodiment 2 applies a normal distribution
function shown in Figure 7 (AS in Figure 7), instead of the cosine
distribution function used
in Embodiment 1, to restore the Luminosity curve for the simple reflection
beam 7 in the
luminosity distribution restoration section 32. The average value in the
function is correlated
with the maximum value in the luminosity curve for the simple reflection beam
7, and the
standard deviations in the function are correlated with the inflection points
of the luminosity
curve for the simple reflection beam 7.
Figure S shows the processing steps in the luminosity distribution restoration
section
32.
The maximum value of the simple reflection beam 7 and the inflection points
obtained in the spot selection section 31 are input into the luminosity
distribution restoration
section 32 (step 120), and the maximum value and the inflection points are
then substituted
in the normal distn'bution function (step 121).
The luminosity data obtained in step 121 are stored (step 122), and luminosity
curve
for the simple reflection beam 7 is ol~ained (step 123).
By using the normal distn'bution function which more closely approximates the
energy distn'bution of the laser light than the cosine distribution function,
more accurate
luminosity distribution restoration is possible.
In this embodiment, because the weighted average position of the luminosity of
beam
7 is equal to the maximum value of beam 7, the maxim value can be used in
obtaining the
focal position of beam 7 in the focal position detection section 33.
Embodiment 3
The distance measuring apparatus in Embodiment 3 applies a Laplace
distn'bution
function shown in Figure 7 (A6 in Figure '~, instead of the cosine
distribution function used
in Embodiment 1, to restore the luminosity curve for the simple reflection
beam 7 in the
luminosity distribution restoration section 32. The average value in the
function is
correlated with the maximum value in the luminosity curve for the simple
reflection beam 7,
and the standard deviations in the function are correlated with the inflection
points of the
luminosity curve for the simple reflection beam 7.
It should be noted that in each of the embodiments (1~3) presented so far, it
is
acceptable to use any distribution function in the restoration process
performed in the
CA 02382372 2002-05-10
17
luminosity distribution restoration section 32, so long as the maximum value
of the
luminosity curve for beam 7 correlates with the average value in the function,
and the
inflection points for beam 7 correlate with the standard deviations in the
function.
By choosing a distn'6ution function to closely approximate the energy
distribution
pattern of a laser beam used in the Laser range sensor, even more accurate
restoration of the
individual luminosity can be performed.
Embodiment 4
Figure 9 shows a schematic arrangement of the distance measuring apparatus
used in
Embodiment 4.
The apparatus in Embodiment 4 differs from that in Embodiment 1 by the absence
of
the luminosity distribution restoration section 32.
That is, in the present embodiment, without performing the restoration process
for
the simple reflection beam ?, the focal position is assumed to be the maximum
value or the
center value of the luminosity curve for beam 7 obtained in spot selection
section 31.
According to this method, because there is no need to restore the luminosity
curve
and to compute weighting for the average position of the curve, focal position
of the simple
reflection beam 7 can be ol~ained more quickly.
Embodiment 5
Figure 10 shows a schematic arrangement of the distance measuring apparatus
used
in Embodiment 5.
The present apparatus differs from that in Embodiment 1 in that the spot
separation
section 30 and the spot selection section 31 are replaced with an extreme
values/inflection
points computation section 36.
When the energy distn'bution data obtained by the detector member 9 relates
only to
simple reflection beam 7 and does not include the 3rd-order reflection beam
16, the spot
separation section 30 and the spot selection section 31 are not necessary.
The distance measuring apparatus in this embodiment receives only simple
reflection
beam and does not receive 3rd-order reflection beam in the detector member 9.
Therefore,
the extreme values/inflection points computation section 36 simply produces
maximum
CA 02382372 2002-05-10
1g
values of the coordinates p,r,,"t, (xM~,X,, y",~,) from the first derivatives,
and obtains
inflection points per, p,r4 in the energy distribution curve from the second
derivatives.
Using the values of the maximum and inflection points, the luminosity
distn'bution
restoration section 32 restores the luminosity curve by following a process
similar to the
process of restoring the luminosity curve when the 3rd-order reflection beam
I6 is included.
In this case, the restored distribution curve closely approximates the
original
distribution curve so that there is no effect of restoration on measurement
error.
Accordingly, the apparatus of this embodiment is applicable when the reflected
beams contain only the simple reflection beam 7 and has no contnbution from
the 3rd-order
reflection bcam 16 so that restoration of the energy distn'bution curve and
focal position of
the light spot can be carried out eflxciently and accurately.
Embodiment 6
The arrangement of the distance measuring apparatus in this embodiment is the
same
as that in Embodiment 1.
The method of calculating the distance is different, and the computational
steps will
be explained using the energy distribution curve produced by the detector
membcr 9 shown
in Figure 3.
In this embodiment also, differentiation is performed by the spot separation
section
30 on the energy distn'bution curve produced by the detector member 9. The
steps are the
same as those explained in Figure 4 and detailed explanations will ix omitted.
As in Embodiment 1, the first derivatives provide for the extreme values and
their
coordinates.
In this embodiment, the maximum values are pM,~x, (x~l, yes,) and pM,~z (xM~,
y"s,~z), and the minimum values are pMU,,i (xMn",, yMm), pMirrz (xMn~rz.
YMn~ra)~ ~d pM~s (xMn~3,
yMlN3)~
Also, from the second derivatives, inflection points p,F,, pl~z, p,~, p,Fa are
obtained.
Next, as in Embodiment 1, the spot separation scction 30 separates each light
spot at
the minimum values pM~NZ (xMINZ, yMU~z).
Next, as in Embodiment I, the spot selection section 31 compares the peak
values
within each light spot with the threshold value, and selects the maximum value
which
CA 02382372 2002-05-10
19
exceeds the threshold value and showing the shortest distance to the object
(right-side of
detector 9) to be the maximum value for the simple refection beam ?.
The steps taken in the spot selection section 31 are the same as the one shown
in
Figure 4, and detailed explanations will be omitted.
In this embodiment, p(x,,,~, y,~) is designated to be the maximum value for
the simple reflection beam 7.
Also, the inflection points per, pø4 are used as the separation points
(dividing position
for each light spot), and the curve is separated at the separation points so
that all the data
outside of the range p~ ~ p~4 are deleted to reduce the effects of multiple
reflection.
When the peak position are close together, the minimum value pM~ (x~, yes)
rather than the inflection point p~ becomes closer to p~ (x, y,"~), therefore,
in such
a case, the minimum value pM~N2 (xM~rz, yMn~ra) is used as the separation
point to divide the
luminous energy distn'bution curve into individual light spots.
This will be explained in more detail with reference to Figures 3, 24 and 25.
If the
superposition of the multiple reflection curve and the simple reflection curve
is only partial
as illustrated in Figure 3, the spot selection section 31 designates the
region bounded by the
inflection point p~ (on the left of per) and the inflection point p~4 (on the
right of p",~)
to be the luminosity for the simple reflection beam 7. On the other hand, if
the two curves
are superimposed such as to hide the respective inflection points, as
illustrated in Figure 24,
then the spot selection section 31 selects the luminosity for the simple
reflection beam 7 to
be the region bounded by the minimum value p~ (on the left of p"~) and the
inflection
point pø4 (on the right of pMnxz). Finally, when the two curves are almost
totally
superimposed upon each other, as illustrated in Figure 25, then the spot
selection section 31
selects the luminosity for the simple reflection beam ? to be the region
bounded by the
inflection point p~ (on the left of p,,,~) and the inflection point p~4 (on
the right of per).
In other words, energy distribution of each light spot is attn'buted in such a
way that
those extreme values and inflection points which are closest to the maximum
value define
the boundary points for separating the light spots.
CA 02382372 2002-05-10
Although the spot selection section 31 selects the energy distribution for the
simple
reflection beam 7, deletion of the energy contribution made by the 3rd-order
reflection beam
16 also deletes a portion of the energy contn'buted by beam 7.
Therefore, the luminosity distribution restoration section 32 restores the
lost portion
of the curve using the normal distribution function or other distribution
function to
approximate the original curve closely.
In this case, the maximum value p~ (x~, y,,,~) for the simple reflection beam
7
corresponds to the maximum value (average) for the normal distribution
function (AS in
Figure 7) and the inflection points p~ (x~, y,~), p~4 (x~a, yea) correspond to
the standard
deviations in the normal distribution function.
From these considerations, by substituting the maximum value of the luminosity
curve and the inflection points for beam 7 into the normal distribution
function, the original
luminosity curve for beam 7 can be restored.
Examination of the luminosity curve for beam 7 shows that the curve is
somewhat
distorted showing asymmetry in the left/right halves (refer to vL and aR in
Figure 3).
Therefore, the left and right standard deviations should be different, and the
raw data cannot
be directly substituted in the normal distribution function.
Therefore, the present method of luminosity distribution restoration relies on
computing the left and right halves of the curve separately using different
values for the
standard deviation in the normal distribution function, and pasting the
restored half curves
together at their common maximum value.
Figure 11 is a flowchart showing the processing steps in the luminosity
distn'bution
restoration section 32.
When the maximum value and the inflection points for the simple reflection
beam 7
obtained in the spot selection section 31 are input into the restoration
section 32 (step 130),
the inflection points are separated into a left-half inflection point and a
right-half inflection
point (steps 131, 132), and the maximum value and the left-side inflection
point for beam 7
(step 133) are obtained.
Similarly, the maximum value and the right-side inflection point for beam 7
are
obtained (step 134). The maximum value and the left-side inflection point are
substituted in
CA 02382372 2002-05-10
21
the normal distnbution function (step 135), and the left-half curve of the
maximum value is
stored (step 136).
Similarly, the maximum value and the right-half inflection point are
substituted in the
distn'bution function (step 135), and the right-half curve of the maximum
value is stored
(step 137).
The left-half and right-half curves obtained in steps 136, 137 are joined
(step 138) to
generate a luminosity curve for the simple reflection beam 7 (step 139).
Figure 12 shows an actual example of the results of processing the energy
distn'bution data, including the 3rd-order multiple reflection, according to
the present
method. In Figure I2, the dotted line relates to the original unprocessed
data, and the solid
line relates to the processed data.
Accordingly, the present method provides a precision result for the simple
reflection
beam 7 that is not adversely affected by the 3rd-order reflection beam 16.
Next, -a weighted average position is computed in the focal position detection
section
33 based on the energy distnbution curve for beam 7 to find the focal position
for beam 7.
Finally, the distance measuring section 34 calculates the distance to the
measuring
object by triangulation using the focal positron obtained for beam 7.
It is acceptable to eliminate the luminosity distnbution restoration section
32, as in
Embodiment 4, i.e., without restoring the energy distribution curve for beam
7, and to
assume that the maximum value or the center value of the energy distn'bution
curve for beam
7 obtained in the spot selection section 31 represents the focal position.
In this case, there is no need to compute the weighted average position so
that the
focal position can be obtained more quickly.
Further, if the Luminous energy distribution on the detector member 9 is
caused only
by beam 7 and does not contain contnbution from beam 16, then, as in
Embodiment 5, the
spot separation section 30 and the spot selection section 31 can be
eliminated, and the
extreme values/inflection points computation section 36 may be provided
instead.
In this case, the extreme values/inflection points computation section 36
simply
obtains the maximum point p""~, (x~,, yhnc,) by first differentiation a~
obtains only the
inflection points per, p~4 by second differentiation.
CA 02382372 2002-05-10
22
The luminosity distn'bution restoration section 32 restores the luminosity
curve using
the values of the maximum and inflection points, in the same way as the method
to include
the 3rd-order multiple reflection 16.
In this case, the restored luminosity curve closely approximates the original
luminosity curve so that there is no effect of restoration on measurement
error.
Embodiment 7
The distance measuring apparatus of Embodiment 7 restores the luminosity of
simple
reflection beam 7 by applying laplace distribution function shown in Figure 7
(A6 in Figure
'~, instead of the cosine distribution function used in Embodiment 6, in the
luminosity
distribution restoration section 32, where the average value of the function
curve is
correlated with the maximum value in the luminosity curve for the simple
reflection beam 7,
and the standard deviations of the function curve is correlated with the
inflection points of
the luminosity curve for the simple reflection beam 7.
It should be noted that in Embodiments 6 or 7, any distribution function is
acceptable
so long as the restoration process performed in the luminosity disstribution
restoration
section 32 can correlate the maximum value of the luminosity curve for beam 7
with the
average value of the function, and the inflection points with the standard
deviations of the
function.
By choosing a distn'bution function to closely approximate the energy
distn'bution
pattern of a laser beam used in the range sensor, even more accurate
restoration of the
luminosity can be performed.
Embodiment 8
Figure 13 shows an arrangement of the distance measuring apparatus.
The apparatus in this embodiment shows an application of the present invention
to a
synchronized scanning range finder.
A light beam (laser beam) 2 emitted from a light source 1 is reflected by a
scanner
mirror 12, and after reflecting from a point Pl on a glossy object 18, a
portion of the beam is
focused, through a scanner mirror 12 and an optical device {lens) 8 on a point
Ql on the
detector member 9.
CA 02382372 2002-05-10
23
According to this apparatus, a 2nd-order reflection beam (15 in Figure 20)
generated
from the light source 2 and reflected from the glossy object 18 does not
reflect on the axial
line of the scanner mirror 12. Therefore, the beam 15 does not focus on the
detector
member 9 so that the 2nd-order reflection is structurally diverted. However,
the 3rd-order
reflection beam 16 will reflect on the axial line of the scanner mirror 12 to
focus on the
detector member 9.
Because the 3rd-order reflection beam 16 focuses on the detector member 9, the
luminous energy distribution will be different than the luminous energy
distribution caused
only by simple reflection beam 7, and correct distance measurcment cannot be
obtained.
Therefore, by applying the techniques descr'bed above for each embodiment the
luminous energy distribution data obtained by the detector member 9, the
effect of beam 16
can be removed so that the luminosity due only to the simple reflection beam 7
can be
obtained.
Similar to Embodiment 4, the luminosity distr~ution restoration section 32 may
be
deleted so that the restoring process in this section is not performed, but
the luminosity
results obtained in the spot selection section 31 is used so that the maximum
value or center
value of the curve can be assumed to be the focal position.
Further, if the resulting luminosity is caused only by beam 7 and does not
contain
contribution from beam 16, then, as in Embodiment 5, the spot separation
section 30 and the
spot selection section 31 can be eliminated, and the extreme values/inflection
points
computation section 36 may be provided instead.
Embodiment 9
Figure 14 shows an arrangemcnt of the distance measuring apparatus.
The apparatus in this embodiment shows an application of the present invention
to a
range sensor of two-dimensional response capability for measuring the
distance.
The basic approach is to consider a case of higher than two-dimensions in the
method of Embodiment 1 which uses a one-dimensional detector 9.
The following method describes the operation of a two-dimensional range
sensor.
CA 02382372 2002-05-10
24
As shown in Figure 14, two scanner mirrors 12, 13 are provided so that
scanning
may be perfonned in two-directions x- and y-directions, and the input-side is
provided with
a two-dimensional detector member 11 replacing the one-dimensional detector
member 9.
This arrangement produces a range sensor which can measure distances in a two-
dimensional space.
In this embodiment, spot separation section 30 performs differentiation steps
on the
luminous energy distn'bution surfaces obtained by the detector member 11,
thereby obtaining
coordinates for the maximum values and inflection points of the luminous
energy distribution
surface, as in Embodiment 1.
Next, as in Embodiment 1, the spot separation section 30 separates each light
spot at
the minimum values pM~ (xMnn, yMnn).
Next, as in Embodiment 1, the spot selection section 31 selects those maximum
values within each light spot which are higher than a threshold value and the
shortest
distance to the object (right-side of the detector member 9) are designated to
be the
maximum value for the simple reflection beam 7.
Next, the luminosity distribution restoration section 32 restores the
luminosity
surfaces using the maximum values in an n-order cosine distribution function
to obtain n-
order energy distribution surfaces for separate light spots.
For example, if a surface having the maximum value is chosen to represent beam
7,
n-order energy distn'bution for beam 7 can be restored using the n-order
distribution
function.
Similarly, it is possible to use Iaplace distn'bution function or normal
distn'bution
function to restore the surface for beam 7.
Also, a faster accurate method for determining the focal position is to use
p",~
(x~~ Ysuxa).
Usually, the maximum value of luminosity for beam 7 can only be represented in
whole pixel units of the detector memlxr (9 or 11), however, by using the
extreme high
values, the maximum value can be defined in terms of fractional-pixel units in
a CCD so that
more accurate distance measurements can be performed.
Also, in this embodiment also, as in Embodiment 4, the luminosity distribution
restoration section 32 may be deleted so that the restoring process is not
perfonned in this
CA 02382372 2002-05-10
section, but the energy distribution results obtained in the spot selection
section 31 is used
so that the maximum value or center value of the surface can be assumed to be
the focal
position.
Further, if the resulting energy distribution is caused only by beam 7 and
does not
contain contn'bution from beam 16, then, as in Embodiment 5, the spot
separation section 30
and the spot selection section 31 can be eliminated, and the extreme
values/inflection points
computation section 36 may be provided instead.
Embodiment 10
Figure 15 shows an arrangement of the distance measuring apparatus. T~vo
scanner
mirrors 12, 13 in Embodiment 9 were used to scan in two directions, x- and y-
directions,
respectively, but in Embodiment 10, the beam output-portion of the mirror 12
is
synchronized with the beam input-portion of the mirror 12 by extending the
mirror 12 in the
x-direction.
By synchronizing the beam input- and output-portions of the mirror 12, the 2nd-
order reflection can be prevented from focusing on the detector member 9, as
in
Embodiment 6.
Also, by synchronizing the beam output-portion of the mirror 12 with the beam
input-portion of the mirmr 12, a one-dimensional detector member 9 can be used
instead of
a two-dimensional detector member 11 to scan a two-dimensional object.
It should be noted that, although the mirror 12 was made longer to acx as one
unit,
but it is also possible to separate the output-side and input-side but to
arrange so that they
are driven synchronously.
Furthermore, in this case, the 3rd-order reflection beam 16 can be processed
in the
same way as a one-dimensional range sensor to produce the same accuracy in
distance
measurements.
In this embodiment also, as in Embodiment 4, the luminosity distn'bution
restoration
section 32 may be deletcd so that the restoring process is not performed in
this section, but
the energy distn'bution results obtained in the spot selection section 31 is
used so that the
maximum value or center value of the surface can be assumed to be the focal
position.
CA 02382372 2002-05-10
26
Further, if the resulting luminous energy distribution is caused only by the
simple
reflection beam 7 and does not contain a contnbution from 3rd-order reflection
beam 16,
then, as in Embodiment S, the spot separation section 30 and the spot
selection section 31
can be eliminated, and the extremc values/inflection points computation
section 36 may be
provided instead.
It should be noted that in all the embodiments presented (excepting Embodiment
5)
even when the 3rd-order reflection beam 16 is not contained in the energy
distnbution data,
it is clear that the method of the present invention can provide accurate
results in distance
measurement.
It should also be noted that the processing sections (3036) or any parts
thereof
shown in Figures 1, 2, 9, 10, 13, 14 and 15 can be operated automatically by
means of a
computer system for reading and executing programs recorded on some computer-
readable
recording medium to execute various processing steps, such as Light spot
separation, light
spot selection, luminosity distribution restoration, focal position detection
and distance
computation. A computer system is meant to include operating systems and
peripherals, and
computer-readable recording medium is meant to include transportable medium
such as
floppy disk, opto-magnetic disk, ROM, CD-ROM and others as well as hard disk
housed
inside the computer. Computer-readable recording medium is also meant to
include such
short-tenn memory devices to operate dynamically in some network circuit
(Internet,
telephone circuits and the like) for sending application programs, as well as
temporary
memory devices such as volat~e memories contained in servers and client
computers in
communication systems. The programs may be applicable only to a portion of the
functions
presented above, or they may operate in conjunction with existing programs pre-
installed in
the computer system.
Also, it should lx noted that, when the distance measuring is to be performed
in real-
time, the computer system is to execute the processing steps necessary for
distance
detennination by having the necessary components, such as beam output section
and beam
input section, operatively included within the system.
CA 02382372 2002-05-10
27
Also, if real-time processing is not necessary, data obtained by separate
means in the
beam output section and beam input section may be stored as digital data and
processed
later in the computer system according to the digital data supplied thereto.
Although the basic concept of the present invention was embodied in various
examples presented, it is obvious that the present invention is not limited by
specifics
embodied in the examples, and many modifications are possible without
departing from the
scope of the analytical principle outlined.
Some of the salient features of the present invention will be summarized in
the
following.
(1) The present invention provides an analytical method for separating a
multiple reflection
beam from a simple reflection beam in light-based distance measuring
techniques, thereby
achieving high precision in distance measurements for those objects
susceptible to producing
multiple reflections because of their shape or surface reflectivity.
(2) The present invention achieves measurement precision, regardless of the
presence of
multiple reflections, by analytically restoring luminosity in energy
distn'bution data created
by a simple reflection beam so as to determine a focal position of a light
spot precisely.
(3) The present invention thus provides a high precision distance measuring
apparatus of
wide applicability in new fields of application which have not been practical
in the
conventional apparatus, because of the loss of precision caused by multiple
reflections.
(4) The present invention is egective even when only simple reflection is
involved because
the precise position of a light spot can be determined more quickly compared
with those
conventional techniques using the maximum value or weighting of light spot.
(S) The present invention extends the precision of distance measurement using
simple
reflection by applying various types of distn'bution functions to better
restore luminosity data
to compensate for distortions in energy distn'bution data to suit different
surface conditions
of the object.
