Note: Descriptions are shown in the official language in which they were submitted.
CA 02802491 2013-01-21
290405-5 CA
METHOD AND APPARATUS FOR
COMPENSATING LATERAL DISPLACEMENTS AND LOW SPEED VARIATIONS
IN THE MEASURE OF A LONGITUDINAL PROFILE OF A SURFACE
TECHNICAL FIELD
This invention relates generally to the inspection of transportation
infrastructures and
in particular to the compensation for lateral movement and low speed
variations of the
measurement instrument.
BACKGROUND OF THE ART
In order to measure surface features and the longitudinal profile of a road to
be
inspected, a number of pavement condition indicators and characteristics are
measured. The
International Roughness Index (IRI) characterizes the pavement condition. FIG.
1 (Prior Art)
shows the longitudinal profile measurement in the ideal case. An aerial view
of a section of a
road to be inspected 100 is shown. The inspection vehicle 102 travels between
the lane
markings 104 along a trajectory 106 following the longitudinal axis 108 of the
road. The
number of measured profiles in the transversal axis 110 captured by the
inspection vehicle
102 with its inspection system 112 can extend from 1 to N where N is the
number of 3D
measurement points available in the transversal axis 110. In most systems, the
number of 3D
points is limited to one or two. In the special case where N is high and the
3D measurement
points are distributed so as to cover the width of the road, an image
representing the road
elevation can be formed. The inspection vehicle 102 travels between the lane
markings 104
and captures the longitudinal profile measurement 116 at a predefined distance
d to one of
the lane markings 104. In the example shown, the predefined distance d is from
the right
hand side marking 118.
In conventional systems, an acquisition instrument called a profilometer is
used for
the measurement of the longitudinal profile of roads. This acquisition
instrument includes
two single point range sensors and Inertial Measurement Units (IMUs) mounted
in the wheel
path of the inspection vehicle. The single point range sensors are used to
measure the
-1--
CA 02802491 2013-01-21
290405-5 CA
distance between the IMUs and the road and the IMUs are used to estimate the
total change
in elevation of the road and the inspection vehicle while in motion. By
subtracting the two
measurements, it is possible to measure only the variations in elevation of
the road surface,
that is, the longitudinal profile. Most of the time, the integration of the
signal from a
vertically oriented accelerometer (the simplest form of IMU) can be used to
track the total
elevation changes of both the road surface and the inspection vehicle. FIG. 2
(Prior Art)
illustrates an acquisition instrument used to measure the longitudinal profile
of the road 100.
The acquisition instrument 200 is equipped with two single point 3D sensors
202 and two
IMUs 204 or two accelerometers mounted on the inspection vehicle 102. The 3D
sensors 202
are usually positioned in the wheel path.
In some systems, the single point range sensors are replaced with multipoint
laser line
profilers that cover a road width of a few inches. These types of laser line
profilers are used
to compensate for different road surface textures such as longitudinally
tinned (striated)
concrete surfaces. FIG. 3 (Prior Art) shows an alternative configuration 300
for the
longitudinal profile measurement using a line of single point 3D sensors 302.
IMUs 304 can
still be present. This configuration adds robustness against texture
variations on the road
surface.
Both the profilers using single point range sensors and the limited width
laser line
sensors are very sensitive to the lateral shift of the inspection vehicle.
To help the driver follow the same trajectory in each run, a guide line is
often painted
on the road surface. Even guided as such, it is very difficult for the
driver/operator to
perfectly align the profiler with the reference line and to do this with
little positional
variations for multiple passes while driving at highway speeds. Lateral
movement will occur.
Since the measurement trajectory is different for each survey even when
captured on the
same road section, the longitudinal profile and the indicators calculated will
also be different.
FIG. 4 (Prior Art) shows the longitudinal profile measurement 400 with a
single point system
when lateral movement is present. Since the inspection vehicle followed a wavy
trajectory
-2-
CA 02802491 2013-01-21
290405-5 CA
402 on the road 100, the resulting longitudinal profile 404 was also measured
along a wavy
trajectory instead of being measured on the ideal straight line trajectory
116.
This non-ideal measurement trajectory 402 limits the performance and
repeatability
of the system for road monitoring applications.
SUMMARY
In methods and apparatus for longitudinal profile measurement, 3D sensors
covering
a large portion or the total width of the surface and feature tracking are
used to compensate
for lateral shifts and low speed variations of the inspection vehicle.
Methods and Systems for measuring a distance to a surface while compensating
for
variations in a transverse position and/or low speed displacement of the
instrument are
provided.
One method includes retrieving a predetermined transversal distance from a
longitudinal feature at which to extract a relevant distance; retrieving a
distance set;
retrieving a position of the longitudinal feature relative to the distance
set; extracting a range
point at the predetermined transversal distance from the longitudinal feature;
adding the
extracted point to a longitudinal distance set.
In another method, if two sensors are provided with an overlap in the
transversal
direction, extracting a range point at a predetermined transversal position;
adding the
extracted range point to a longitudinal distance set; retrieving a pitch angle
of the instrument;
calculating a local slope of the surface using an overlapping transversal
point, the pitch angle
and the separation length; calculating a height variation using the local
slope and a
longitudinal separation.
According to one aspect of the present invention, there is provided a system
for
measuring a distance to a surface along a longitudinal direction of the
surface using an
acquisition instrument while compensating for variations in a transverse
position of the
acquisition instrument, the surface having a longitudinally-aligned feature.
The system
-3-
CA 02802491 2013-01-21
290405-5 CA
comprises an acquisition instrument including a multipoint range sensor
acquiring the
distance between the acquisition instrument and the surface, the multipoint
range sensor
acquiring the distance in a field of view of the acquisition instrument at a
multitude of
transversal points, thereby acquiring a distance set, the field of view having
a transversal
dimension and a longitudinal dimension along the longitudinal direction of the
surface, the
transversal dimension being longer than the longitudinal dimension; a
translation mechanism
for displacing the acquisition instrument to allow the acquisition instrument
to acquire the
distance set at a plurality of positions along the longitudinal direction; a
processor for:
retrieving a position of the longitudinally-aligned feature of the surface
relative to the field of
view of the acquisition instrument; retrieving a predetermined transversal
distance from the
longitudinally-aligned feature at which to extract a relevant distance from
the distance set;
extracting a range point in the distance set at the predetermined transversal
distance from the
position of the longitudinally-aligned feature; adding the extracted range
point to generate a
longitudinal distance set at a constant transversal distance from the
longitudinal feature.
In one embodiment, the processor further being for retrieving an image,
wherein the
retrieving the position of the longitudinally-aligned feature comprises
detecting a location of
the longitudinal feature in the image.
In one embodiment, the retrieving the position of the longitudinally-aligned
feature
comprises detecting a location of the longitudinal feature in the distance set
generated by the
multipoint range sensor using the distance between the instrument and the
surface at the
longitudinally-aligned feature.
In one embodiment, the acquisition instrument further comprises an elevation
sensor
for measuring a total elevation of both the surface and the acquisition
instrument at least at
the predetermined transversal distance from the position of the longitudinally-
aligned feature;
the processor further being for subtracting the distance between the
acquisition instrument
and the surface acquired by the acquisition instrument from the total
elevation measured by
the elevation sensor to determine a surface elevation of the surface and for
adding the surface
-4-
CA 02802491 2013-01-21
290405-5 CA
elevation to generate a surface elevation set at a constant transversal
distance from the
longitudinal feature.
In one embodiment, the acquisition instrument includes a pitch finder, the
pitch finder
being adapted to measure a pitch angle of the acquisition instrument in the
longitudinal
direction versus gravity, wherein the multipoint range sensor includes two
multipoint range
sensors, the two multipoint range sensors being a first sensor with a first
field of view and a
second sensor with a second field of view, the first field of view partly
overlapping the
second field of view in the transversal direction at an overlap, the first
field of view being
separated by a separation length from the second field of view at the overlap
in the
longitudinal direction; wherein the processor is further adapted to determine
a surface
elevation of the surface using the pitch angle and an overlapping transversal
point in the
overlap in the first field of view and in the second field of view and the
separation length and
for adding the surface elevation to generate a surface elevation set at a
constant transversal
distance from the longitudinal feature.
According to another broad aspect of the present invention, there is provided
a
method for measuring a distance to a surface along a longitudinal direction of
the surface
using an acquisition instrument while compensating for variations in a
transverse position of
the acquisition instrument, the surface having a longitudinally-aligned
feature. The method
comprises retrieving a predetermined transversal distance from the
longitudinally-aligned
feature at which to extract a relevant distance; for each position of a
plurality of positions
along the longitudinal direction, retrieving a distance set including a
multitude of transversal
points, the transversal points each being a distance between the acquisition
instrument and
the surface along a transversal direction; retrieving a position of the
longitudinally-aligned
feature of the surface relative to the distance set; extracting a range point
in the distance set at
the predetermined transversal distance from the position of the longitudinally-
aligned feature;
adding the extracted range point to generate a longitudinal distance set at a
constant
transversal distance from the longitudinal feature.
-5--
= CA 02802491 2013-01-21
290405-5 CA
In one embodiment, retrieving the position of the longitudinally-aligned
feature
comprises detecting a location of the longitudinal feature in an image.
In one embodiment, the detection a location of the longitudinal feature in an
image
includes using an intensity of the longitudinally-aligned feature in one of a
grey-scale image,
a color image and a range image.
In one embodiment, retrieving the position of the longitudinally-aligned
feature
comprises detecting a location of the longitudinal feature in the distance set
using the
distance between the instrument and the surface at the longitudinally-aligned
feature.
In one embodiment, the method further comprising retrieving a total elevation
of both
the surface and the acquisition instrument at least at the predetermined
transversal distance
from the position of the longitudinally-aligned feature; subtracting the
distance between the
acquisition instrument and the surface from the total elevation to determine a
surface
elevation of the surface; adding the surface elevation to generate a surface
elevation set at a
constant transversal distance from the longitudinal feature.
In one embodiment, the method further comprises combining the surface
elevation set
and the longitudinal distance set to create a longitudinal 3D profile.
In one embodiment, retrieving the distance set includes retrieving a first
distance set
and a second distance set, at least a portion of the transversal points of the
first distance set
being aligned transversally with at least a portion of the transversal points
of the second
distance set thereby creating a transversal overlap of the first and second
distance sets, the
first distance set and the second distance set being acquired at separate
positions along the
longitudinal direction, the separate positions being separated by a separation
length;
retrieving a pitch angle of the acquisition instrument in the longitudinal
direction versus
gravity; calculating a local slope of the surface using an overlapping
transversal point in the
transversal overlap in the first distance set and in the second distance set,
the pitch angle and
the separation length; calculating a height variation using the local slope
and a longitudinal
distance between consecutive ones of the plurality of positions along the
longitudinal
-6-
CA 02802491 2013-01-21
290405-5 CA
direction; adding the height variation to generate a surface height set at a
constant transversal
distance from the longitudinal feature.
In one embodiment, the method further comprises combining the surface height
set
and the longitudinal distance set to create a longitudinal 3D profile.
According to another broad aspect of the present invention there is provided a
system
for measuring a distance to a surface along a longitudinal direction of the
surface using an
acquisition instrument while compensating for low speed variations of the
acquisition
instrument, the system comprising: an acquisition instrument including: a
pitch finder, the
pitch finder being adapted to measure a pitch angle of the acquisition
instrument in the
longitudinal direction versus gravity; two multipoint range sensors, the two
multipoint range
sensors including a first sensor and a second sensor, the first multipoint
range sensor
acquiring the distance between the acquisition instrument and the surface in a
first field of
view at a first multitude of transversal points, thereby acquiring a first
distance set, the
second multipoint range sensor acquiring the distance between the acquisition
instrument and
the surface in a second field of view at a second multitude of transversal
points, thereby
acquiring a second distance set, the first and second field of view having a
transversal
dimension and a longitudinal dimension along the longitudinal direction of the
surface, the
transversal dimension being longer than the longitudinal dimensionõ the first
field of view
partly overlapping the second field of view in the transversal direction at an
overlap, the first
field of view being separated by a separation length from the second field of
view at the
overlap in the longitudinal direction; a translation mechanism for displacing
the acquisition
instrument to allow the acquisition instrument to acquire the first and second
distance set at a
plurality of positions along the longitudinal direction; a processor for:
extracting a range
point in the first distance set at a predetermined transversal position;
adding the extracted
range point to generate a longitudinal distance set; determining a surface
elevation of the
surface using the pitch angle and an overlapping transversal point in the
overlap in the first
field of view and in the second field of view and the separation length;
adding the surface
elevation to the longitudinal distance set to generate a surface elevation
set.
-7-
CA 02802491 2013-01-21
290405-5 CA
According to still another broad aspect of the present invention, there is
provided a
method for measuring a distance to a surface along a longitudinal direction of
the surface
using an acquisition instrument while compensating for low speed variations of
the
acquisition instrument, the method comprising: for each position of a
plurality of positions
along the longitudinal direction, retrieving a first distance set including a
first multitude of
transversal points and a second distance set including a second multitude of
transversal
points, the transversal points each being a distance between the acquisition
instrument and
the surface along a transversal direction, at least a portion of the
transversal points of the first
distance set being aligned transversally with at least a portion of the
transversal points of the
second distance set thereby creating a transversal overlap of the first and
second distance
sets, the first distance set and the second distance set being acquired at
separate positions
along the longitudinal direction, the separate positions being separated by a
separation
length; extracting a range point in the first distance set at a predetermined
transversal
position; adding the extracted range point to generate a longitudinal distance
set; retrieving a
pitch angle of the acquisition instrument in the longitudinal direction versus
gravity;
calculating a local slope of the surface using an overlapping transversal
point in the
transversal overlap in the first distance set and in the second distance set,
the pitch angle and
the separation length; calculating a height variation using the local slope
and a longitudinal
distance between consecutive ones of the plurality of positions along the
longitudinal
direction; adding the height variation to the longitudinal distance set to
generate a surface
height set.
BRIEF DESCRIPTION OF THE DRAWINGS
Having thus generally described the nature of the invention, reference will
now be
made to the accompanying drawings, showing by way of illustration example
embodiments
thereof and in which:
FIG. 1 (Prior Art) shows the longitudinal profile measurement trajectory in
the ideal
case;
-8--
CA 02802491 2013-01-21
290405-5 CA
FIG. 2 (Prior Art) illustrates an acquisition instrument used to measure the
longitudinal profile of the road which uses two single-point 3D sensors;
FIG. 3 (Prior Art) shows an alternative configuration for the longitudinal
profile
measurement instrument using a line of 3D sensors;
FIG. 4 (Prior Art) shows the longitudinal profile measurement trajectory with
a single
point system when lateral movement of the acquisition instrument is present;
FIG. 5 shows an example configuration for the acquisition instrument which
allows to
capture a full width transversal profile;
FIG. 6 illustrates the extraction of the 3D point on the ideal trajectory
using the 3D
transversal profile captured by the example acquisition instrument of FIG. 5
and feature
tracking;
FIG. 7 shows an example method used to compensate the lateral shifts of the
inspection vehicle;
Fig. 8 shows an intensity profile captured by an example system;
FIG. 9 shows the overlap between the fields of view of the sensors of an
example
configuration;
FIG. 10 shows how the information from the overlap can serve to determine the
slope
and the change in elevation of the collected data in a single acquisition;
FIG. 11 shows how the information from the overlap can be used in consecutive
acquisitions to estimate the slope;
FIG. 12 shows a vehicle climbing a hill while acquiring data and the relevant
angles
used for the calculation of the estimated slope; and
FIG. 13 shows an example method used to obtain the elevation profile for the
longitudinal profile when the fields of view of the sensors overlap.
¨9¨
CA 02802491 2013-01-21
290405-5 CA
It will be noted that throughout the appended drawings, like features are
identified by
like reference numerals.
DETAILED DESCRIPTION
The proposed method detects the position of longitudinal features such as lane
markings present on the road surface and uses this information to compensate
for the
variations in the transverse position of the inspection vehicle and therefore
the acquisition
instrument with respect to these longitudinal features. A transversely
oriented multipoint
range sensor is mounted on the survey vehicle in order to measure the distance
between the
vehicle and the surface. The multipoint range sensor can also be adapted to
detect the
presence of lane markings. The field of detection of the range sensor can have
partial or full
lane width.
Although described in relation with a road surface which bears lane markings
and on
which cars and trucks can circulate, the present method and system can be
applied to any
type of surface, such as a road, an airport runway, a tunnel lining, a train
track, etc. The
translation mechanism which displaces the sensors to acquire distance
information at a
plurality of positions along the longitudinal direction can be a car or truck
if the surface is a
road but can also be any type of vehicle, man driven or robotized, such as a
train wagon, a
plane, a subway car, a displaceable robot, etc.
In order to calculate longitudinal profiles, the multipoint laser range sensor
can also
be equipped with accelerometers or IMUs to measure the elevation changes of
the road and
the vertical oscillations of the inspection vehicle. FIG. 5 shows an example
configuration of
the inspection vehicle 502 and profiler acquisition instrument 500 for the
generation of a
lateral movement compensated longitudinal profile. The system includes two
high speed 3D
sensors 504 having fields of view 508 covering the width of the road 510. The
system can
also include IMUs 506. This configuration allows referencing of a measurement
with respect
to a tracking feature in the transversal profile.
¨ 10 ¨
CA 02802491 2013-01-21
290405-5 CA
The high speed 3D sensors can be any type of multipoint range sensors, such as
triangulation based laser line profilers, scanning point laser profilers,
lidar based scanning
point laser profilers, etc. The multipoint range sensors acquire distance
information at a
multitude of transversal points in their field of view to create a distance
set.
The IMUs are a type of elevation sensors. Global positioning system (GPS)
receivers
are another type of elevation sensors. The elevation sensors can generally be
multi-axis
accelerometers or vertically oriented single axis accelerometers.
The longitudinally-aligned tracking feature could already be present in the
road
infrastructure or could be added for the purpose of tracking the trajectory of
the acquisition
instrument. Examples of existing features are lane markings and reflectors.
Road surface
features such as ruts, texture or road side transitions (drop-off, edge,
curb), a joint, a concrete
slab edge, a road wheel path position, a road rut shape, a rail, a rail tie
etc. could also be used
as tracking features. In the case of ruts, the center of the rut corresponds
to the wheel path
and can be the measurement point for the IRI. Examples of added features are
painted
markings such as dots or lines used for guiding the measurement process.
The elevation and positional information could also be provided by an
elevation
sensor, for example a Global Navigation Satellite System (GNSS) such as GPS,
GLONASS
or Galileo.
Using feature tracking and a predefined distance relative to the feature one
can extract
the desired 3D measurement from the complete or almost complete transversal 3D
profile.
This process is repeated for each successive transverse profile to create the
longitudinal road
profile where the lateral shift of the inspection vehicle has been compensated
for.
FIG. 6 illustrates the extraction of the 3D point on the ideal trajectory 600
using the
3D transversal profile captured by the example acquisition instrument of FIG.
5 and feature
tracking. It compensates for the lateral shift, also known as the variation in
transverse
position, of the inspection vehicle. In this example case, the lane marking on
the right-hand
side 118 was tracked to compensate for the vehicle's lateral movement. The
vehicle
¨ 11 ¨
CA 02802491 2013-01-21
290405-5 CA
trajectory 402 is still wavy. However, the tracking feature is localized 604
and the 3D point
606 located at a distance d from the tracking feature can be extracted from
the 3D transversal
profile 608. The resulting longitudinal profile where lateral shifts have been
compensated for
610 is along a proper longitudinal trajectory.
FIG. 7 shows an example method 700 used to compensate the lateral shift of the
inspection vehicle. The transversal 3D profile is acquired 702. The tracking
feature is located
in either the transversal 3D profile or an intensity profile or in both using
signal processing
algorithms 704. The 3D point located at a predetermined distance "d" from the
tracking
feature is extracted from the transversal 3D profile 706. The extracted 3D
point is added to
the longitudinal profile of the road 708. The four steps are repeated for each
successive
transversal 3D profile acquired in the survey as the vehicle travels and
therefore moves the
acquisition instrument in order to obtain the complete longitudinal profile.
The intensity profile can be created from the transversal 3D profile, for
example using
a line by line 2D intensity profile or can be an intensity image obtained by
an additional
sensor, such as a still camera or a video camera.
In a simplified version of this example method 700, the position of the
longitudinally-
aligned feature of the surface relative to the field of view of the
acquisition instrument is
retrieved for each of a plurality of longitudinal positions. The predetermined
transversal
distance from the longitudinally-aligned feature at which to extract a
relevant distance from
the distance set is retrieved. A range point is extracted in the distance set
at the
predetermined transversal distance from the position of the longitudinally-
aligned feature.
The extracted range point is the relevant distance from the distance set. The
extracted range
point is added to generate a longitudinal distance set at a constant
transversal distance from
the longitudinal feature.
It will be readily understood that the position of the longitudinally-aligned
feature of
the surface relative to the field of view of the acquisition instrument can be
retrieved at some
of the longitudinal positions and extrapolated to be used at other
longitudinal positions. This
is particularly useful when the longitudinally-aligned feature is discontinued
in some sections
- 12 -
CA 02802491 2013-01-21
290405-5 CA
along the longitudinal direction. For example, a dashed line on a road surface
would be an
example of a discontinued longitudinally-aligned feature for which the
position of the feature
needs to be extrapolated from visible portions of the feature.
Using the acceleration and angular rate measured by the IMU and proper signal
processing, one can estimated the pitch and roll of the inspection vehicle. By
combining the
vehicle orientation information (pitch and roll) and the 3D measurements from
the sensors,
the road shape can be estimated and thus the elevation profile at the selected
location can be
extracted.
The tracking feature can be detected in the transversal 3D profile using one
or both of
the intensity and the range information captured in the 3D profile.
In the case where the tracking feature to be extracted is a marking painted on
the
road, the intensity data from the sensor or a camera can be used. For example,
the 3D sensors
may be able to determine the intensity of the light reflected back from the
surface. This
intensity data can be transformed into an image in grey-scale. Alternatively,
the intensity can
be from a color or a black and white obtained using an external camera or
device or a range
image. Generally, the marking will have a higher intensity than the pavement.
A simple
threshold operation can thus be applied to extract the location of the
marking.
Alternatively, the height of the paint for the painted lane marking could be
differentiated from the height of the surrounding road surface. If the
longitudinal feature is a
joint of a concrete slab on a concrete road, the longitudinal joint with a
height lower than the
surrounding surface, could be detected and tracked as the tracking feature. In
the case of a
rut, the presence of the deepest point in a rut on an asphalt road surface
could serve as the
longitudinal feature to be tracked.
In order to determine the distance d, the lane width can be supplied by the
user or
measured by detecting the lane marking on the road. The distance between the
two wheel
paths of the inspection vehicle is also known. From these two values, the
distance d from the
lane marking at which to take the inspection data can be computed as (Lane
_Width -
¨ 13 ¨
CA 02802491 2013-01-21
290405-5 CA
Wheel_path_distance) I 2. The distance d could also be supplied by the user to
suit the
requirements of the application.
As will be readily understood, once the tracking feature has been detected and
d is
known, it is possible to extract the 3D point in the 3D profile using signal
processing
algorithms.
Fig. 8 shows an intensity profile 800 captured by an example system. The
intensity of
the image is graphed as a function of the pixel index in the transversal
direction. The
intensity threshold 802 for the detection of the lane marking is identified at
an intensity of
175 gray levels, 255 being the maximum value in the image. The captured
intensity profile
804 contains a section 806 for which the intensity is higher than the pre-
determined intensity
threshold. Because the system is known to have imaged the lane marking, this
section which
has a higher intensity than the threshold is identified to be the lane marking
808. The chosen
pixel 810 at a pre-determined distance d from the feature identified in the
intensity profile is
also identified. The pixel index 812 for the chosen pixel can be extracted.
Then, the pixel
index can be used to extract the 3D point in the 3D transversal profile and
add it to the
longitudinal profile.
For example, in one test trial, the difference between the reference IRI value
measured with a walking profiler and an example system for measuring the IRI
installed on a
truck was reduced from 10% to 3% using the present compensation for lateral
movement of
the acquisition instrument using feature tracking.
As will be readily understood, the compensation for the variation in
transverse
position of the acquisition instrument can be done in real-time, as the data
is being acquired
by the multipoint range sensor. Alternatively, the compensation can be
performed off-line,
after acquisition along the longitudinal direction has ended and data has been
retrieved from
the acquisition instrument. It will be understood that the connection between
the acquisition
instrument and the processor which calculates the compensation and applies it
to the acquired
data can be a wired or wireless connection. The processor can be provided as
part or external
to the acquisition instrument. Additionally, the communication between the two
devices can
- 14 -
CA 02802491 2013-01-21
290405-5 CA
be carried over a network. Processing of the data can be split in sub-actions
carried out by a
plurality of processors for example using cloud computing capabilities.
In order to compensate for low speed variations of the translation mechanism,
it is
possible to measure the elevation profile of the road without using
accelerometers provided
there is a pitch finder instrument (for example a gyroscope, GPS or GNSS)
which measures
the pitch of the acquisition instrument 502 in the longitudinal direction (or
direction of
translation) relative to gravity and provided there is an overlap between the
field of view of
the sensors 504 in the instrument. This can be useful at low speed (for
example at a speed
less than 25 km/hr) where the weak vertical accelerations of the vehicle are
not accurately
measured by the accelerometers.
This low speed compensation can be performed independently of the compensation
for variations in the transversal direction, without tracking a longitudinal
feature or can be
combined with it to yield a longitudinal distance set which is compensated for
both the lateral
movements of the acquisition instrument and the low speed displacement along
the
longitudinal direction.
FIG. 9 shows a schematic representation 1100 of the fields of view of the two
sensors
of the instrument which overlap and shows how this overlap can serve to
determine the
elevation. The overlap of the fields of view is shown at 1102. The right hand
side of the field
of view of the left sensor 1104 overlaps with the left hand side of the field
of view of the
right sensor 1106. The overlap is a transversal overlap, meaning that a
plurality of points
acquired by the multipoint range sensor is acquired by both sensors. The
sensors are
separated by a separation length do in the longitudinal direction.
In FIG. 9, the fields of view of the 3D sensors are shown to be at an angle to
the
travel direction and not strictly aligned with the transversal direction. They
are still provided
along the transversal direction within the meaning of this specification. This
is useful in
preventing that each sensor captures the laser trace of the other sensor. It
also allows a better
capture of transversally aligned defects in the road surface and allows a
better calculation of
the slope. The angle of the 3D sensors with respect to the transversal
direction can be any
- 15 -
CA 02802491 2013-01-21
290405-5 CA
angle chosen between 00 and 45 . They could be provided at 150 to the
transversal
direction, for example. Also, as it apparent from FIG. 9, the 3D sensors can
be provided with
an overlap in the transversal direction and placed at a separation length from
one another in
the longitudinal direction. As will be readily understood, even if only one 3D
sensor was
being used, it could still be provided at an angle to the transversal
direction. As will also be
readily understood, even if two 3D sensors are not provided with an
overlapping field of
view, they could still be provided at an angle to the transversal direction.
Moreover, two 3D
sensors could be provided with an overlapping field of view even if the
present method for
compensating for low speed displacement of the acquisition instrument is not
performed.
As shown in Fig. 10, the overlap 1102 allows measuring 1200 the local slope
(0')
between the instrument 502 and the surface 100 using the distance of the
overlap do and the
difference between the heights of the captured data Ah (Ah = h1-h2) at each
sensor. The
height h2 captured by the left sensor is subtracted from the height h1
captured by the right
sensor.
FIG. 12 shows the relationship between the pitch of the instrument versus
gravity
(Opitch), the slope between the instrument and the road (0') and the real
local slope of the road
(0) 1400. Using pitch finder instrument 1402 (for example a gyroscope, GPS or
GNSS) to
measure the pitch (Opitch) and adding the slope (0') measured by the sensors
gives the local
slope of the road versus gravity (0). From this slope and knowing the distance
traveled (di)
between two successive profiles, one can estimate the real height variation of
the road (hi).
The elevation profile can be created by repeating this process for each
successive transversal
profile as follows:
Ah = hi-h2
0' = atan(Ah / do)
0 = 000+0'
hi = di x tan(0)
- 16 -
CA 02802491 2013-01-21
290405-5 CA
Fig. 11 further illustrates the estimation of the elevation profile using the
field of view
overlap. Elevations hi; and h21 are measured simultaneously by left and right
sensors
respectively. The local slope of the road surface can be estimated using these
two
measurements and knowing the distance traveled (di) between two successive
profiles, the
elevation change (111) can also be estimated. The elevation profile will then
be:
Elevation profile = [0, h1, h1 +h12, h11+11,2+h13, = = .]
FIG. 13 shows an example method for obtaining the height information 1300. The
transversal 3D profile is acquired 1302. The instrument (or vehicle) pitch is
acquired 1304.
The slope 0' is calculated using hi, h2 and do 1306. The local slope of the
road 0 is calculated
by adding 0' to Opiich 1308. Using the estimated slope 0 and the traveled
distance di, the height
variation hi is calculated 1310. The estimated height variation is added to
the longitudinal
profile 1312.
Example
In an example practical application, the lane of the road to be inspected
typically has
a width of 3.6 m. The length of the road along the longitudinal direction can
be anywhere
from a few meters to tens of kilometers. The predefined distance d from the
lane marking at
which to take inspection measurements is 90 cm. The inspection vehicle on
which are
installed the 3D sensors will travel at speeds up to 100 km/hr.
In an example system for the inspection, the 3D sensors have a transversal
field of
view at the road surface of 2 m with a longitudinal width for their field of
view of 1 mm.
They are installed at a height of about 2 m, on an inspection vehicle. The
inspection vehicle
can travel at speeds up to 100 km/h. The 3D sensors have a sampling rate of 5
600 profiles
per second in the longitudinal direction. The sampling spacing can be 1 to 5
mm and is
adjustable. There are 4096 transversal sampling points with a transversal
field of view of 4 m
and a transversal resolution of 1 mm. The depth accuracy is 0.5 mm. The
overlap do is 50 cm.
In this example system, the 3D sensors are laser profile scanners for 2D
profiles.
AcuityTM is a manufacturer of such non-contact laser scanning profilometers.
In this example
- 17 -
CA 02802491 2013-01-21
290405-5 CA
system, if IMUs are included, they can be obtained, for example, from the
manufacturer
STMicroelectronicsTm. An example IMU which would adequate for the present
system is
model LSM330D. The LSM330D is a system-in-package featuring a 3D digital
accelerometer and a 3D digital gyroscope. The LSM330D has dynamically user-
selectable
full scale acceleration range of 2 g/ 4 g/ 8 g/ 16 g and angular rate of
250/ 500/ 2000
deg/sec.
The embodiments described above are intended to be exemplary only. The scope
of
the invention is therefore intended to be limited solely by the appended
claims.
- 18 -
