MACHINE VISION FOR ICE LAYER THICKNESS MEASUREMENTS
I. Hermanto & R. E. Gagnon*
Canpolar East Inc.
44 Austin Street, St. John's, NF, Canada
A1B 4C2
Tel: (709)722-6067
Email:info@canpolar.com
Web: http://www.vetech.com
R. E. Gagnon*
Institute for Marine Dynamics
National Research Council of Canada
St. John's, Newfoundland, Canada, A1B 3T5
ABSTRACT
A patented technique for remotely measuring layer thickness of transparent fluids and solids on surfaces has been incorporated into Canpolar East's Smart Camera (VE-262) and been validated for measurement of ice layer thickness. This application addresses the need for reliable detection of icing on aircraft. The thickness measurement technique is carried out by directing a laser beam at the surface with the ice layer. A pattern of illumination appears on the surface as a result of internal reflection of the laser light within the ice layer. The layer thickness is calculated by a simple formula involving the size of the image pattern and the refractive index of layer material. The system was tested using a wedge-shaped ice sample with a maximum thickness of about 10 mm. The data indicated a system measurement accuracy of 0.1 mm.
1.0 INTRODUCTION
Icing on aircraft components has lead to fatal crashes throughout the history of modern aviation and remains to be adequately addressed. Detection of icing is important to alert pilots of the need for mitigating measures while in flight and also to assess de-icing requirements before takeoff. Several techniques are currently available to carry out ice layer thickness measurements1. The majority of these sensors are, however, based on either mechanical or acoustical techniques of a contact type. This paper reports a remote thickness measurement system consisting of a patented method2 for remote measurement of transparent layer thickness coupled to a machine vision system with built in image analysis software. The measurement technique, prior to addition of the machine vision component, has already been used to measure ice accumulation on a model air foil and a model helicopter rotor3, and de- and anti-icing fluid thickness during simulated takeoff in icing tunnel experiments using a model wing4. Here we present ice layer thickness measurements which demonstrate the effectiveness, accuracy and simplicity of the complete system.
2.0 DESCRIPTION OF ICE THICKNESS MEASUREMENT TECHNIQUE
The technique makes use of the refractive property of ice,
or liquid, to remotely detect and measure its thickness on a
surface. A laser beam of any visible wavelength or infrared
laser with a wavelength outside of the region where ice
strongly absorbs (i.e. not greater than 1 m) is used. The
beam is directed towards the surface on which a layer of ice
is present (see Figure 1). The angle of incidence is not
important and can be very large. The beam traverses the ice
layer and impinges on the surface, producing an intense
bright spot from which light scatters in all directions (for
diffuse reflecting surfaces). All of the light scattered from
the spot and striking the ice/air interface at an angle of
incidence less than a certain angle from the surface, passes
through the interface. Those rays incident at angles greater
than get internally reflected from the ice/air interface to
strike and illuminate the surface again. Consequently the
view from above, normal to the surface, shows a bright spot
where the laser first strikes the surface, in the center of a
dark circular region with a bright perimeter (diameter D)
that drops in intensity as the distance from the center
increases. The view of the surface from other than normal
will show an ellipse rather then a circle. The angle is a
function of only the refractive index (n) of ice, 1.312.
According to the simple geometry of the situation, as
illustrated in Figure 1, the thickness (H) of the ice is given
by
where
For water ice 
is approximately 50o.
To remotely view the spot where the laser strikes the surface, and to measure the diameter of the bright circle, a video camera with a telephoto lens is used. Ordinary CCD video cameras are sensitive to infrared radiation in the range specified above, if an infrared laser is used rather than a visible laser source. The camera can be positioned at the same location as the laser or at some other site. The diameter of the bright circle can be measured directly off a video monitor or from captured video images using suitable software. The distance from the surface at which measurements can be made depends on the power of the telephoto lens on the video camera. It is important that the diameter of the laser beam be narrow. This affects the sharpness of the bright circle from which measurements are made. For very thin ice, i.e. less than 500 m, the perimeter of the bright circle is less than the laser spot itself and cannot be measured. However, the diameter of the circular zone comprised of the rays of light that get internally reflected at angles less than theta can be measured and the ice thickness correspondingly estimated.
The technique can be used to determine the thickness of ice on moving objects. In cases where objects with the ice layer are moving quickly and periodically through the same space, e.g. propellers, wind turbines, helicopter rotors, fans, etc., they have the appearance of being stationary, and the method and apparatus described above can be used, where the measurements are made directly from a video monitor or from captured video images corresponding to the times when the periodically moving surface appears in the field of view. The latter method would be used in the case of slower moving surfaces.
3.0 ICE THICKNESS MEASUREMENT EXPERIMENT
3.1 EXPERIMENTAL SETUP
To test the accuracy and sensitivity of the ice thickness
measurement system, including the machine vision and
image analysis algorithm, a series of thickness
measurements was made on a laboratory-grown piece of
freshwater ice. The ice was columnar grained with no
internal air bubbles present. The ice was cut into a wedge
shape with a wedge angle of 4.46o . The base of the wedge
was about 15 cm in length. This was frozen onto a light
colored flat surface (see Figure 2). The diode laser was
mounted near the Smart Camera VE-262 (see Figure 3) and
directed at the ice wedge. The VE-262 was located about
56 cm directly above the ice, viewing downwards so that
the spot where the laser beam impinged on the ice was in
the central viewing area. Measurements were made of the
ice thickness as the ice wedge was moved horizontally in a
series of 7.62 mm steps. This effectively increased the ice
thickness in a linear and precise manner. The images were
elliptical because of the wedge shape (see Figure 4) of the
ice, where the long axis was along the wedge axis. The
short axis of the elliptical image, along which the ice
thickness is constant, was used for the measurements.
The Smart Camera is a complete vision system with built-in 486 PC, DSP frame grabber, hard disk, VGA video and other communication interfaces. It comes with DOS and WindowsTM operating system plus image processing library.
Figure 3. VE-262 Smart Camera
Figure 4. Raw(left) and Processed (right) Images
3.2 IMAGE ANALYSIS ALGORITHM
Figure 5 shows the image processing flow diagram for the ice thickness measurement. The purpose of image analysis is to extract the diameter of the darker ring (see Figure 4). The first stage in the post processing is the removal of noise. This should be carried out in such a way so that edge detail is preserved. The noise removal technique used to accomplish this is a 3x3 median filter. This filter was applied 5 times. The next stage is to accentuate the separation between the darker and the brighter regions by first growing the brighter area of the image with a grayscale erosion operator 3 times and then growing the darker region back 3 times with a grayscale dilation operator. The image was then binarized using a user defined threshold level. (Note that real-time measurement requires some kind of automatic thresholding). One iteration of a binary opening operation was applied to the binarized image to further clean up single pixel objects and finally the blob analysis was applied to the image. During the analysis, blob pixels whose area is smaller than 20 pixels are not processed. Blob features such as the feret_x (length of blob in the x-direction) were then extracted. This diameter value is then converted from pixels to millimeters using the calibration conversion factor.
All of the image processing functions mentioned above are available as icons from the VE-Tools visual programming software which comes with the VE-262 Smart Camera. The WindowsTM Graphical User Interface was also created using the VE-Tools. VE-Tools allows application developers to rapidly create WindowsTM based image processing applications without having to write a single line of code.
3.3 TEST RESULTS
Two sets of measurements were obtained (see Table 1), each consisting of 9 or 10 points. Figure 6 shows the results of both sets of the measurements. Least squares linear fits were obtained for each data set and, as can be seen from the Figure 6, the fits were excellent (r > 0.999). The relative accuracy of the data is indicated by the standard error in the regression lines, corresponding to about 0.05 mm. The absolute accuracy is obtained from the ice wedge angle determined from the slopes of the two data sets. The slopes yielded wedge angles of 4.43o and 4.53o. Hence the calculated wedge angle from either data set agrees to within 1.5% with the actual value, where this level of accuracy approaches the accuracy at which the wedge angle could be measured in the first place. This corresponds to an absolute thickness measurement accuracy of 0.1 mm for the ice thicknesses measured here, which includes any systematic errors in the image analysis technique.
4.0 CONCLUSION
A novel ice thickness measurement system had been implemented and tested successfully using a low cost laser diode as the illuminating source and the VE-262 Smart Camera with the VE-Tools visual programming software. In laboratory tests the system was shown to provide measurement accuracy on the order of 0.1 mm for ice thicknesses up to 1 cm. Potential applications are the measurement of ice on fixed and moving aircraft components, the inspection of airport runways for ice and water conditions and thickness measurement of layers of industrial fluids and solids in factories.
5.0 REFERENCES
1. FAA, "Aircraft ice detectors and related technologies for onground and inflight applications". Report DOT/FAA/CT-92/27, FAA Technical Center, Atlantic City, NJ, (1993).
2. R.E. Gagnon. "Method and Apparatus for Remote Detection and Thickness Measurement of Ice or Liquid Layer". US Patent 5,400,144, March 21, 1995.
3. R.E. Gagnon and D.L. Marcotte. "Icing Thickness Measurements from an Airfoil and Model Helicopter Rotor Using a New Remote Technique", Proceedings of the American Helicopter Society International Icing Symposium `95, Montreal, Canada, September, 1995, 417-428.
4. M.M. Oleskiw, P.J. Penna, R.S. Crabbe and M.E. Beyers "Full-Scale Wind-Tunnel Simulation of Takeoff
Performance Degradation with Contaminated Fluid Runback", Proceedings of the American Helicopter Society
International Icing Symposium `95, Montreal, Canada, September, 1995, 179-190.
44 Austin Street, St. John's, NF, Canada
A1B 4C2
Phone: (709) 722-6067
Fax: (709) 722-1138
http://www.vetech.com/
e-mail: info@canpolar.com