DAAAM INTERNATIONAL SCIENTIFIC BOOK 2022 pp.

121-132 Chapter 10

TOOL WEAR MEASUREMENT USING

MACHINE VISION SYSTEM
MUNDAR, G. & ZUPERL, U.
Abstract: When designing a machining process with cutting, one of the key influencing
factors is tool wear, which occurs as a result of the action of various forces during the
machining process. In accordance with the modernization of modern production, our
work includes the development of an optical measuring cell for the control and
measurement of cutting tools. This covers the planning, design and manufacture of the
measuring cell, which enables the attachment of the camera and its movement in the x
and y directions, as well as the appropriate clamping of the tool. It also includes the
implementation of the algorithm for capturing the image from the camera and its
processing with edge detection, converting the camera's coordinate system into the
tool's coordinate system, and converting pixels to millimetres. The final output
parameter of the measurement is the value of the x and y components of the tip of the
cutting edge.

Key words: tool wear, machine vision, camera, SICK AppStudio, measurement system

Authors´ data: MSc. Munđar, G[oran]*; Prof. Dr. Sc. Župerl, U[roš]*, * University
of Maribor, Faculty of mechanical engineering, Laboratory for Mechatronics,
Smetanova ulica 17, 2000 Maribor, Slovenia, [email protected],
[email protected]

This Publication has to be referred as: Mundar, G[oran] & Zuperl, U[ros] (2022).
Tool Wear Measurement Using Machine Vision System, Chapter 10 in DAAAM
International Scientific Book 2022, pp.121-132, B. Katalinic (Ed.), Published by
DAAAM International, ISBN 978-3-902734-34-1, ISSN 1726-9687, Vienna, Austria
DOI: 10.2507/daaam.scibook.2022.10

121
 Mundar, G. & Zuperl, U.: Tool Wear Measurement Using Machine Vision System
1. Introduction

The modernization of modern production dictates investments in new upgrades

and the integration of advanced technologies into production processes, as well as the
introduction of measures with as little human intervention as possible. When designing
a machining process with cutting, one of the key influencing factors is tool wear, which
occurs as a result of the action of various forces during the cutting process. The result
of tool wear is first of all a reduced quality of the treated surfaces, and then a lower
accuracy of achieving the desired dimensions of the workpiece (Teti et al., 2010). In
order to ensure high-quality processing and thus greater competitiveness, we want the
impact of tool wear to be as small as possible. Nowadays, the problem is solved by the
CNC operator removing a specific tool from the machine tool every now and then,
taking the tool to a measuring device, where the actual size of the tool is measured.
Operator then updates the tool database with corrections and these tool corrections are
then taken into account in further processing (Mohanraj et al., 2020).
One of the most innovative monitoring applications is Tool condition monitoring
(TCM), which is inevitable for reducing machine downtime (Denkena et al. 2009). One
of the main causes of downtime is excessive wear and breakage of cutting tools.
Damaged tool can therefore decrease quality of machined surface, due to unevenly
distributed forces during cutting process. With appropriate TCM techniques (Rehorn
et al., 2005) cutting speeds can be increased by 10-50 %.
In accordance with the modernization and automation of modern production, the
idea arises that tool wear could be measured directly during the processing itself. The
tool would be placed in the area inside the machine tool where the machine vision
system is located and automatic measurement and correction of the tool would be
carried out without human intervention. This would save time, achieve a higher level
of automation, reduce the possible influence of the human factor and ensure accurate
processing. Based on statistical data and artificial intelligence, it would also be possible
to predict in advance when cutting tool will be so worn that it should be replaced (Dutta
et al., 2013).
Basically, TCM systems can be divided into two groups: direct techniques and
indirect techniques (Li & Tian, 2013). In the direct techniques conditions such as flank
wear width, crater depth and crater area are measured directly in off-line method
(machining process must be stopped during measurement) using 3D surface profiler,
electron microscope or optical microscope. Previously mentioned condition can also
be measured using in-process methods (it does not require stoppage of the machining
process) with CCD camera. In indirect TCM techniques the following signals of cutting
process are measured: force, current, power, surface finish, acoustic emission, etc.
Measurement of those signals allow conclusions to be drawn about the degree of tool
wear using in-process methods. These TCM systems are typically based on comparing
a reference signal from an optimized cutting process to the actual process signal
acquired from previously mentioned sensors (Wang & Gao, 2006).
These techniques have been implemented mainly by using various technologies
such as acoustic emissions, cutting force, spindle current and vibration sensors (Pfeifer
& Wiegers, 2000).

122
DAAAM INTERNATIONAL SCIENTIFIC BOOK 2022 pp. 121-132 Chapter 10
However, there are also some very serious limitations when using these methods.
To overcome these limitations, much research is currently being conducted to
determine the degree of tool wear by analysing different images acquired using
different optical sensors such as lasers, CCD and CMOS cameras, and thermal IR
cameras. There are a wide range of applications that combine optical sensors with
digital image processing and machine vision which are used for quality control, tool
wear measurement, workpiece surface measurement, etc. in machining processes such
as milling (Župerl et al., 2022).
As part of the work, we were involved in the development of an optical system
for the control of cutting tools in the Sick AppStudio software package and
implemented it in a prototype measuring model, which enables reliable measurements
of the dimensions of rotationally-symmetric cutting tools, tool wear and the detection
of possible damage to the tool. The measuring cell, consisting of aluminium profiles,
allows the tool to be clamped and fixed in the desired place, while allowing the tool to
rotate around its axis, so that each tile on the tool can be measured by rotating the tool.
A high-speed camera with appropriate lighting is fixed perpendicular to the work plate.
The position of the camera can be set according to the size of the tool we want to
measure by moving the camera via guides in the x and y directions.

2. Measuring cell

It was necessary to design, draw and manufacture a measuring cell that would
allow the camera to be attached and moved in the x and y directions. With this, we
were able to cover the entire measuring range and measure various tools regardless of
their size. It was also necessary to find a solution for clamping the tool, which will
ensure that it is always fixed at the same, precisely determined place, which we
determined according to the focal length of the camera. At the same time, it was
necessary to allow the tool to rotate around its axis in order to be able to measure each
individual cutting edge on the tool. The focal length of the camera lens had to be taken
into account in order to obtain a sufficiently high-quality image, which will then be
suitable for further processing. Since the measurement of tool wear with the help of a
camera is also significantly affected by light disturbances from the surroundings, it was
necessary to ensure constant lighting and to prevent the penetration of external light
into the measuring cell.

2.1 Design of the measuring cell

The 3D model of the measuring cell was designed in the Solidworks software
tool based on the requirements dictated by the application and described in the previous
paragraph. The tool had to be clamped and fixed in the desired position, while allowing
it to rotate on its axis, which would allow each cutting edge on the tool to be measured.
It was necessary to attach a camera perpendicular to the tool, and to install a backlight
light behind the tool. We had to find a solution so that the camera can move in the x
and y directions in order to measure different tools of different sizes and make sure that
the distance between the camera and the cutting edge of the tool is a constant focus
distance. The 3D model of the measuring cell that was developed is shown in Fig. 1.

123
 Mundar, G. & Zuperl, U.: Tool Wear Measurement Using Machine Vision System

Fig. 1. 3D model of the measuring cell

2.2 Tool holder design

For ease of machining, the tool holder was made of aluminium and was then
inserted into a radial bearing with housing. The tool holder is tapered for the SK 50
taper, which enables stable positioning and clamping of the tool. It is mounted with a
radial bearing and allows the tool to rotate freely in order to be able to measure each
cutting edge on the tool.

Fig. 2. a) 3D model of tool holder, b) Actual implementation of tool holder

124
DAAAM INTERNATIONAL SCIENTIFIC BOOK 2022 pp. 121-132 Chapter 10
The movement of the tool holder is possible in the x direction via the four screws
with which the bearing housing is attached to the aluminium profiles, or in the y
direction by moving two aluminium profiles along the T-groove. In principle there is
no need to move the tool, as this position can be set via camera movement, where the
position can be set more precisely and continuously.
We determined the position of the tool in the z direction at the beginning and
blocked it. The 3D model of the tool holder is shown in Fig. 2a. The actual
implementation of the clamping in the measuring cell is shown in Figure 2b.

2.3 Camera mount design

The linear movement of the camera in the vertical and horizontal directions was
implemented via round guides, which were attached to the aluminum housing. The
position of the camera is determined using polyethylene carts, which were made from
PETG material using 3D printer. Releasing and fixing the camera position is set via
four threaded buttons that are screwed to the guide in the position in which we want to
fix the camera position. With this, we ensured accurate and consistent changes in the
camera position. To ensure a constant focus distance, the movement in the z direction
is disabled. The model of the camera attachment is shown in Fig. 3, where the two
possible directions in which the camera can be moved are indicated.

Fig. 3. Model of the camera attachment

2.4 Calibration grid installation

The main purpose of using the calibration grid is to convert the camera coordinate
system into the tool coordinate system. We had to install the calibration grid and
prevent movement in the x and y directions and its rotation. It was also necessary to
align the coordinate origin of the calibration grid with the coordinate origin of the tool.
We were able to ensure accurate measurements and their reproducibility only with
high-quality clamping.

125
 Mundar, G. & Zuperl, U.: Tool Wear Measurement Using Machine Vision System
3. Machine vision system

This chapter will present the hardware and software used in the development of
a machine vision system for determining cutting edge wear.

3.1 Hardware
The built system consisted of the following hardware:
• Camera SICK picoCam304x
• Sensor integration machine SICK SIMx4000
• 24 V power supply
• LED backlight
• Telecentric lens

The tool's image will be captured by the SICK camera, which has a resolution of
2048 px x 2048 px and enables the capture of images at a speed of 19 fps. The camera
will be fitted with a telecentric lens that allows for 27mm x 27mm image capture.
The camera will be connected to the SICK SIMx4000 sensor integration machine
via an Ethernet communication interface. All machine vision algorithms will be
implemented inside the SICK SIMx4000, and at the same time the machine will also
take care of sending the data to online cloud, which can then be accessed via a personal
computer, tablet or phone. The 24 V power supply will provide power for the following
components: the camera, the sensor integration machine, and the LED backlight.

3.2 Software
SICK AppStudio software tool was used to create an application for determining
the wear of cutting edges. SICK AppStudio is used for developing customer-specific
applications on programmable SICK devices. It includes Flow Editor and Lua script
programming technologies for creation of sensor applications.
As part of the research, we used SICK AppStudio to create an application that
took care of capturing the image via the camera, processing the image, calculating the
diameter of the tool and sending the data to the cloud. At the same time, as part of the
development of the application, we also created a user interface that will allow users
to perform measurements more easily in the measuring cell. The developed application
will be described in more detail in the next chapter.

4. Application for determining tool wear

Tool wear measurement is monitored by a camera, which transmits the captured

image to the sensor integration machine via an Ethernet connection. We have divided
the program and user interface into three sections, which are implemented step by step.
The first part consists of finding and setting the position of the camera according to the
tool that is clamped in the tool holder. Vertical and horizontal lines are helpful in this
step. We locate the tip of the machining tool at the intersection of these two lines. At
this point, we fix the position of the camera via threaded buttons.

126
DAAAM INTERNATIONAL SCIENTIFIC BOOK 2022 pp. 121-132 Chapter 10
The second part consists of converting the camera coordinate system to the tool
coordinate system and finding the rotation angle of the image. A calibration grid with
a 20-mm grid, which was installed at the place where of tool holder, was used for
coordinate system conversion.
The third part consists of finding the cutting edge of the tool and performing
measurements of the cutting edge in the x and y directions. To analyse the measurement
of tool wear, it was necessary to properly pre-process the captured image and then,
using the functions included in AppStudio, obtain information about the x and y
components of the tool cutting edge.

4.1 Step 1: Positioning the camera

In the first step, it is necessary to determine the area in which the captured image
should be located. This was done by using a vertical and horizontal line as shown in
Fig. 4. It is then necessary to physically position the tool in the 4th quadrant of the
image, so that the tip of the tool is almost exactly at the intersection of the two lines.
To draw a line in the horizontal and vertical directions, it was necessary to determine
the starting and ending points, which are specified in the number of pixels. Here, the
first component represents the number of pixels in the x direction and the second in the
y direction.

Fig. 4. First step of tool measurement procedure

4.2 Step 2: Converting the coordinate system

The second step consists of linking the coordinate systems of the camera and the
tool and rotating the image. To convert the camera's coordinate system into the tool's
coordinate system, we used a 20 mm x 20 mm calibration grid. Firstly, it was necessary
to remove the tool from the measuring cell and install the mentioned calibration grid
inside it. After pressing the "Find measuring grid" button, the application captures the
image of the calibration grid and finds the intersection of the vertical and horizontal
lines in the image. The application then finds the angle of rotation between the
horizontal line in the image and the bottom edge of the image.

127
 Mundar, G. & Zuperl, U.: Tool Wear Measurement Using Machine Vision System
The coordinates of the intersection of the found lines and the angle of rotation
are displayed on the screen, as shown in Fig. 5a. The application then searches for any
numbers in the image. The numbers written on the grid represent the x and y distances
of the intersection on the calibration grid, located to the right below the numbers. The
upper number represents the x and the lower y distance of the intersection based on
tool coordinate system origin. The application in the image checks whether the
numbers are located on the right or on the left side of the found vertical line and
accordingly connects the coordinate system of the intersection in the image with the
coordinate system on the calibration grid. At the end of the second step, the application
rotates the image so that the horizontal line found is parallel to the bottom and top
edges of the image, as shown in Fig. 5b.

Fig. 5. Second step of tool measurement procedure

4.3 Step 3: Measurement the cutting edges

The main objective of the third step is to find the cutting edge of the tool and the point
with the largest diameter. Firstly, it was necessary to remove the calibration grid from
the cell and re-install the tool in it. The tool must be rotated so that the first cutting
edge of the tool is clearly visible in the image. After pressing the "Measure diameter"
button, the application measures the diameter of the tool. When the button is pressed,
the image is rotated by the angle calculated in the second step described in chapter 4.2.
The application then draws a small rectangle on the image, which can be seen in the
Fig. 6. Within this rectangle, each point of the cutting edge of the tool is then located.
Due to the use of background lighting, which provides an intense contrast between the
background and the observed object, it is possible to easily determine the cutting edge
of the tool, which is black in the image. After successfully finding the cutting edge
points, the application determines the point on the tool with the smallest x coordinate.
This is the point that represents the maximum diameter of the tool. The tool diameter
is calculated using equation (1):

128
DAAAM INTERNATIONAL SCIENTIFIC BOOK 2022 pp. 121-132 Chapter 10
𝐷 = 2 ∙ (𝑟0 ± 𝑘 ∙ |𝑥1 − 𝑥0 |) (1)

where 𝐷 is the largest diameter (expressed in mm) of the cutting edge of the tool,
𝑟0 is the distance in the x direction between the vertical line from step 2 and the centre
of the tool holder (expressed in mm), 𝑘 is the resolution of the measurement system,
𝑥0 is the x coordinate on the image (expressed in px) of the intersection of the lines
from step 2 and 𝑥1 is the x coordinate on the image (expressed in px) of the point on
the tool with the largest diameter.
The resolution of the measurement system can be calculated from the resolution
of the camera and the field of view of the lens. The resolution of the camera we used
in the system is 2048px x 2048px. The field of view of the lens is 27 mm x 27 mm.
Equation 2 shows the calculation of the resolution of the used measurement system:

𝑘 = 𝐹𝑂𝑉𝑥 ⁄𝑟𝑐 = 27 𝑚𝑚⁄2048 𝑝𝑥 = 0.01318 𝑚𝑚/𝑝𝑥 = 13.18 𝜇𝑚/𝑝𝑥 (2)

where 𝐹𝑂𝑉𝑥 is the field of view of the lens in the horizontal direction and 𝑟𝑐 is
the resolution of the camera in the horizontal direction. We can see that the resolution
of the developed system is 13,18 µm/px.

Fig. 6. Third step of tool measurement procedure

Using equations 1 and 2, the application calculates the tool diameter. Then the
application draws on the screen the found cutting edge, the point with the largest
diameter and prints the calculated diameter of the tool in mm as shown in Fig. 6.

4.4 Application control panel

The application control panel was created using the html programming language.
The control panel enables the connection between the user and the measuring system.
The control panel can be accessed by entering the correct IP address in a browser on a
computer, tablet or phone when the device is connected to the same network as the
SIMx4000 sensor integration machine.

129
 Mundar, G. & Zuperl, U.: Tool Wear Measurement Using Machine Vision System
The control panel of the application consists of four images, three buttons and
instructions for performing measurements. It is shown in Fig. 7.
The left image named "Live image" shows the live image captured by the
camera. By pressing the "Grid" button, the vertical and horizontal lines can be turned
on or off.
The upper middle image named "Measuring grid" shows the image of the
calibration grid obtained by pressing the button "Find measuring grid" after placing the
calibration grid in the measuring cell. The coordinates of the found intersection point
and the image rotation angle are displayed on the image. In the middle image below,
with the name "Rotated measuring grid", a rotated image of the calibration grid and the
coordinates of the found intersection are shown.
The right image named "Diameter measurement" shows the image of the tool
when the "Measure diameter" button is pressed after calibration grid was successfully
found in the previous step and the tool was placed back into the measuring cell. The
image also shows the calculated diameter in mm.

Fig. 7. Application control panel

In order to successfully measure all the cutting edges of the tool, it was necessary
to manually rotate the tool and measure each cutting edge separately by pressing the
"Measure diameter" button.

5. Results

The measurement results obtained with the developed measuring system were
compared with measurements using a caliper and with measurements on the Zoller
“smile 420” optical measuring system (***).

130
DAAAM INTERNATIONAL SCIENTIFIC BOOK 2022 pp. 121-132 Chapter 10
We compared the results of the measurements of the four cutting edges of the
tool and compared the results when using three different measurement systems. As a
reference, we used the Zoller measuring system, which allows a more accurate
measurement than caliper. The results of the measurements are shown in Table 1.

Measuring system Measured diameter [mm] Average Error [%]

[mm]
Zoller “smile 59.580 59.576 59.572 59.55 59.570 Reference
420”
Caliper 59.45 59.40 59.53 59.38 59.44 0.219 %
Our system 59.552 59.522 59.542 59.467 59.521 0.082 %
Tab. 1. Results of measurements

From the results, we can see that using the developed measuring system we
achieved better results than using a calliper. The relative measurement error of the
constructed system compared to the Zoller measuring system was 0.082%. The relative
error when using calliper was 0.219%, which is 2.6 times more than the error of our
developed system.

6. Conclusion

When we talk about the modernization of production and the fulfilment of the
requirements dictated by Industry 4.0 and Industry 5.0, which is fast approaching, the
need to achieve higher productivity, reducing wear and damage on the tool are crucial
for ensuring competitiveness on the market.
The test optical measuring cell for tool wear control that we have developed
meets all the needs for tool wear control, based on which we can perform tool
correction on a CNC machine. Regardless, it has certain drawbacks and allows for
additional expansions and upgrades. Based on the measurement accuracy calculation
we performed in Chapter 5, we found that our system does not give satisfactory results
in terms of measurement accuracy for serious tool wear measurements. However, the
resolution of the camera is quite sufficient for the needs of testing the tool condition
with the help of machine vision.
The optical tool wear monitoring system also offers other options for upgrading
the existing measuring cell. Using a motor drive to move the camera in the x and y
directions and using limit switches would simplify the determination of the tool
coordinate system and the conversion of the camera coordinate system to the tool
coordinate system, since the position in which both limit switches have a signal would
represent the reference position. Therefore, there would be no need to use a calibration
grid, as we did now. With each movement of the camera, system would know exactly
where it is located. Also, with the motor drive of the tool rotation, the measuring cell
would be fully automated and there would be no need for human intervention in the
measuring cell for each individual measurement.

131
 Mundar, G. & Zuperl, U.: Tool Wear Measurement Using Machine Vision System
The implementation of a machine vision system to control tool wear in the
machining process on a CNC machine would also require the connection of the optical
system to the CNC machine controller and the installation of measuring system in the
CNC machine itself.
Future activities will be carried out to evolve the presented tool wear
measurement system with automation, such as automatic camera movement and
automatic tool rotation, which will simplify and speed up the measurement process for
the operator. The presented system will serve as a basis for the creation of a cognitive
cyber-physical control system for controlling the state of the tool during the machining
process, which we want to develop in the future.

7. References

Denkena, B., Bluemel, P., Roebbing, J., & Kroening, S. (2009). Condition-Based
Maintenance of Machine Tools. Annals of DAAAM & Proceedings.
Dutta, S., Pal, S. K., Mukhopadhyay, S., & Sen, R. (2013). Application of digital image
processing in tool condition monitoring: A review. CIRP Journal of Manufacturing
Science and Technology, 6(3), 212-232.
Li, D., & Tian, J. (2013). An accurate calibration method for a camera with telecentric
lenses. Optics and Lasers in Engineering, 51(5), 538-541.
Mohanraj, T., Shankar, S., Rajasekar, R., Sakthivel, N. R., & Pramanik, A. (2020).
Tool condition monitoring techniques in milling process—a review. Journal of
Materials Research and Technology, 9(1), 1032-1042.
Pfeifer, T., & Wiegers, L. (2000). Reliable tool wear monitoring by optimized image
and illumination control in machine vision. Measurement, 28(3), 209-218.
Rehorn, A. G., Jiang, J., & Orban, P. E. (2005). State-of-the-art methods and results in
tool condition monitoring: a review. The International Journal of Advanced
Manufacturing Technology, 26(7), 693-710.
Teti, R., Jemielniak, K., O’Donnell, G., & Dornfeld, D. (2010). Advanced monitoring
of machining operations. CIRP annals, 59(2), 717-739.
Wang, L., & Gao, R. X. (Eds.). (2006). Condition monitoring and control for intelligent
manufacturing. Springer Science & Business Media.
Župerl, U., Stepien, K., Munđar, G., & Kovačič, M. (2022). A Cloud-Based System
for the Optical Monitoring of Tool Conditions during Milling through the Detection of
Chip Surface Size and Identification of Cutting Force Trends. Processes, 10(4), 671.
*** ZOLLER Inc., Presetting and Measuring Machine »smile«: The Entry-Level for
the Professional Presetting and Measurement of Standard Tools, Accessed on: 2022-
09-09

132