Device to evaluate cleanliness of fiber optic connectors using image processing and neural networks

International Journal of Electrical and Computer Engineering (IJECE)
Vol. 11, No. 4, August 2021, pp. 3093~3105
ISSN: 2088-8708, DOI: 10.11591/ijece.v11i4.pp3093-3105  3093
Journal homepage: http://ijece.iaescore.com
Device to evaluate cleanliness of fiber optic connectors using
image processing and neural networks
Victor Fernandez, Javier Chavez, Guillermo Kemper
School of Electronic Engineering, Faculty of Engineering, Universidad Peruana de Ciencias Aplicadas, Lima, Peru
Article Info ABSTRACT
Article history:
Received Aug 19, 2020
Revised Dec 18, 2020
Accepted Jan 13, 2021
This work proposes a portable, handheld electronic device, which measures
the cleanliness in fiber optic connectors via digital image processing and
artificial neural networks. Its purpose is to reduce the evaluation subjectivity
in visual inspection done by human experts. Although devices with this
purpose already exist, they tend to be cost-prohibitive and do not take
advantage of neither image processing nor artificial intelligence to improve
their results. The device consists of an optical microscope for fiber optic
connector analysis, a digital camera adapter, a reduced-board computer, an
image processing algorithm, a neural network algorithm and an LCD screen
for equipment operation and results visualization. The image processing
algorithm applies grayscale histogram equalization, Gaussian filtering,
Canny filtering, Hough transform, region of interest segmentation and
obtaining radiometric descriptors as inputs to the neural network. Validation
consisted of comparing the results by the proposed device with those
obtained by agreeing human experts via visual inspection. Results yield an
average Cohen's Kappa of 0.926, which implies a very satisfactory
performance by the proposed device.
Keywords:
Cleanliness evaluation
Fiber optic connectors
Image processing
Neural networks
This is an open access article under the CC BY-SA license.
Corresponding Author:
Guillermo Kemper
School of Electronic Engineering, Faculty of Engineering
Universidad Peruana de Ciencias Aplicadas
Lima 15023, Perú
Email: guillermo.kemper@upc.pe
1. INTRODUCTION
Throughout the ages, humanity has faced the engineering challenge of transporting large quantities
of information from one place to another, which has been faced through diverse means. During the 30s, the
introduction of the coaxial and multipair cable technologies enabled transporting signals during many years.
Later on, the increasing telecommunications demand and digitalization trends resulted in the development of
a new technology which took off during the 60 s, known as fiber optics [1]. Meanwhile, other data transport
technologies appeared such as microwaves and satellite communications, which use atmospheric properties
for data transfer. However, each technology has its own operating conditions. For example, microwaves are
severely affected by weather conditions and have limited capacity depending on the requirement [2]. On the
other hand, the fiber optic communications requires that the connectors used to join the cables are completely
clean and free of any kind of dirt to ensure the quality of the services. Otherwise, there will be a negative
impact on the transmitted signal, which will impair the performance of the link [3, 4].
With this in mind, industry has developed technology both for identifying dirty connectors and for
cleaning them. IEC 61300-3-35:2015 is the standard that sets the quality requirements for fiber optic
connector terminations [5]. That is how the main problem to solve is develop a device that include strong

 ISSN: 2088-8708
Int J Elec & Comp Eng, Vol. 11, No. 4, August 2021 : 3093 - 3105
3094
image algorithms to evaluate optical fiber connectors with high precision and low cost. In the scientific
literature there are some proposals to solve the problem, Duffy et al. study in [6] automatized solutions to
evaluate and clean fiber optic connectors. Rehman and Mozaffar [7] use wavelets to evaluate the dirtiness of
connectors, while Filipenko in [8] use the interference method to analyze them. Commercially, companies
such as VIAVI solutions [9] and EXFO [10] manufacture and commercialize inspection and evaluation
equipment to verify the state of the fiber optic connector. Proposals in the scientific literature do not use
image processing and neural networks, which can improve the method’s precision. Likewise, the proposal by
Duffy et al. uses a programmable logic controller (PLC), which is not the best alternative for a portable
solution. On the other hand, commercial solutions imply making large investments which may not be viable
for small and medium-sized telecommunications networks companies.
In the face of all these problems and drawbacks, the present work proposes a low-cost portable
equipment capable of successfully evaluating and identifying the dirtiness of fiber optics connectors using an
optical microscope, a single-board computer, open-source tools, image processing algorithms and neural
networks. The device has a near-perfect assertiveness with evaluation results showing a Cohen’s Kappa
agreement index of 0.926. Furthermore, the equipment is well-suited to fieldwork requirements due to its
portability.
2. DESCRIPTION OF THE PROPOSED DEVICE
Figure 1 shows the block diagram of the proposed algorithm. The monocular of the microscope has
an adapter fixed to it, which safely holds the digital camera that takes the RGB image of the fiber optic
connector. The following paragraphs describe the image processing steps and algorithms employed to
achieve the desired results.
Figure 1. Proposed method flowchart
2.1. Optical microscope
It is composed of a manual device which allows visualization of the connector’s ferrule and the
contact area. Figure 2 shows the composing parts of a manual microscope. The eyepiece permits
visualization of the fiber through an optical zoom up to 400X. The focus control handles the manual focus to
improve image definition. The light source entrance helps the user note if the illumination LED is on or off.
2.2. Camera-microscope adapter and prototype device
The camera-microscope adapter consists of a piece, 3D printed in black PLA. This design is
exclusive to the proposed solution and helps in creating a dark enclosure to have a stable, constant
illumination when imaging takes place. Figure 3 shows a schematic of the adapter. The top piece is a
removable lid that covers the digital camera, while the rectangular hole is the entrance for an HDMI
connector to the camera. Figure 4 shows the camera installed in the rectangular area of the adapter see
Figures 4(a) and 4(b). Figure 4(c) shows the HMDI cable connected to the camera. Figure 4(d) shows a
picture of the bottom view, in which the camera lens is observed.
The adapter must comply with some minimum requirements to be compatible with the monocular in
the microscope. Per the manufacturer, the minimum focal distance is of 3.6 mm. Then, the designed adapter
will achieve a focal distance of 8 mm and be able to take images of the full fiber optic for subsequent
analysis. Figure 5 shows a picture of the adapter, the microscope, the ensemble and the labeled parts.
Likewise, the Figure 5(d) shows the complete box that contains the processor, the touch screen and the
program that will allow the evaluation. The complete prototype has characteristics:
 Easy to connect as seen in the Figure 5. The equipment allows connection through the HDMI cable and
the camera sends the image to the processor. All the data processed in the single board computer will be
displayed on the touch screen in real time.

Int J Elec & Comp Eng ISSN: 2088-8708 
Device to evaluate cleanliness of fiber optic connectors using image processing and... (Victor Fernandez)
3095
 The portability of the solution is based on its ease of transport: lightweight and easy to handle.
 The prototype uses an AC-DC power adapter. However, it has the option to deploy an internal battery.
Figure 2. Optical microscope description
Source: microscope quick start guide
Figure 3. Adapter schematic diagram
(a) (b) (c) (d)
Figure 4. Camera-microscope adapter: (a) Camera installed, (b) Adapter with the top piece, (c) HDMI cable
connected, (d) Bottom view
(a) (b)
(c) (d)
Figure 5. Adapter and prototype equipment: (a) Both elements separated, (b) Ensemble,
(c) Parts and components for image acquisition, (d) Prototype box

 ISSN: 2088-8708
3096
2.3. Image acquisition
A 25x24x9 mm, 8 Megapixel digital camera with a Sony IMX219 sensor images the fibers. To
allow the user real-time image visualization, the camera configuration was set to RGB (true color, 24 bits).
Because of the small sizes of the dirt objects, the chosen image resolution was of 3280x2464 pixels. This
translated into a spatial resolution of 0.0625 μm2
per pixel, an appropriate size to adequately evaluate the
objects. The default illumination intensity level was used since it is supplied by the microscope and kept
stable by the adapter’s enclosure. Figure 6(a) shows an example of an image acquired with the ensemble.
2.4. Image processing and descriptors
The following section details the steps taken to analyze the acquired image and obtain the three
selected image descriptors to train the neuronal network and determine the level of cleanliness or dirt in the
regions of interest.
2.4.1. Image cropping
The image acquired in step 2.3 is cropped, taking into account the fiber core and cladding areas
(area of interest for processing). The resulting image has a smaller size and its digitized primary components
of light are expressed as 𝐼𝑅(𝑥, 𝑦), 𝐼𝐺(𝑥, 𝑦) and 𝐼𝐵(𝑥, 𝑦). The values of this components are 8 bit-integers
([0,255]) according with true color format.
2.4.2. Grayscale conversion
In order to reduce computational load without diminishing the algorithm’s performance, the image
from Figure 6(b) is converted into an 8-bit grayscale image, according to (1) [11].
𝐼1(𝑥, 𝑦) = 0.299𝐼𝑅(𝑥, 𝑦) + 0.587𝐼𝐺(𝑥, 𝑦) + 0.114𝐼𝐵(𝑥, 𝑦) (1)
where 𝐼𝑅(𝑥, 𝑦), 𝐼𝐺(𝑥, 𝑦) and 𝐼𝐵(𝑥, 𝑦) are the primary components (red, green and blue) of the cropped image,
and 𝐼1(𝑥, 𝑦) is the grayscale image as shown in Figure 6(c).
(a) (b) (c)
Figure 6. Example results of the grayscale conversion: (a) Acquired image example, (b) Image with primary
components 𝐼𝑅(𝑥, 𝑦), 𝐼𝐺(𝑥, 𝑦) and 𝐼𝐵(𝑥, 𝑦), (c) Image 𝐼1(𝑥, 𝑦)
2.4.3. Histogram equalization
Image 𝐼1(𝑥, 𝑦) as shown in Figure 7(a) goes through a histogram equalization process [12]. The
resulting image is labeled 𝐼2(𝑥, 𝑦) Figure 7(b). Histogram equalization showed better results than
enhancement via histogram stretching, because it improved the later segmentation between the cladding
region and the rest of the image.
2.4.4. Gaussian filtering
Image 𝐼2(𝑥, 𝑦) is filtered with a 5x5 Gaussian mask [13] to eliminate undesired edges and contrast
regions. Such mask is defined by (2) [14].
ℎ1(𝑥, 𝑦) =
1
2𝜋𝜎2 𝑒
−(𝑥2+𝑦2)
2𝜎2
(2)
In this case we consider 𝜎 = 1 to obtain satisfactory results. The filtering process is expressed
through convolution between 𝐼2(𝑥, 𝑦) and ℎ1(𝑥, 𝑦) mask:
cladding cladding
cladding

3097
𝐼3(𝑥, 𝑦) = 𝐼2(𝑥, 𝑦) ∗ ℎ1(𝑥, 𝑦) (3)
Figure 7(c) illustrates the input image 𝐼2(𝑥, 𝑦) and the output image 𝐼3(𝑥, 𝑦).
(a) (b) (c)
Figure 7. Example results of the histogram equalization and Gaussian filtering: (a) Image 𝐼1(𝑥, 𝑦),
(b) Image 𝐼2(𝑥, 𝑦), (c) Image 𝐼3(𝑥, 𝑦)
2.4.5. Image thresholding
Image 𝐼3(𝑥, 𝑦) is thresholded in order to obtain the cladding segmentation mask, according to (4):
𝐼4(𝑥, 𝑦) = {
255, 𝐼3(𝑥, 𝑦) > 𝜇1
0, 𝑎𝑛𝑦 𝑜𝑡ℎ𝑒𝑟 𝑐𝑎𝑠𝑒
(4)
Any pixel with a higher value than the threshold 𝜇1 = 104 will be set to 255. This threshold was
chosen because it yielded the best results for different types of fibers and levels of cleanliness. Figure 8(a)
illustrates the output of the thresholded image, 𝐼4(𝑥, 𝑦).
2.4.6. Canny filtering
Canny filtering is applied to image 𝐼4(𝑥, 𝑦) to obtain the circle of the fiber cladding. The Canny
filtering involve: smoothed using a Gaussian kernel, obtaining of edge strength from Sobel filtering,
calculation and quantization of the edge direction, nonmaximum suppression and thresholding with
hysteresis [15]. In this case, two hysteresis thresholds (𝜇𝑀𝐴𝑋 = 110 and 𝜇𝑀𝐼𝑁 = 50) are set to obtain the
shapes and borders of the objects in the thresholding. The result of this filtering is 𝐼5(𝑥, 𝑦), shown in
Figure 8(b). The chosen hysteresis thresholds enabled reducing false contours for different types of fibers and
levels of cleanliness.
(a) (b)
Figure 8. Example results of the thresholding and filtering image: (a) Image 𝐼4(𝑥, 𝑦), (b) Image 𝐼5(𝑥, 𝑦)
2.4.7. Circle detection via Hough transform
The Hough transform for circle detection [16] is applied to image 𝐼5(𝑥, 𝑦). In this case a
3-dimensional accumulator was used, storing the 2 coordinates of the center and the radius. In the
implementation of Hough's transform, the parameter “dp” was set to 3, in order to resolve the search of a
large circle with low distortion levels. Gradski [17] defines this parameter as the resolution of the
accumulator image, which enables the creation of an accumulator with a lower resolution than the original
image. Essentially, a larger “dp” value translates into a smaller accumulator matrix. The chosen value
resulted in successfully detecting the circle which exactly belongs to the cladding, and greatly reduced the

 ISSN: 2088-8708
3098
false detection of circles. A larger “dp” value might have reduced the detected circle quantities to zero in
some cases. The transform requires setting a minimum distance between the centers of the circles. If this
distance is too small, false circles may be detected. If it is too large, circles may go undetected. It is known
that the circle to be detected has a radius of 250 pixels, so this value serves as the starting point for the
following computation.
If the value were chosen to be less than or equal to 250, most detected circles would be
superimposed over each other. If the value were chosen to be 500 (the circle diameter), some circles tangent
to the true circle would be detected. Figure 6 shows that the cladding is almost in the middle of the image,
and the region of interest is not large, so lost circles outside of this region would not be relevant for this
process. Thus, the distance between centers of circles is defined to be 1000, which reduced the quantity of
possible circles in the selection process. The results of the transform are the circle radius 𝑟𝑐 and the
coordinates of the circle center (𝑥𝑐, 𝑦𝑐). With these coordinates and the circle radius 𝑟𝑐, the detected circles
are drawn over the original image such that the regions of interest are defined. Figure 9 shows an example
where results correspond to 3 case: Clean fiber as shown in Figure 9(a), dirty fiber as shown in Figure 9(b)
and very dirty fiber as shown in Figure 9(c).
(a) (b) (c)
Figure 9. Example results of the transform: (a) Clean fiber, (b) Dirty fiber, (c) Very dirty fiber
2.4.8. Regions of interest segmentation
The segmentation of the regions of interest takes the circle coordinates and radius from the previous
step and generates a mask 𝑍1(𝑥, 𝑦) the same size of 𝐼1(𝑥, 𝑦), according to the (5).
𝑍1(𝑥, 𝑦) = {
1, (𝑥 − 𝑥𝑐)2
+ (𝑦 − 𝑦𝑐)2
≤ 𝑟𝑐
2
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
(5)
where (𝑥𝑐, 𝑦𝑐) corresponds to the center coordinates of the found circle and 𝑟c, the radius of the found circle.
Using the (6), the mask 𝑍1(𝑥, 𝑦) segments the region of interest in 𝐼1(𝑥, 𝑦):
𝐼6(𝑥, 𝑦) = 𝑍1(𝑥, 𝑦). 𝐼1(𝑥, 𝑦) (6)
where 𝐼6(𝑥, 𝑦) is the output image after masking. Figure 10(a) shows an example of 𝐼6(𝑥, 𝑦).
Since the fiber cleanliness must also be evaluated around the outer neighborhood of the cladding, the
outer ring was also segmented. Thus, mask 𝑍3(𝑥, 𝑦) was computed according to the (7).
𝑍3(𝑥, 𝑦) = 𝑍2(𝑥, 𝑦) − 𝑍1(𝑥, 𝑦) (7)
where 𝑍2 is defined according to (8).
𝑍2(𝑥, 𝑦) = {
1, (𝑥 − 𝑥𝑐)2
+ (𝑦 − 𝑦𝑐)2
≤ (2𝑟𝑐)2
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
(8)
Finally, the segmentation of the grayscale pixels in the cladding exterior ring is computed according to (9).
𝐼7(𝑥, 𝑦) = 𝑍3(𝑥, 𝑦). 𝐼1(𝑥, 𝑦) (9)
Figure 10(b) shows an example of 𝐼7(𝑥, 𝑦).

3099
(a) (b)
Figure 10. Example results of the segmentation: (a) Image 𝐼6(𝑥, 𝑦) (segmented cladding), (b) Image 𝐼7(𝑥, 𝑦)
(segmented outer ring)
2.4.9. Obtaining descriptors
These steps evaluate the segmented regions 𝐼6(𝑥, 𝑦) and 𝐼7(𝑥, 𝑦) to compute the first two
descriptors: cladding and ring variance. Furthermore, they also evaluate image 𝐼1(𝑥, 𝑦) to find the third
descriptor: quantity of contours in the image. The first step consists on calculating the histograms of 𝐼6(𝑥, 𝑦)
and 𝐼7(𝑥, 𝑦), without considering pixels with zero intensity (black pixels), such that histograms only contain
only gray pixels. Figures 11 and 12 show these histograms.
Figure 11. Segmented cladding histogram Figure 12. Segmented ring histogram
These histograms serve to calculate the variances [18] in each region (cladding or ring). These
variances are defined as σ𝐶𝑙𝑎𝑑𝑑𝑖𝑛𝑔
2
and σ𝑟𝑖𝑛𝑔
2
according to (10).
𝜎2
=
∑ (𝑋𝑖−𝑋
̅)2
𝑚
𝑖=1
𝑚−1
(10)
where 𝑋𝑖 represents a pixel value of a region, 𝑚 the number of pixels of a region and 𝑋
̅ the average value of a
region (cladding or ring).
The variances of each region are radiometric descriptors that allow to detect the dirt presence level.
In this case a region with high level of dirt also will have high variance value. The third descriptor is the
number of contours (𝑁𝑢𝑚𝐶𝑜𝑛𝑡) identified in the 𝐼1(𝑥, 𝑦) image. This descriptor was chosen because the
number of contours increases significantly when the dirt level is high. It helps to differentiate between
different levels of dirt or cleanliness.
To obtain this descriptor, the steps 2.4.2, 2.4.3, 2.4.4, 2.4.5 and 2.4.6 are applied to 𝐼1(𝑥, 𝑦). The
difference is in the hysteresis thresholds used in the Canny filtering step: 𝜇𝑀𝐴𝑋 = 200 and 𝜇𝑀𝐼𝑁 = 10. These
hysteresis values attempt to increase the threshold range such that more borders will be found in the whole
image, instead of only in the cladding region. This translates into searching for more borders and shapes that
could correspond to dirtiness. Figure 13 shows the output image 𝐼8(𝑥, 𝑦).
The next step consists of finding the quantity of contours in image 𝐼8(𝑥, 𝑦). Figure 14 shows a
flowchart to achieve this objective. Figure 14 shows that Suzuki’s algorithm [19] analyzes image 𝐼8(𝑥, 𝑦) to
then discard redundant points and obtain a list with the coordinates of all the relevant contours. Then, the
integer variable 𝑁𝑢𝑚𝐶𝑜𝑛𝑡 is calculated from this list’s length.

 ISSN: 2088-8708
3100
Figure 13. Image 𝐼8(𝑥, 𝑦) Figure 14. Flowchart for contour quantification
Figure 15 illustrate the basic steps of Suzuki’s algorithm. The algorithm scans the image from left to
right, labels edges as outer or hole edges and establishes the hierarchy between the discovered pixels. This
repeats for each image row, from top to bottom. The algorithm considers contours as continuous and smooth
curves. Finally, the steps achieve the following 3 descriptors: σ𝐶𝑙𝑎𝑑𝑑𝑖𝑛𝑔
2
, σ𝑟𝑖𝑛𝑔
2
and 𝑁𝑢𝑚𝐶𝑜𝑛𝑡.
2.5. Identification of the cleanliness level by neural networks
A trained model of neural networks was used to identify the cleanliness level of the fiber optic
connectors [20]. The model inputs consist on the three previously calculated descriptors, and the output is a
3-element array which indicates the cleanliness of the fiber optic connector. Figure 16 illustrates how the
descriptors are input into the neural network and the model outputs the binary array as described above.
[0,0,1]: CLEAN state [0,1,0]: DIRTY state [1,0,0]: VERY DIRTY state
Figure 15. Suzuki’s algorithm basic steps Figure 16. ANN diagram
The dataset consisted of 99 pictures of fiber optic connectors, evenly distributed into 3 groups of 33
photos for each possible output. The dataset was further divided into training and evaluation sets:
 Training set (70% of the total dataset)
 Evaluation set (30% of the total dataset)
The objective was to achieve a precision of at least 90%. To achieve this, the network was set
according to the following characteristics:
 Network size: Sánchez [21] mentions that the importance of the hidden layer size to the neural network
model fit has been stated many times in the scientific literature, but no conclusive results have been
demonstrated. Nonetheless, there are plenty guiding criteria to start building said models. Some of the
criteria considered in this work are described by Del Carpio [22]:
a) Quantity of hidden layers: more layers add complexity to the network and make the training slower. If
data are linearly separable, hidden layers are not necessary, but in this case, data are not linearly
separable.

3101
b) Quantity of neurons in each hidden layer: these neurons are related to the input variables, such that
this quantity must not be larger than twice the input size. If this is not enough to achieve acceptable
results, more neurons are added to the output layer.
Another criterion is directly related to the final adjustment error and desired precision. These criteria
guided the model building and simulation, which was done with TensorFlow (a free and open-source
software library for dataflow and differentiable programming developed by Google). This simulation
helped in defining the final network size: 3 hidden layers with 12, 4 and 3 neurons, which improve the
loss during training and testing. Scikit-learn [23] explains each neuron in the hidden layer transforms the
values from the previous layer with a weighted linear summation, followed by an activation function. The
output layer receives the values from the last hidden layer and transforms them into output values.
 Alpha=0.06, for overfitting correction. This value contributes to keeping the precision above 90% by
avoiding overfitting. Scikit-learn use this parameter as a regulator which penalizes overfitting by
restricting the weight magnitude.
 ReLU (rectified linear unit) activation function for the hidden layer.
 Training algorithm ‘adam’. Scikit-learn defines ‘adam’ as a stochastic gradient-based optimizer, which
can automatically adjust the amount to update parameters based on adaptive estimates of lower-order
moments
 Maximum iterations=11000
With these hyperparameters, the model achieved a training precision of 97% and a testing precision
of 92%. Figure 17 presents the training and testing algorithm flowchart. The first step is building the network
model, defining neurons, layers, and iterations. Next, training takes place, where loss is analyzed at the end:
if it is lower than 0.3, the model is tested, otherwise, training resumes. The 0.3 loss ensured that by the end of
the training, precision was of at least 90%. Finally, if precision is higher than 91%, the model is saved,
otherwise training resumes.
Figure 17. Neural network training flowchart

 ISSN: 2088-8708
3102
3. RESULTS AND DISCUSSION
First of all, as result, the device reaches the denomination of a low-cost device. Also, the novel
proposal device permits a high-quality image acquisition in order to solute the issues registered in previous
solutions and proposals [7]. In the other hand, to test the accuracy of the device, two human experts evaluated
the cleanliness/dirtiness of 27 SC/UPC fiber optic connectors through survey-type formats. This evaluation
was performed by visual inspection of images acquired with optical microscope. However, for validation,
only cases where both experts agreed in their evaluation were considered.
Furthermore, the images of the same connectors were evaluated by the proposed algorithm (output
of the neural network algorithm). Using the Cohen’s Kappa Index [24] which is served to measure the
agreement level according between two or more observers according to the values in Table 1 [25, 26], the
agreement level according between experts and the device proposed are defined by the magnitude of the
index K [27].
Table 1. Kappa Cohen index interpretation
Kappa Cohen Index Interpretation Kappa index Range
Poor coincidence K ≤0.20
Fair agreement 0.20 <K ≤0.40
Moderate agreement 0.40 <K ≤0.60
Satisfactory agreement 0.60 <K ≤0.80
Near-perfect agreement K >0.80
Figure 18 shows the flowchart to compute K. A first agreement evaluation of the 27 connectors
resulted in 6 of them where human experts had differing results, due to subjective perception. After
eliminating those 6 conflicting elements, the evaluation set was reduced to 21 connectors (i.e., the device was
evaluated 21 times), which results are indicated in Table 2. Next, Table 3 shows the confusion matrix
between expert evaluation and the proposed algorithm’s evaluation. These results are shown as proportions in
Table 4. The metric K [24] is computed according to (11).
𝐾 =
𝑝𝑜−𝑝𝑐
1−𝑝𝑐
(11)
where:
𝑝𝑜 = proportion of evaluated samples where judges agree
𝑝𝑐 = proportion of evaluated samples expected at random
Image adquisition
Run detection
algorithm
Save results
Experts Survey
Expert 1
Expert 2
Save results
Eliminate
subjectivity
Calculate K index
Figure 18. Procedure to compute K
Calculating (11) for the confusion matrix at Table 4 yields:
𝐾 =
[0.190+0.429+0.333]−[(0.190×0.238)+(0.429×0.429)+(0.381×0.333)]
1−[(0.190×0.238)+(0.429×0.429)+(0.381×0.333)]
= 0.926

3103
The agreement results from Table 5 is interpreted according to Table 1, such that:
𝐾 > 0.8 → 𝑁𝑒𝑎𝑟 𝑝𝑒𝑟𝑓𝑒𝑐𝑡 𝑐𝑜𝑛𝑐𝑜𝑟𝑑𝑎𝑛𝑐𝑒
Next, the index is calculated for each possible result: clean, dirty, and very dirty. Results are shown
in Table 6. Note that the proposed method achieves accurate results when compared to human experts.
Table 2. Results to avoid subjectivity
# Image ID Connector Expert 1 Expert 2 Experts’ coincidence Prototype Result
1 Connector 95 Dirty Dirty Yes Dirty
3 Connector 63 Very Dirty Very Dirty Yes Very Dirty
4 Connector 2 Clean Clean Yes Clean
18 Connector 4 Dirty Dirty Yes Clean
Table 3. Confusion matrix
Experts
Clean
Very
Dirty
Dirty Total
Prototype
results
Clean 4 0 1 5
Very Dirty 0 9 0 9
Dirty 0 0 7 7
Total 4 9 8 21
Table 4. Confusion matrix of proportions
Experts
Clean
Very
Dirty
Dirty Total
Prototype
results
Clean 0.190 0.000 0.048 0.238
Very Dirty 0.000 0.429 0.000 0.429
Dirty 0.000 0.000 0.333 0.333
Total 0.190 0.429 0.381 1.000
Table 5. Index of agreement result
Index Value
Agreement index(K) 0.926
# Valid cases 21
Table 6. Individual agreement index result
Index Value
Agreement
index(K)
Clean 0.859
Dirty 0.529
Very Dirty 1.000
# Valid cases 21
4. CONCLUSION
The conclusions and final remarks are a low-cost device, capable of evaluating the cleanliness of
fiber optic connectors with similar results than human experts, has been successfully developed. The
microscope adapter achieved stable illumination and safely holds the digital camera. The image processing
algorithm computes adequate variables, which reflects on the Cohen’s Kappa index values being close to
perfect agreement. Nonetheless, when analyzing results individually, the index for dirty connectors is 0.529,
which is classified as Moderate Agreement. Although this can be considered an acceptable result, it will be
necessary to enhance training such that the agreement value can be higher for this particular result.

 ISSN: 2088-8708
3104
The algorithm takes 3 seconds to output the image classification. The digital camera microscope
adapter is practical and innovative. The proposed system can be implemented in companies and organizations
that already have the optical microscope, further reducing the system cost and increasing the market
acceptance. The algorithm might fail when analyzing fiber optic connectors with a broken or damaged
ferrule. Since this case is not considered in the algorithm design, the result might not agree with the
connector being clean, dirty, or very dirty. As opportunities for improvement, there is the development of a
case that can withstand the inclemency of field work and complement the algorithm developed with instances
that allow us to distinguish what type of dirt is present and perhaps carry out an automatic cleaning, using
electronic tools.
ACKNOWLEDGEMENTS
The authors would like to thanks “Universidad Peruana de Ciencias Aplicadas” for technical and
economic support.
REFERENCES
[1] Y. Roriguez, “Fibra Óptica,” Santa Fe: El Cid Editor, 2009.
[2] E. H. Perez, “Introducción a las Telecomunicaciones Modernas,” Mexico DF: Limusa, 2004.
[3] Chougule, Sharada V., and Mahesh S. Chavan, “Investigation of Connector Performance and Damage to Optical
Fiber Connectors with Dust During High-Power Transmission,” in Journal of Lightwave Technology, vol. 36,
no. 17, pp. 3586–3593, Sep. 2018.
[4] J. Simatupang, “Theoretical Analysis of Backreflections in Bidirectional Wavelength Division Multiplexing-
Passive Optical Networks,” Bulletin of Electrical Engineering and Informatics (BEEI), vol. 2, no. 1, pp. 45–52,
Mar. 2013.
[5] M. Brown, “Achieving IEC Standars Compliance for Fiber Optic Connector Quality,” VIAVI, [Online]. Available:
https://www.viavisolutions.com/en-us/literature/achieving-iec-standards-compliance-fiber-optic-connector-quality-
white-paper-en.pdf
[6] J. Duffy, D. Hashimoto, J. Sloan, and G. Bellegarde, “Development of an Automated Cleaning System for Multi-
ferrule Fiber Optic Connectors,” in Conference on Optical Fiber Communication/National Fiber Optic Engineers
Conference, San Diego, CA, 2008, pp. 1–5, doi: 10.1109/OFC.2008.4528037.
[7] A. Rehman and W. B. Mozaffar, “Optical Fiber Connector Surface Defect Detection Using Wavelets,” in 2007
International Conference on Information and Emerging Technologies, Karachi, 2007, pp. 1–6.
[8] A. I. Filipenko, “Fiber connector end face analysis by interference method,” in Proceedings of LFNM'2000. 2nd
International Workshop on Laser and Fiber-Optical Networks Modeling (Cat. No.00EX419), Kharkiv, Ukraine,
2000, pp. 74–76, doi: 10.1109/LFNM.2000.854045.
[9] “Fiber Inspection & Cleaning,” VIAVI, 2020. [Online]. Available: https://www.viavisolutions.com/en-us/product-
category/enterprise-networks/test-certification/fiber-inspection-cleaning
[10] “Fiber Inspection,” EXFO, 2020. [Online]. Available: https://www.exfo.com/es/productos/pruebas-de-redes-de-
campo/inspeccion-de-fibra/
[11] UIT-R, “Parámetros de codificación de televisión digital para estudios con formatos de imagen normal 4:3 y de
pantalla ancha 16:9,” Recomendación UIT-R BT.601-7, 2011. [Online]. Available:
https://www.itu.int/dms_pubrec/itu-r/rec/bt/R-REC-BT.601-7-201103-I!!PDF-S.pdf
[12] C. Platero, “Apuntes de Visión Artificial,” Universidad Politécnica de Madrid, 2007, Art. no. 120.
[13] J. García, “Estudio de Filtros Óptimos para la eliminación de ruido en imágenes digitales usando algoritmos de
multirresolución no lineales,” Repositorio Digital de la Universidad Politécnica de Cartagena, pp. 3, 2009.
[14] R. L. Cañar and R. Escandón, “Implementation of a prototype for capture and digital processing of thermal images
acquired from a UAV,” Enfoque UTE, vol. 9, no. 1, pp. 1–11, Mar. 2018.
[15] E. A. Sekehravani, E. Babulak, and M. Masoodi, “Implementing canny Edge detection algorithm for noisy image,”
in Bulletin of Electrical Engineering and Informatics, vol. 9, no. 4, pp. 1404–1410, Aug. 2020.
[16] H. Ohmaid et al., “Iris segmentation using a new unsupervised neural approach” in IAES International Journal of
Artificial Intelligence (IJ-AI), vol. 9, no. 1, pp. 58–64, Mar. 2020.
[17] G. Bradski and A. Kaehler, “Learning OpenCV,” 2008. [Online]. Available:
https://www.bogotobogo.com/cplusplus/files/OReilly%20Learning%20OpenCV.pdf
[18] C. Salvatierra, “Fundamentos de procesamiento digital de imágenes de satélite,” 2004. [Online]. Available:
http://ffyl1.uncu.edu.ar/IMG/pdf/Notas_clase_Procesamiento_digital.pdf
[19] S. Suzuki and K. Abe, “Topological Structural Analysis of Digitized Binary Images by Bordes Following,” CVGIP,
vol. 30, no. 1, pp. 32–46, 1985.
[20] S. Haykin, “Neural Networks: A Comprehensive Fundation,” 2nd edition USA, Prentice Hall, 1999.
[21] P. Sánchez and J. Velásquez, “El rol del algoritmo de entrenamiento en la selección de modelos de redes
neuronales,” Revista U.D.C.A Actualidad & Divulgación Científica, vol. 14, no. 1, pp. 149–156, 2011.
[22] C. Del Carpio, “Método de reconocimiento de patrones basado en procesamiento digital de imágenes y redes
neuronales orientado al registro automático de evaluaciones académicas,” M.S. thesis, Universidad Nacional de
Ingeniería, Facultad de Ingeniería Eléctrica y Electrónica, Lima, Perú, 2015.

3105
[23] Scikit-Learn, “Varying regularization in Multi-layer Perceptron,” 2019. [Online]. Available: https://scikit-
learn.org/stable/auto_examples/neural_networks/plot_mlp_alpha.html
[24] J. Cohen, “A Coefficient of Agreement for Nominal Scales,” Educational and Psychological Measurement,
vol. 20, no. 1, pp. 37–46, Apr. 1960.
[25] S. M. Vieira, U. Kaymak, and J. M. C. Sousa, “Cohen's kappa coefficient as a performance measure for feature
selection,” in International Conference on Fuzzy Systems, Barcelona, 2010, pp. 1–8.
[26] J. Lazar, J. Feng and H. Hochheiser, “Research Methods in Human-Computer Interaction,” 2nd Edition, Morgan
Kaufmann, 2017.
[27] J. R. Landis and G. G. Koch, “The measurement of observer agreement for categorical data,” Biometrics, vol. 33,
no. 1, pp. 159–174, Mar. 1977.
BIOGRAPHIES OF AUTHORS
Victor Fernandez, Electronic Engineer of the Peruvian University of Applied Sciences (UPC),
he obtained a Diploma Project Management and Agile Methodologies from Telefonica
Foundation. He has been part of telecommunications project for more than 4 years. He currently
works on development of new technologies and IT applications.
Javier Gustavo Chavez, Electronic Engineer of the Peruvian University of Applied Sciences
(UPC), he obtained a Diploma on Telecommunication Management in Graduate School of the
Peruvian University of Applied Sciences (EPG-UPC). Optical DWDM Network Specialist. He
has been part of optical fiber projects for 7 years. He currently works on optimizing and
supporting the National Optical Fiber Network of Perú.
Guillermo Kemper received the B.S. degree in Electrical Engineering in 1994 from the Antenor
Orrego Private University (UPAO), Trujillo, Peru. In years 1996 and 2001 he obtained a Master
and a PhD degree in Electrical and Communications Engineering respectively. Both degrees
from the State University of Campinas (UNICAMP), Brazil. He has taken part for more than 3
years in the research agreement between the CPqD-Telebras (Brazilian Telecommunications
Company) and UNICAMP, developing and designing audio and video coders. Currently, he is
working as a research professor at the Electrical Engineering Program of the Peruvian University
of Applied Sciences (UPC) in Lima, Peru. In addition, he is a researcher at the National Institute
of Research and Training in Telecommunications (INICTEL). His research interests include
voice, audio, images and video processing, as well as digital communications and digital
television.

Device to evaluate cleanliness of fiber optic connectors using image processing and neural networks

More Related Content

What's hot (20)

Similar to Device to evaluate cleanliness of fiber optic connectors using image processing and neural networks (20)

More from IJECEIAES (20)

Recently uploaded (20)

Device to evaluate cleanliness of fiber optic connectors using image processing and neural networks