Coffee Classifier Machine PRESENTATION.pdf

A coffee classifier machine is designed to sort and classify coffee beans based on various attributes such as size, shape, color, and quality.

1. Researcher’s:
DIANA C. MANOY | KYREL JOY P. PABLO
ACCEPATABILITY OF THE
DEVELOPED COFFEE BEAN
CLASSIFIER MACHINE
Adviser:
CELIA ROSE J. NOTA

2. OVERVIEW
Introduction
01
Literature Review
02
Results and Analysis
04
Discussion
05
Research
Methodology
03
Conclusion
06

3. This innovative approach has the potential to significantly simplify the traditionally
time-consuming operation of bean classification. It was created with the goal of
changing the sorting and grading of coffee beans. According to Riedel (2019) the
essential topic of this inquiry is whether, in the current environment of coffee
production, such a technique is not only technically feasible but also commercially
and socially Agree. To fully resolve this challenge, this thesis will conduct a holistic
investigation that considers socio-cultural ramifications, economic viability, and
technological utility. We intend to present a comprehensive assessment of the Coffee
Bean Classifier Machine through extensive testing, practical case studies, and
stakeholder interviews, highlighting its potential to 1 transform industry practices while
remaining consistent with the principles and customs that guide coffee production.
The researchers aim to bridge the existing gap between technological
advancements and their efficient integration into traditional farming methods. An
artificial image-processing agent that can be able to look at characteristics of the
coffee beans for example its color and size. Furthermore, the successful integration of
a coffee bean classifier would benefit the industry greatly in terms of efficient
timescale, cost of labor and standard of product output. Automation of the beam
classification process would reduce the effects of human related errors, making sure
that only the best quality of beans is available in the market for sale. We can argue
that the integration of technology and traditional farming practices will yield results
that will save the future of coffee production.
INTRODUCTION

4. OBJECTIVES OF THE STUDY
General Objectives
This study generally aims to develop a coffee bean
classifier machine that will identify the defected and accepted
coffee bean.
Specific Objectives
Specifically, this study aims to:
1.1 Create a diverse dataset of defected coffee bean images
for training the classifier.
1.2 Classify coffee beans into defective beans using a wide
camera.
1.3 Automatically sort the defected beans using the solenoid
and separate the classified coffee beans respectively.
1.4Evaluate the machine device in terms of:
4.1 . Functionality
4.2 . Accuracy, and
4.3. Acceptability

5. Related Literature
●YOLOv8 on a Custom Dataset – Roboflow: Pre-Training
●Google Colab (Google Colaboratory): Final Training
●Python Version 3.12
●Image Processing
●Defects Coffee Beans
●Accepted Coffee Beans
●Amd A10 Radeon 7 Processor
●
Related Studies
●Classification of Defects in Robusta Green Coffee Beans Using YOLO
●Coffee Beans Using Image Analysis Techniques
●Design and Development of a Coffee Bean Selector Using the Yolo Algorithm
●Design and Performance Test of the Coffee Bean Classifier
●Development and Testing of Green Coffee Bean Quality Sorter using Image Processing and Artificial Neural
Network
●Grading of Green Coffee Beans for Specialty Coffee using Image Processing Techniques

6. Related Studies
●Normal and Peaberry Coffee Beans Classification from Green Coffee Bean Images Using
Convolutional Neural Networks and Support Vector Machine
●Performance Evaluation of Coffee Bean Binary Classification Through Deep Learning
Techniques
●Pre-processing of Coffee Beans
●Quality Assessment of Coffee Beans Through Computer Vision And Machine Learning
Algorithms
●Quality Sorting of Green Coffee Beans From Wet Processing by Using The Principle Of
Machine Learning

8. Perspective Plan

9. The research design for the thesis titled "Acceptability of
the Developed Coffee Bean Classifier Machine" will employ
quantitative methods. To evaluate and acquirea information,
the researchers used evaluation sheets to the respondents
of the study. The evaluation sheets mostly included
questions that determine whether the device achieves its
functionality and acceptability of the machine, in terms with
the accuracy in the objectives the researchers were the one
who will assess and provide accurate evaluation together
with their statistician who advise and give them knowledge.

10. The designed acceptability of the developed coffee
bean classifier machine among coffee farmers and
processors will be evaluated through the objective
gathering of data made possible by the quantitative
research approach. Utilizing a methodical
methodology, the research seeks to produce reliable
and accurate information that may guide further
advancements and modifications to the machine's
performance and design.

11. The researchers collected the evaluation forms from
the participants and tally the results after evaluating
the Acceptability of The Developed Coffee Bean
Classifier Machine. Based on the respondents'
responses, the data has been tabulated. The device
evaluation results in terms of functionality,
acceptability and accuracy were statistically
processed using frequency counts and percentage
distribution.

12. RESULT AND
DISSCUSSION
The researchers created a diverse dataset of coffee
bean images for training a classifier. The process
involved manually adjusting settings and scaling the
images to lower resolution. The wide camera enabled
effective classification into accepted and defect
categories. A solenoid for automatic sorting further
streamlined the classification process. The device
demonstrated 100% precision in functionality and
acceptability, with an accuracy rate of 86%. The
meticulous process of data collection, image
preparation, and technological application resulted in
a reliable and efficient coffee bean classifier machine,
meeting industry standards. The process of creating
a diverse dataset and training the classifier was time-
consuming but ultimately successful.

13. The figure shows the actual sorting
of coffee beans using a solenoid.
The solenoid quickly engages to
remove defective beans from the
mainstream when the camera
detects their location, guaranteeing
that only defective beans go on with
the process.
This figure shows the fifty
accepted coffee beans that used
during the classification of the
device. In the first trial, the
device detected only fifty
accepted coffee beans; thirty-
seven (37) entered the accepted
chamber, and 13 entered the
defective chamber.
The figures show the actual image of the
coffee bean classification that was
captured using a wide camera. The
camera recognized the coffee beans
that had been accepted and rejected.
After the coffee beans drop on a
straight conveyor belt, a camera
immediately detects them and shows
their location.

14. The output for the coffee beans that the camera detected to be
acceptable and defective is shown in Figure 9. In accepted coffee
beans one hundred (100) beans are used, and one hundred (100)
defective beans are used. During the training, the class named for
accepted is CB1 which is Coffee Beans 1 and for defective is CB2
which is Coffee Bean 2. The camera detects the CB1 and CB2 but
when it comes to sorting the solenoid sort only the defective coffee
beans. The result of the device in the accepted coffee beans
chamber out of 100 beans 13 defective coffee beans entered and
for defective out of 100 beans 18 accepted coffee beans entered in
a chamber of defective.
Figure 8 shows the output of the second
trial for defective coffee beans. In the
second trial, the device detected only
fifty defective coffee beans; thirty (30)
entered the defect chamber, and twenty
(20) entered the accepted chamber.

15. The results of evaluation of the device in terms of functionality.
The result showed that the device can classify the coffee beans
into accepted category obtained 15 YES answers out of 15 having a
percentage of 100 and was interpreted as functional.
The same table shown that the device was having 15 YES answers
out of 15 having a percentage of 100 in the question about the
device can classify the coffee beans into defective category.
Also, the third question "The device can sort the coffee beans into
accepted category" got 13 YES answered out of 15 having a
percentage of 86.37.
Next, 13 coffee processors and farmers answered YES out of 15 in
the question the device can sort the coffee beans into defective
category having a total percentage of 86.37.
Lastly, the same table shown that the device obtained 15 YES
answers out of 15 having a percentage of 100 in terms of the
device can easily operate by the user.
The overall total percentage of YES answers was 94.55 it implied
that the device was functional as evaluated.
The results of evaluation of the device in terms of functionality. The
result showed that the device can classify the coffee beans into
accepted category obtained 15 YES answers out of 15 having a
percentage of 100 and was interpreted as functional. The same table
shown that the device was having 15 YES answers out of 15 having a
percentage of 100 in the question about the device can classify the
coffee beans into defective category. Also, the third question "The
device can sort the coffee beans into accepted category" got 13 YES
answered out of 15 having a percentage of 86.37. Next, 13 coffee
processors and farmers answered YES out of 15 in the question the
device can sort the coffee beans into defective category having a total
percentage of 86.37. Lastly, the same table shown that the device
obtained 15 YES answers out of 15 having a percentage of 100 in terms
of the device can easily operate by the user. The overall total
percentage of YES answers was 94.55 it implied that the device was
functional as evaluated.

16. The results of the device's accuracy examination were displayed
in Table 14. One hundred beans were accepted as part of our trial,
while another hundred were found to be defective. Out of 100
accepted beans, 13 defective beans have been entered in
accepted chamber, and 18 accepted beans have been entered in
defective chamber out of 100. The device's accuracy rate for
coffee beans that are acceptable beans is 87%, while the device's
accuracy percentage for defective beans is 82%.

17. The summary of the evaluation of the Coffee Beans Classifier Machine.
The total result in terms of functionality of the device obtained 94.55%
which was interpreted as functional. It implied that 73 out of 75 answered
YES in the evaluation. The same table showed the result in terms of
acceptability of the device acquired an interpretation of acceptable with
a percentage of 100 for YES. It indicated that 75 out of 75 answered YES
as the Image processing was very crucial it couldn’t easily detect the
beans depending on capturing the beans and annotating. Regarding the
device's accuracy, the researchers used an actual confusion matrix for
evaluation, the device obtained 84.5% which was interpreted as accurate.
This indicated that the study objectives were met and the device was
functional, acceptable, and accurate as evaluated.

18. •The Coffee Bean Classifier Machine is a device designed to
efficiently classify and sort coffee beans, ensuring consistency and
quality in coffee production. The machine uses a wide camera to
provide high-definition images of coffee beans for classification, and
a solenoid to automatically sort beans based on classification results.
The machine's functionality and acceptability were evaluated,
resulting in a 100% precise functional and acceptable device.
However, the construction of the machine faced challenges, including
incompatible camera modules, a complex setup process, and
inefficiencies. The machine's initial design was not compact and did
not fit well with the overall structure, leading to redesigns. The
stepper motor was replaced with a solenoid for faster and more
precise sorting capabilities. The feeder/funnel design was redesigned
to prevent bean stacking and the Raspberry Pi processor was
switched to a desktop computer for faster processing and more
accurate results. Proper training and adjustment of camera settings
were necessary for accurate classification.

19. •Based to the opinions of the respondents and the findings of this
research the acceptability of the developed coffee bean classifier
machine is effective as most of the objectives have been met.
Coffee bean sorting and classification using the coffee bean
classifier machine has shown encouraging results in terms of
functionality, acceptability, and accuracy. With the use of a
solenoid for sorting, a desktop computer for processing, and
thorough design, the system successfully classifies coffee beans
into acceptable and defective categories. The success of the
machine depends on selecting suitable parts and taking design
issues seriously. Difficulties with the initial camera module and
stepper motor were resolved by thorough examination and
modifications. The coffee bean classification machine can
guarantee constant coffee production quality, save labor
expenses, while improving accuracy significantly. More
improvements can result with more recommendation and refining.

20. Based on the results of the study and conclusion arrived at, and
with the goal to perfect the coffee bean classification machine, we
would recommend to those who want to execute out our thesis
project:

1. Used high specs of processors to avoid delay during classify the
coffee beans.
2. Used a High-End Camera can Increases classification accuracy
and speed by providing clear, detailed high-resolution pictures.
3. Enhance the sorting and preparation of coffee beans it's
advisable to use compressor by creating air jets for accurate and
efficient sorting.
4. Future researchers design the feeder to be adjustable, allowing
for different bean sizes and weights. This flexibility can ensure that
the feeder can handle various types of beans without jamming or
misalignment.

21. THANK YOU

22. RESOURCE PAGE