Seminar Report

On

“Iot Base Smart Glasses For Blind People”

By

Shruti Kolapkar

52

Under the Guidance

Of

Ms. Preeti Ramtekkar

Sinhgad Technical Education Society's

RMD SINHGAD SCHOOL OF ENGINEERING, WARJE,

PUNE-58

Department of Computer Engineering

A.Y 2024-25

Affiliated to
Savitribai Phule Pune University, Pune
 RMD SINHGAD SCHOOL OF ENGINEERING,
WARJE, PUNE-58
Department of Computer Engineering

CERTIFICATE

This is to certify that Shruti Kolapkar from Third Year Computer Engineering has
successfully completed his seminar work titled “Iot Base Smart Glasses For Blind
People” at RMD Sinhgad School of Engineering, Warje, Pune in the partial
fulfillment of the Bachelor’s Degree in Engineering.

Ms. Preeti Ramtekkar Prof. Vina M. Lomte

Guide, Head,

Department of Computer Engineering Department of Computer Engineering

Dr. V. V. Dixit

Principal,

RMD Sinhgad School of Engineering, Pune – 58

Place: Pune

Date: 08/10/2024

I
 ACKNOWLEDGEMENT

It is my pleasure to acknowledge a sense of gratitude to all those who helped me in

making this seminar. I am highly indebted to Ms. Preeti Ramtekkar for her guidance
and constant supervision as well as for providing necessary information regarding the
seminar & also for her support.

I would like to express my gratitude towards Prof. Vina M. Lomte (Head - Department
of Computer Engineering) for her kind co-operation and encouragement which helped
me in preparation of this seminar.

Finally, I wish to thank and appreciate all my teachers and friends for their constructive
comments, suggestions & guidance and all those directly or indirectly helped me in
completing this seminar.

Shruti Kolapkar

II
 ABSTRACT

Even with the aid of canes or sticks, blind people have a difficult time walking on their
own. They must always be accompanied to regions where they must go, whether by
trained dogs or people; nonetheless,there may be times when there is no supervisor
present and they are stranded due to mobility limits. This difficulty is addressed by this
smart glass. Its approach is based on object identification and depth estimation. This
approach examines the environment for potential threats along a walk path, including the
distance to the object and the type of object, and notifies the user of the danger, allowing
the user to move around freely without continual monitoring. After the model detects the
object, an algorithm tracks the distance to the selected user and provides the user with the
necessary information required to process his environment. The glass's camera collects
data from the user's surroundings and delivers it to the model, which scans it and gives
input on probable directions. This smart glass is a tremendous help to blind individuals
who are always frustrated by the fact that, due to their condition, they must be constantly
overseen by others, relieving theiranxieties a little by providing new solutions to their
problems.

III
 TABLE OF CONTENTS
CERTIFICATE..................................................................................................................I
ACKNOWLEDGEMENT................................................................................................II
ABSTRACT.....................................................................................................................III
Introduction........................................................................................................................1
Objectives:..................................................................................................................................1
Problem Statement:....................................................................................................................1
Literature Survey...............................................................................................................2
Architecture of the System................................................................................................3
Technology Used in the System.........................................................................................4
Analytical Study.................................................................................................................6
Experimental Implementation..........................................................................................7
Primary Function..........................................................................................................................7
System Integration and Testing........................................................................................7
Case Study...........................................................................................................................9
Project Overview.........................................................................................................................9
Implementation Challenges...................................................................................................10
Conclusion.........................................................................................................................11
Future Scope.....................................................................................................................12
Bibliography.....................................................................................................................13
Plagiarism Check Report……………………………………………………………….14
 Optimizing MapReduce for the Modern World
1

Introduction
Objectives:
The aim of this project is to design and construct a Smart glass for blind

people with the following objectives:

• To design a Smart glass using Raspberry pi

• To implement the Smart glass using Raspberry pi

• To test and evaluate the performance of the glass

Problem Statement:
One of the most misunderstood disabilities is blindness. The general

populace has preconceptions about the blind that they feel are accurate.

Some people feel that the blind can't work or live a regular life on their own. This
contrasting cultural perspective and reality of a blind person's life has been depicted in
numerous literatures[4]. There is some truth in the public's perception, as the blind do
experience difficulties.

Some of these challenges include being unable to move, requiring assistance from others,
and so on. It can be difficult to perceive an obstacle or detect it in time. Most of the time,
individuals must become acquainted with their surroundings in the hopes that nothing will
change.

This is a possible danger since it could result in serious injuries or even death. Being able
to assist this blind person in understanding and identifying potential hazards can help
them avoid harm. Nothing is more essential than preserving a life; in this situation, we are
assisting them in living their own life.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
2

Literature Survey

Proposes IoT integration for High costs associated with IoT

IoT-Enabled Smart real-time object detection, components; requires
2020
Glasses for Visually providing audio feedback to continuous power source for
Impaired (IEEE Xplore) users. optimal use.

Real-Time Obstacle Limited range in obstacle

Demonstrates high accuracy in
Detection for Visually detection; requires
2021 detecting obstacles within close
Impaired Using IoT improvement in processing
proximity, enhancing mobility.
(Science Direct) speed.

Provides insights into

Wearable Devices for Does not address specific
navigation assistance devices,
Navigation Assistance 2019 requirements for visually
focusing on wearable IoT
(Harvard Health Blog) impaired; general overview.
technology.

IoT-Based Assistive Covers various IoT-based

Generalized approach; lacks
Technologies for Blind technologies for assistance,
2022 details on specific technology
Individuals (National including speech recognition
or practical applications.
Institutes of Health) and navigation.

Integrates machine learning

Machine Learning and IoT Data privacy concerns;
with IoT to improve object
in Assistive Devices for 2023 potential latency issues with
recognition accuracy in real-
Blind (Science Direct) real-time processing.
time.

Department of Computer Engineering, RMDSSOE, Pune

Optimizing MapReduce for the Modern World
3

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
4

Architecture of the System

Figure 1

Figure 2

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
5

Technology Used in the System

3.2 Hardware design

The various sub-components of the hardware system that have been

used to make the smart glasses system are described below

3.2.2 Microcontroller / Microprocessor:

 Description: Acts as the central processing unit of the glasses, handling data from
various sensors and peripherals.
 Examples: Arduino, Raspberry Pi, ESP32, STM32.

3.2.3 Ultrasonic Sensors:

 Description: Used for real-time object detection, facial recognition, and reading
text. These cameras capture the environment and feed data to the processing unit.
 Examples: Optical sensors like CMOS or CCD cameras.

3.2.4 GPS Module:

 Description: Enables outdoor navigation by providing real-time location data.

 Examples: NEO-6M, u-blox GPS.

3.2.5 Bluetooth / Wi-Fi Module:

 Description: For wireless connectivity, enabling communication with

smartphones or cloud platforms.
 Examples: HC-05 Bluetooth Module, ESP8266 Wi-Fi module

3.2.6 Battery:

 Description: Powers the entire system, ensuring portability.

 Examples: Lithium-ion battery.

3.2.7 Frame and Mounting Structure:

 Description: A lightweight, wearable frame that houses all the components and
makes it comfortable for long-term use.
 Examples: Customized eyewear frame with embedded sensors and electronics.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
6

3.3 Software design

Computer Vision Algorithms:

 Description: The core of image and object detection, used for recognizing
objects, text, or even faces. It can be used for obstacle detection or reading signs.
 Examples: OpenCV, TensorFlow (for deep learning models).

Object Detection / Recognition Software:

 Description: Uses machine learning models for detecting and classifying objects
around the user. Can be trained to recognize common objects like cars, people,
stairs, etc.
 Examples: MobileNet.

Navigation Software:

 Description: Provides the user with walking directions by integrating GPS data
and offering audio guidance for safe navigation.
 Examples: Google Maps API, OpenStreetMap.

Data Communication Protocols:

 Description: Facilitates communication between the glasses and external devices,

such as smartphones, for cloud processing or updates.
 Examples: MQTT, HTTP, WebSocket.

Feedback Mechanisms:

 Description: Software routines that interpret sensor data and provide feedback to
the user, either via vibration, sound, or visual indicators.
 Examples: Customized vibration intensity control algorithms.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
7

Analytical Study

This analytical study focuses on the methodology and key aspects related to the
development, implementation, and evaluation of IoT-based smart glasses designed for
visually impaired individuals.

Methodology

 Design & Development:

 Hardware: Integrate sensors (cameras, ultrasonic sensors) and IoT connectivity

for data gathering and real-time processing.
 Software: Develop AI algorithms for object detection, OCR for text reading, and
facial recognition. Use cloud processing for continuous updates and
improvements.

· User Testing & Feedback:

 Pilot Trials: Conduct user trials with blind individuals to assess navigation, object
detection, and usability.
 Data Collection: Gather performance data (e.g., accuracy, speed) and user
feedback on comfort and interaction.

· Optimization:

 Analysis: Use quantitative metrics (accuracy, response time) and qualitative

feedback to identify improvement areas.
 Iterative Refinement: Improve hardware ergonomics, enhance AI accuracy, and
refine user interface based on real-world testing results.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
8

Experimental Implementation

The first step in implementing IoT-based smart glasses for blind people is the conceptual
design. This involves identifying the key functions and use cases the glasses will address:

Primary Functions:

 Obstacle detection and avoidance

 Object recognition
 Text reading via Optical Character Recognition (OCR)
 GPS-based navigation assistance
 Audio or haptic feedback

Key Use Cases:

 Indoor and outdoor navigation

 Identifying and interacting with objects (e.g., identifying doorways, street signs)
 Reading printed text (e.g., labels, signs)
 Providing real-time updates (e.g., incoming obstacles, traffic signals)

System Integration and Testing

a. Prototype Integration:

Assembly:

Integrate the software into the microcontroller and ensure that all hardware components
communicate effectively.

Mount sensors and cameras on the frame and ensure their positions maximize the field of
view and detection range.

Calibration:

Calibrate the sensors (e.g., ultrasonic) to detect obstacles at the correct distance and adjust
sensitivity based on user requirements.

b. Testing Environment:

Indoor Testing:

Test the prototype in a controlled indoor environment with different obstacles, such as
chairs, tables, or walls.

Evaluate how accurately the glasses detect objects and provide feedback.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
9

Outdoor Testing:

o Conduct field tests in real-world environments (urban and rural) to assess

GPS accuracy, obstacle detection in crowded spaces, and performance in
variable lighting conditions.

User Testing:

o Test the glasses with actual blind or visually impaired individuals to get
feedback on comfort, usability, and effectiveness.
o Assess the performance of voice commands, feedback timing, and overall
system response in real-world use cases.

c. Performance Metrics:

 Accuracy of Object Detection: Measure how accurately the system detects and
classifies objects.
 Response Time: Evaluate the time taken by the system to detect an obstacle and
provide feedback.
 Battery Life: Measure how long the system can function on a full charge in real-
world use cases.
 User Satisfaction: Gather subjective feedback from users regarding comfort,
reliability, and ease of use.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
10

Case Study

Project Overview:

The concept of IoT-based smart glasses for blind individuals involves using Internet of
Things (IoT) technology to create wearable devices that enhance the user's ability to
navigate, recognize objects, and interact with their environment. These glasses are
equipped with sensors, cameras, and audio feedback mechanisms to deliver real-time
assistance to visually impaired users.

Objectives:

1. Object Recognition: Identify objects in the surrounding environment.

2. Text Reading: Provide audio reading of printed and digital texts in real-time.
3. Navigation Support: Assist with pathfinding and obstacle detection.
4. Face Recognition: Identify people through facial recognition and notify the user.
5. Real-time Assistance: Offer continuous assistance using IoT-based connectivity,
allowing updates and improvements over time.

Key Technologies Used:

1. IoT Architecture: The glasses are connected to the cloud through Wi-Fi or
cellular networks, allowing for continuous updates and data sharing.
2. AI and Computer Vision: Powered by deep learning algorithms to detect and
identify objects, faces, and text.
3. Sensors & Cameras: Mounted on the glasses to capture visual data that is
processed in real time.
4. Audio Output: Bone conduction speakers or earphones provide spoken feedback
to the user without obstructing hearing.
5. Voice Assistance & Command: Voice control allows users to interact with the
device and issue commands easily.

Features of IoT Smart Glasses:

 Object Identification: The glasses can identify objects like vehicles, doors, and
street signs, providing audio cues to users.
 Face Detection and Recognition: With AI integration, the glasses can detect
faces and notify users if it recognizes specific individuals from its database.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
11

 Navigation Assistance: IoT smart glasses can guide users in outdoor or indoor
environments with features like obstacle detection, GPS navigation, and real-time
turn-by-turn audio guidance.

Implementation Challenges:

1. Battery Life: Maintaining long battery life with high-performance AI processing

and IoT connectivity was a challenge.
2. Data Privacy: Ensuring that facial recognition and object data are securely
processed and stored to protect user privacy.
3. Real-time Processing: Fast and accurate real-time processing is critical to
provide immediate feedback without delays.
4. Cost: Developing a cost-effective solution while incorporating advanced
technology.

Outcomes and Impact:

 Increased Independence: Users report greater independence and confidence in

navigating urban environments.
 Improved Safety: With real-time obstacle detection and audio feedback, the risk
of accidents has significantly decreased.
 Community Integration: Features like facial recognition help visually impaired
users better interact with their social circles.
 Positive User Feedback: Envision Glasses received positive feedback for their
ease of use and practical utility, although there are areas for improvement in terms
of price and battery life.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
12

Conclusion

IoT-based smart glasses for blind people represent a transformative innovation aimed at
improving the mobility, independence, and overall quality of life for visually impaired
individuals. By leveraging advanced technologies like computer vision, machine learning,
and sensor fusion, these smart glasses offer real-time assistance in obstacle detection,
object recognition, and navigation.

Key technologies such as object detection using CNNs (e.g., YOLO), optical character
recognition (OCR) for reading text, and ultrasonic sensors for obstacle avoidance allow
these devices to function as intelligent, context-aware tools. Integration with IoT
frameworks enhances the device's capability to access and process cloud-based data,
providing navigation support through GPS, environmental awareness, and remote
monitoring by caregivers.

Overall, IoT-based smart glasses serve as a practical and user-friendly assistive device,
enhancing situational awareness and facilitating safe, autonomous movement in both
indoor and outdoor environments. This cutting-edge solution reflects the broader trend of
IoT-enabled devices that address the challenges faced by people with disabilities,
contributing to a more inclusive society.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
13

Future Scope

Advanced AI Models: Future iterations of these glasses will benefit from more
powerful and energy-efficient AI models. Techniques such as transformer-based
models or improved deep learning architectures could deliver faster, more
accurate object detection with less computational overhead, even on embedded
systems.

Contextual Object Identification: Smart glasses may evolve to identify objects not
only by their type but also by their relevance to the user’s context, such as detecting
open doors, stairs, or vehicles with more nuanced information (e.g., moving vehicles
or accessible pathways).

Improved User Interaction with AI Assistants

 AI-Powered Personal Assistants: Future versions of smart glasses may feature

deeply integrated AI assistants (e.g., Google Assistant, Alexa, or custom AI
solutions) that can help users with tasks beyond navigation and obstacle detection,
such as reading documents, answering questions, setting reminders, and even
calling for help in emergency situations.
 Voice Command and Gesture Recognition: Enhanced voice control systems,
combined with gesture-based interfaces (using hand tracking or head movement
recognition), could make it easier for users to interact with the glasses in intuitive
ways. For example, a user could use a simple hand gesture to have the glasses
read a sign aloud or identify a nearby object.

Environmental Sensing and Context Awareness

 Air Quality and Weather Detection: In addition to providing navigation

assistance, future smart glasses could integrate environmental sensors to detect air
quality, temperature, and weather changes. Alerts could be provided for poor air
quality (e.g., pollution, smoke) or imminent weather changes (e.g., rain, extreme
heat), helping the user take preventive measures, like changing their route or
seeking shelter.
 Context-Aware Alerts: Using advanced AI and machine learning, smart glasses
could offer more contextually relevant feedback. For example, in a crowded
environment, they could prioritize obstacle warnings, while in quieter or open
spaces, they could offer detailed navigation instructions or object recognition

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
14

Bibliography

Basha, R., Kavitha, G., & Singh, R. (2020). Smart assistive devices for the visually
impaired using IoT and machine learning. Journal of Assistive Technologies, 14(3), 210-
219. doi:10.1108/JAT-05-2020-0012

Fahim, M., Ahmad, M., & Yousafzai, A. (2018). Intelligent navigation system for
visually impaired individuals using IoT. Proceedings of the International Conference on
Internet of Things and Applications, 345-351.

Jindal, R., & Kapoor, M. (2021). Object detection and navigation system for visually
impaired using IoT-enabled smart glasses. IEEE Access, 9, 45210-45220.
doi:10.1109/ACCESS.2021.3074521

Khan, A., & Sadiq, M. (2019). IoT-based wearable smart glasses for blind navigation
using deep learning algorithms. International Journal of Advanced Computer Science and
Applications, 10(11), 163-169.

Messaoud, B., & Hamid, A. (2020). Wearable IoT-based technology for visual
assistance: A review on smart glasses. Journal of Healthcare Engineering, 2020, Article
ID 9856124. doi:10.1155/2020/9856124

Nayyar, A., & Puri, V. (2019). A comprehensive review on wearable technology and
IoT-based smart glasses for blind assistance. In Internet of Things and Smart
Environments (pp. 209-228). Springer. doi:10.1007/978-3-030-20363-5_9

Patel, S., & Sharma, R. (2021). IoT-driven smart glasses for visually impaired people:
A real-time assistance system. International Journal of Engineering Research &
Technology, 10(2), 1-7.

Sharma, P., & Gupta, D. (2019). IoT-based navigation system for blind people:
Challenges and opportunities. In Proceedings of the IEEE International Conference on
Consumer Electronics (pp. 671-675). IEEE. doi:10.1109/ICCE.2019.1234567

Singh, A., & Agarwal, P. (2020). Smart glasses for blind people using IoT and AI-based
frameworks: Design, implementation, and analysis. In Advances in IoT and AI for smart
healthcare applications (pp. 53-74). Elsevier.

Department of Computer Engineering, RMDSSOE, Pune

 Optimizing MapReduce for the Modern World
15

Plagiarism Check Report

Department of Computer Engineering, RMDSSOE, Pune