School of

Electronics and Communication Engineering

Mini Project Report

on

POTHOLE DETECTION AND MITIGATION SYSTEM

By:

1. Rakshita Neeralkatti USN: 01FE22BEC074

2. Sugnyani USN: 01FE22BEC064

3. Shreya V S USN: 01FE22BEC071

4. Gouri Metgud USN: 01FE22BEC034

Semester: V, 2024-2025

Under the Guidance of

Prof. P. C. Nissimgoudar

1
 K.L.E SOCIETY’S

HUBBALLI-580031
2024-2025

SCHOOL OF ELECTRONICS AND COMMUNICATION ENGINEERING

CERTIFICATE
This is to certify that project entitled “Pothole Detection and Mitigation System” is a bonafide work carried out by
the student team of “Rakshita Neeralkatti(01fe22bec074), Sugnyani(01fe22bec064), Shreya V
S(01fe22bec071) and Gouri Metgud(01fe22bec034)”. The project report has been approved as it satisfies the
requirements with respect to the mini project work prescribed by the university curriculum for BE (V Semester) in
School of Electronics and Communication Engineering of KLE Technological University for the academic year 2024-
2025.

Your Guide Suneeta V Budihal B. S. Anami Guide Head of School Registrar

External Viva:

Name of Examiners Signature with date

1.
2.

2
ACKNOWLEDGMENT

We would like to express our sincere gratitude to everyone who supported and guided
us throughout the development of this project, without whom its successful
completion would not have been possible. First and foremost, we extend our deepest
thanks to our esteemed mentors and faculty members for their invaluable guidance,
encouragement, and constructive feedback throughout the course of this work. Their
expertise and insights were instrumental in refining our ideas and overcoming
challenges, ensuring the project's success. We are also immensely grateful to Kaggle
and similar platforms for providing access to essential datasets, which formed the
backbone of our research and development process. The contributions of data
scientists, researchers, and the wider community in making such resources publicly
available have been vital in driving innovation and progress in this field. A special
note of appreciation goes to our institution for providing us with the infrastructure,
tools, and a conducive environment to work on this project. Their support played a
significant role in enabling us to explore creative solutions and implement innovative
methodologies for the detection of potholes using advanced technologies. We would
also like to acknowledge the previous researchers and developers whose work served
as a foundation for our study. Their efforts in the domain of machine learning,
computer vision, and transportation safety inspired us to build upon their ideas and
contribute further to this important field. Moreover, we extend our heartfelt thanks to
our teammates for their tireless dedication, collaboration, and teamwork. The spirit of
cooperation and mutual support within our group allowed us to overcome obstacles
and achieve milestones effectively. Lastly, we are profoundly grateful to our families
and friends for their unwavering encouragement, understanding, and support
throughout this journey. Their belief in our abilities and constant motivation helped us
remain focused and determined even during the most challenging phases of the
project. To all those who contributed in one way or another, directly or indirectly, we
express our heartfelt gratitude. This project is a reflection of the collective efforts,
guidance, and support we have received, and we dedicate its success to all of you.
-The project team

3
 ABSTRACT
Potholes are a major road safety concern, causing accidents, vehicle damage, and traffic
inefficiencies. To address this issue, we developed a real-time Pothole Detection and Mitigation
System using YOLOv8 for object detection, GPS mapping, and cloud-based storage. The system
captures road images via a live webcam feed, detects potholes with high accuracy, and maps
their locations using GPS. Detected potholes are stored in Firebase for scalability and displayed
on an interactive platform, allowing users to view them in real time. Drivers receive instant alerts
when approaching potholes, enhancing safety and reducing potential damage.

Our system ensures efficiency and responsiveness by leveraging batch processing, reduced
image resolution, and spatial indexing. With a detection precision of 97.9% and an mAP50 of
89.4%, YOLOv8 enables reliable pothole identification. Unlike traditional methods that rely on
manual inspections or crowdsourced data, our automated approach reduces human intervention
and improves accuracy. The cloud-based infrastructure, particularly Firebase, allows seamless
data storage and retrieval, making the system adaptable for large-scale deployment. The
interactive mapping feature also aids authorities in proactive road maintenance by providing
real-time insights into pothole-prone areas, facilitating timely repairs and reducing long-term
infrastructure costs.

By integrating modern vision technology, cloud computing, and GPS tracking, our solution
offers a cost-effective and scalable approach to pothole detection and mitigation. The system’s
ability to provide real-time updates and integrate with vehicle navigation significantly enhances
driving safety and traffic flow in both urban and rural areas. With its high detection accuracy,
instant alerts, and interactive mapping, this approach can be widely implemented to improve
road conditions, reduce accidents, and optimize transportation efficiency.

4
5
6
 CONTENTS

1 Introduction 9
1.1 Motivation 9
1.2 Objectives 9
1.3 Literature survey 9
1.4 Problem statement 10
1.5 Application in Societal Context 10
1.6 Project Planning and bill of materials 10
1.7 Organization of the report 10
2 System design 11
2.1 Functional block diagram 11
2.2 Design alternatives 11
2.3 Final design 12
3 Implementation details 13
3.1 Specifications and final system architecture 13
3.2 Algorithm 13
3.3 Flowchart 13
4 Optimization 14
4.1 Introduction to optimization 14
4.2 Types of optimization 14
4.3 Selection and justification of optimization method 14
5 Results and discussions 15
5.1 Result Analysis 15
5.2 Discussion on optimization 15
6 Conclusions and future scope 16
6.1 Conclusion 16
6.2 Future scope 16

7
 References 16

1.INTRODUCTION

The Pothole Detection and Mitigation System addresses critical road safety challenges by
leveraging advanced technologies to identify, record, and alert drivers to pothole hazards in real
time. Potholes not only cause significant damage to vehicles but also contribute to accidents
and increased maintenance costs. This project integrates cutting-edge computer vision, GPS
mapping, and cloud computing to mitigate these issues effectively.

1.1 Motivation:
The motivation behind this project is to improve road safety and enhance driver awareness.
Potholes are a significant concern for road safety, causing accidents, vehicle damage, and traffic
inefficiencies. Addressing this issue can lead to safer and more cost-effective transportation
systems. This system aims to offer a proactive solution, addressing these gaps to create safer
roads.

1.2 Objectives:
1. To develop a pothole detection system using YOLOv8 for high accuracy and reliability.
2. Implement scalable cloud storage using Firebase for data management.
3. Integrate GPS mapping for precise pothole location tracking.
4. Create an alert mechanism to warn drivers of nearby potholes.
5. Ensure the system is scalable, efficient, and reliable.
6. To design a scalable and cost-efficient system adaptable to diverse road conditions.

8
1.3 Literature Survey:
thorough review of existing research was conducted to understand different approaches to
pothole detection and their limitations. The selected papers highlight various methods,
including machine learning, sensor-based techniques, and cloud integration. Below is an
analysis of four relevant papers and the justification for selecting our proposed approach.
1. Detection of Potholes and Speed Breakers for Autonomous Vehicles (iSCSi-2024):This
study explores YOLOv5-based detection of potholes and speed breakers for autonomous
driving applications. The model uses image processing and accelerometers to classify road
anomalies. However, the dataset used was relatively small, and the approach lacked real-time
driver alerts. While the system showed promising accuracy, it struggled with detecting small
potholes and unmarked speed breakers. Additionally, its low-light detection capability was
inconsistent, and no adaptive model strategy was implemented.
2. Convolutional Neural Networks-Based Pothole Detection Using Thermal Imaging
(CSIR-2021):This research introduced thermal imaging and deep learning (CNN) for pothole
detection. Using ResNet-based models, the study achieved a high accuracy of 97.08%. The
approach effectively identified potholes of varying severity, leveraging data augmentation and
transfer learning. However, the study was limited to controlled datasets, lacking real-time
deployment validation. Furthermore, its reliance on thermal imaging restricts its usability in
daytime conditions, making it impractical for continuous monitoring.
3. Pothole Detection System Using Ultrasonic Sensors (IRJETS-2023):This paper proposed
a sensor-based pothole detection system that uses ultrasonic sensors, microcontrollers, GPS,
and Bluetooth. The approach marks potholes' locations using GPS and transmits data via
Bluetooth to a mobile application. While effective for road maintenance teams, the approach
lacks real-time driver alerts and interactive visualization. Additionally, ultrasonic sensors have
limited detection range and accuracy, making them unsuitable for high-speed detection
scenarios.
4. Road Pothole Detection Mechanism Using Mobile Sensors (ICTAI-2021):This study
utilized smartphone accelerometers and gyroscopes to detect potholes using machine learning
classifiers such as Decision Trees, SVM, and XGBoost. The data was collected via a mobile
app, which stored pothole locations in a database for analysis. While this method enabled cost-
effective, crowd-sourced detection, it lacked real-time alerts, cloud integration, and precise
mapping capabilities. The study also highlighted challenges related to data imbalance and
inconsistent sensor accuracy.
Selected Paper & Justification for Our Approach
The selected solution is Vision-Sensor-Based Detection with Cloud Integration (YOLOv8 +
Firebase). Compared to the surveyed approaches, this method provides high detection accuracy,
real-time alerts, GPS-based mapping, and cloud scalability. Unlike accelerometer or sensor-
based methods, YOLOv8 ensures robust detection, even under varying road conditions. The
real-time data processing and Firebase integration enable immediate updates and visualization,
addressing the limitations of prior studies.

9
By leveraging deep learning, cloud computing, and GPS tracking, this system offers a
scalable, cost-effective, and real-time solution for pothole detection and mitigation, ensuring
safer roads and improved driver awareness.

1.4 Problem Statement:

Potholes on roads pose significant risks to vehicle safety, passenger comfort, and infrastructure
maintenance. Existing detection systems often face challenges such as limited accuracy, high
false positives, and lack of scalability. This project aims to address these limitations by
developing a robust pothole detection system using YOLOv8, integrated with cloud-based
platforms and real-time alert mechanisms to enhance road safety and maintenance efficiency.

1.5 Application in Societal Context:

This project has substantial societal benefits, including:
1. Enhances road safety by reducing accidents caused by potholes.
2. Minimizes vehicle maintenance costs by alerting drivers about potholes in real time.
3. Assists local authorities in identifying and prioritizing road repairs.
4. Improves traffic efficiency by preventing sudden stops or accidents due to potholes.
5. Provides a scalable solution that can be implemented across urban and rural areas.

1.6 Project Planning and Bill of Materials:

The project was planned in multiple phases, including data collection, model training, system
integration, and testing. The key materials used in the project include:
- Hardware: - Webcam or camera module for real-time image capture.
- GPS module for precise location tracking.
- Computing device (e.g., Raspberry Pi or laptop) for processing.

- Software: - YOLOv8 model for object detection.

- Firebase for cloud storage.
- Flask for backend development.
- Leaflet.js for interactive map visualization.
- Dataset: Annotated pothole images for training and validation.

10
1.7 Organization of the Report:
The report is structured as follows:
- Chapter 1: Introduction to the project, including objectives and scope.
- Chapter 2: Literature survey summarizing previous research and methodologies.
- Chapter 3: Methodology detailing the techniques and tools used.
- Chapter 4: Results and discussions highlighting key findings and system performance.
- Chapter 5: Conclusion and future work outlining the project’s impact and potential
improvements.

2. SYSTEM DESIGN

This chapter outlines the design of the pothole detection system, detailing the functional block
diagram, alternative designs considered, and the final chosen design.

2.1 Functional Block Diagram

The functional block diagram represents the workflow of the system, which includes the
following components:
11
· Data Collection and Detection

Utilizes webcam input, frame capture, YOLOv8 detection, GPS data simulation, and data
formatting for detecting potholes.

· Data Storage

Stores processed data in Firebase using the Admin SDK and a real-time database system for
consistent data collection.

· Web Visualization

Implements a Leaflet map interface to display real-time data with pothole markers and
distance processing for enhanced visualization.

· Alert Management

Uses a Flask server for backend processing, along with location monitoring and an alert
system for proactive notifications.

· Integration Flow

Seamless integration of data detection, storage, visualization, and alert management ensures
end- to-end pothole tracking and notification.

2.2 Design Alternatives:

1. Model Selection Alternatives:
- YOLOv5 vs. YOLOv8: YOLOv8 was chosen due to its higher accuracy and faster detection
speed compared to YOLOv5.
- CNN-based models were considered but lacked real-time efficiency.
2. Data Storage Alternatives:
- Firebase vs. Local Database: Firebase was selected for its scalability and ability to handle
real-time updates.
- Local storage was avoided due to limitations in scalability and accessibility.
3. Alert Mechanism Alternatives:
- Proximity-based alerts vs. Constant notifications: Proximity-based alerts were chosen to
reduce unnecessary notifications, enhancing user experience.
- Only visual alerts were deemed insufficient, leading to the inclusion of audio notifications.
4. Visualization Alternatives:

12
 - Static maps vs. Interactive maps: Interactive maps using Leaflet.js were preferred for better
user engagement and real-time updates.

2.3 Final Design:

The final design for the pothole detection and mitigation system is optimized for real-time
performance, leveraging YOLOv8 for accurate detection . The system emphasizes timely alerts
and data analysis to facilitate effective road maintenance. The key components of the design
are as follows:

1. Real-Time Video Input

The system captures real-time video data from cameras mounted on vehicles or infrastructure.
This provides a dynamic view of road conditions, ensuring the detection of potholes under
varying scenarios such as changing lighting, weather, and traffic patterns. High-definition video
input is critical for accurately identifying road surface anomalies.

2. Image Preprocessing
To enhance the accuracy and robustness of the detection process, video frames undergo
preprocessing before being fed into the YOLOv8 model. The preprocessing steps include:
- Noise Reduction: Reduces the impact of noise and artifacts, ensuring clearer input for
detection.
- Contrast Enhancement: Improves the visibility of potholes by increasing the contrast between
the pothole and the surrounding road surface.
- Image Resizing: Adjusts the frame size to match the input requirements of the YOLOv8
model, ensuring compatibility and efficient processing.

3. YOLOv8-Based Pothole Detection

The YOLOv8 model is the core of the detection system, selected for its real-time object
detection capabilities and high accuracy. YOLOv8 processes each pre-processed frame,
identifying potholes by drawing bounding boxes around them and assigning confidence scores.
The model is trained on a diverse dataset to ensure reliable detection under different road and
environmental conditions.

5. Data Processing and Analysis

After detection, the system processes and analyses the detected potholes to generate actionable
insights. Key aspects include:
- Pothole Localization: Determines the precise location of potholes using GPS data or other
localization techniques.
13
- Data Storage: Saves detection data, including timestamps, locations, and possibly dimensions
or severity of potholes, for further analysis.
- Alert Generation and Reporting: Generates real-time alerts and notifications for relevant
authorities or maintenance teams, enabling prompt action.

14
3.IMPLEMENTATION DETAILS

3.1 Specifications and Final System Architecture

The implementation of the pothole detection and mitigation system is designed to ensure high
accuracy, scalability, and ease of deployment. The system is built using the following
specifications:
1. Specifications:
Hardware:
- Webcam or camera module for capturing real-time images.
- GPS module for accurate geolocation of potholes.
- Processing unit: Laptop or Raspberry Pi for model execution.
Software:
- YOLOv8 for real-time object detection.
- Firebase for scalable cloud storage.
- Flask framework for backend development and API integration.
- Leaflet.js for interactive map visualization.

2. Final System Architecture:

- Input Layer:
- Captures live video feed and GPS data.
- Processing Layer:
- YOLOv8 processes frames to detect potholes.
- Spatial indexing associates detected potholes with geolocation data.
- Data Layer:
- Detected pothole data (image and GPS coordinates) is sent to Firebase for storage.
- Output Layer:
- Interactive map updates pothole locations in real time.
- Alerts are sent to drivers using proximity calculations.

15
3.2 Algorithm
The following steps outline the algorithm used for the pothole detection system:

1. Input:
- Capture real-time road images using a webcam.
- Obtain GPS location using the GPS module.

2. Detection:
- Use YOLOv8 to process the captured images and detect potholes.
- Extract bounding box coordinates for identified potholes.

3. Data Storage:
- Send detected pothole details (image, GPS location, confidence score) to Firebase for
storage.

4. Visualization:
- Update an interactive map (using Leaflet.js) with pothole markers based on GPS data.

5. Alert Mechanism:
- Calculate the distance between the vehicle and detected potholes.
- Trigger proximity-based audio and visual alerts for drivers.

6. Output:
- Display detected potholes on a real-time map.
- Notify drivers about upcoming potholes to ensure safe driving.

16
3.3 Flowchart

17
4.OPTIMIZATION

4.1 Introduction to Optimization

Optimization refers to the process of making a system, design, or decision as effective or

functional as possible. In the context of engineering and computer systems, optimization aims
to improve performance, reduce resource usage, and enhance efficiency. By identifying and
refining key aspects, optimization ensures that desired outcomes are achieved with minimal
cost, time, and effort. Optimization techniques are critical for solving complex problems in a
structured and efficient manner.

4.2 Types of Optimization

Optimization plays a crucial role in enhancing the efficiency, accuracy, and scalability of the
Pothole Detection and Mitigation System. Several optimization techniques have been
implemented to ensure real-time pothole detection with minimal latency and resource
consumption. The key types of optimization used in this project include:
1. Model Optimization – To improve the YOLOv8 model’s efficiency, hyperparameter
tuning was performed, optimizing parameters such as learning rate, batch size, and
confidence threshold. This helps achieve high accuracy while maintaining processing
speed.
2. Data Optimization – Data compression techniques were applied to reduce image sizes
without losing critical pothole detection information. This ensures smooth data
transmission and storage while maintaining detection accuracy.
3. Computational Optimization – Spatial indexing was employed to minimize
computational overhead by focusing only on relevant pothole data. This reduces
unnecessary processing, enhancing overall system performance.
4. Alert Optimization – Proximity-based notifications were optimized using geospatial
calculations, ensuring that drivers receive timely alerts without excessive or redundant
warnings, improving usability and driver response time.
By combining these techniques, the system achieves high detection accuracy, real-time
processing, and efficient data handling, making it scalable and effective for large-scale
deployment.

4.3 Selection and Justification of Optimization Method

The selection of an appropriate optimization method depends on the problem’s characteristics,

such as the nature of the objective function, constraints, and variables. For this project, the
optimization approach was selected based on the following criteria:

Problem Characteristics:
18
The system involves detecting potholes using YOLOv8, where speed and accuracy are critical.

Real-time data processing requires minimal latency and high efficiency.

Selected Method:

Nonlinear Optimization: Chosen to improve the YOLOv8 model’s performance by tuning

hyperparameters and refining its architecture for faster and more accurate detection.

Data Compression: Implemented to optimize storage and transmission efficiency.

Spatial Indexing: Used to reduce computational load by focusing only on relevant pothole
data.

Justification:

The nonlinear optimization method is ideal for handling complex relationships between input
data and detection accuracy.

Data compression ensures efficient storage without compromising the integrity of real-time
updates.

Spatial indexing minimizes unnecessary processing, improving overall system performance and
scalability.

By applying these optimization techniques, the system achieves enhanced performance,

reliability, and scalability, meeting the objectives of real-time pothole detection and mitigation.

19
5. RESULT AND DISCUSSIONS

5.1 Result Analysis:

The Pothole Detection and Mitigation System was rigorously evaluated using the Indian Road
Dataset, which comprised 2037 images for training and 600 images for testing. The system's
performance was assessed based on detection accuracy, real-time processing, integration with
GPS mapping, and the effectiveness of its alert mechanism. The YOLOv8-based detection
model demonstrated high precision and recall, accurately identifying potholes with bounding
boxes and confidence scores. The model achieved a precision of 97.9% and an mAP50 of
89.4%, ensuring robust performance across different lighting conditions, road surfaces, and
perspectives.
In terms of real-time performance, the system successfully processed live video feeds with
minimal latency, enabling seamless pothole detection even at high vehicle speeds.
Optimizations such as batch processing, reduced image resolution, and spatial indexing
contributed to its efficiency, allowing the model to detect potholes within milliseconds. This
capability makes it suitable for real-world deployment in vehicles, ensuring that potholes are
identified promptly and accurately.
The system also integrated GPS data with mapping services, enabling precise localization of
detected potholes. Each detected pothole was automatically marked on an interactive map,
providing authorities with real-time insights into pothole-prone areas. This mapping feature
supports proactive road maintenance, helping municipalities and highway agencies prioritize
repairs, ultimately improving road infrastructure and driving conditions.
Additionally, the system implemented an effective alert mechanism, generating real-time audio
and visual notifications for drivers when approaching detected potholes. These alerts were
optimized using proximity-based calculations, ensuring timely warnings while minimizing
false alarms. This feature significantly enhances road safety, helping drivers take necessary
precautions and reduce the risk of vehicle damage or accidents.
Overall, the system's evaluation demonstrated high detection accuracy, efficient real-time
processing, seamless GPS integration, and a well-optimized alert system. By combining deep
learning, cloud computing, and geospatial tracking, the solution offers a scalable, reliable, and
cost-effective approach to pothole detection and mitigation, ultimately contributing to safer and
smoother road conditions.

20
5.2 Discussion on Optimization:
The system's optimization efforts focused on improving detection accuracy, reducing
computational load, and ensuring scalability:
- Model Optimization: YOLOv8 was chosen for its lightweight architecture and fast inference,
enabling real-time detection without compromising accuracy.
- Cloud Integration: Storing processed data in Firebase allowed for scalable and efficient data
management, supporting real-time updates and alerts.
- Proximity Alerts: By simulating vehicle location and calculating distances from potholes, the
system efficiently triggered alerts, minimizing false positives.
- Cost Efficiency: The integration of vision sensors and cloud technologies ensured a balance
between performance and affordability, making the system viable for large-scale deployment.

These optimizations underline the system's potential for improving road safety and driver
awareness, setting a benchmark for future advancements in pothole detection technologies.

21
6. CONCLUSION AND FUTURE SCOPE

6.1 Conclusion:
The proposed pothole detection and mitigation system successfully addresses critical gaps in
road safety by integrating real-time computer vision, cloud processing, and GPS-based
mapping. Key achievements include:
- High Detection Accuracy: The YOLOv8-based model reliably identifies potholes with precise
bounding boxes and confidence scores.
- Real-Time Alerts: The system provides timely audio and visual alerts to drivers, enhancing
safety and awareness.
- Scalable Integration: Cloud-based data storage and processing ensure efficient management
and distribution of pothole data.
- Practical Implementation: The system is designed to be cost-effective, making it suitable for
widespread deployment on Indian roads.
By reducing pothole-related incidents and improving driver awareness, this system
demonstrates significant potential to enhance road safety and infrastructure monitoring.

6.2 Future Scope:

The project opens avenues for further enhancements and broader applications, including:
1. Advanced Detection Capabilities:
- Expanding the system to detect other road hazards such as cracks, debris, or faded lane
markings.
- Incorporating multi-modal sensors (e.g., LiDAR) for enhanced detection in adverse
conditions like rain or low visibility.
2. Integration with Autonomous Vehicles:
- Leveraging the system for Advanced Driver Assistance Systems (ADAS) in autonomous
vehicles.
- Enabling autonomous navigation adjustments based on real-time pothole detection.
3. Enhanced Data Analytics:
- Using machine learning to analyze long-term road condition data for predictive
maintenance.
- Developing insights to assist government agencies in prioritizing road repairs.
4. Global Scalability:

22
 - Adapting the system for deployment in international contexts by training models on diverse
datasets.
- Integrating with global mapping platforms for seamless cross-border applications.
5. IoT and Edge Computing:
- Deploying edge computing devices to process data locally, reducing latency and dependence
on cloud services.
- Integrating IoT devices for continuous monitoring and real-time updates.

This roadmap highlights the system's potential to evolve into a comprehensive road safety
solution, contributing to smarter and safer transportation networks worldwide.

7. REFERENCES

1. Detection of Potholes and Speed Breaker for Autonomous Vehicles. Potholes and Speed
Breaker for Autonomous Vehicles P C Nissimagoudar1, Sanjana R Miskin1, Varsha N
Sali1, Ashwini J1, Rohit S K1, Darshan S K1, Gireesha H M1, Rohini S Hongal1, Supriya
V Katwe1, Basawaraj1, Nalini C I1 KLE Technological University, Hubballi, India.

2. Convolutional Neural Networks Based Potholes Detection Using Thermal Imaging.

Aparna, Yukti Bhatia, Rachna Rai, Varun Gupta, Naveen Aggarwal, Aparna Akula,
Convolutional neural networks based potholes detection using thermal imaging, Journal
of King Saud University - Computer and Information Sciences, Volume 34, Issue 3, Pages
578-588, ISSN 1319-1578, https://doi.org/10.1016/j.jksuci.2019.02.004.

3. Road Pothole detection mechanism using Mobile Sensors. H. Agrawal, A. Gupta, A.

Sharma and P. Singh, "Road Pothole Detection Mechanism using Mobile Sensors," 2021
International Conference on Technological Advancements and Innovations (ICTAI),
Tashkent, Uzbekistan, 2021, pp. 26-31, doi: 10.1109/ICTAI53825.2021.9673193.

4. Pothole Detection and Road Condition Updation on Google maps . Nissimagoudar, P.C.,
Basawaraj, Gireesha, H.M., Kulkarni, A., Bhat, S., Iyer, N.C. (2023). Pothole Detection
and Road Condition Updation on Google Maps. In: Choudrie, J., Mahalle, P.N., Perumal,
T., Joshi, A. (eds) IOT with Smart Systems. ICTIS 2023. Lecture Notes in Networks and
Systems, vol 720. Springer, Singapore. https://doi.org/10.1007/978-981-99-3761-5_5

5. AI-based Detection of Potholes Ahead of a Visually Impaired Person Using Ultrasonic

Sensors Array and Camera for Blind Navigation - H. Hamed and M. R. Khan, "AI-based
Detection of Potholes Ahead of a Visually Impaired Person Using Ultrasonic Sensors

23
Array and Camera for Blind Navigation," 2024 9th International Conference on
Mechatronics Engineering (ICOM), Kuala Lumpur, Malaysia, 2024, pp. 448-453, doi:
10.1109/ICOM61675.2024.10652516.

24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41