Techno-Societal 2024
“AI and ML in Water Quality Analysis: A Review
of Techniques, Challenges, and Future Directions”
Paper ID-316
Name of Presenter -1. Prof. Sabanaz S. Peerzade
2. Mahima Kukadapwar
3. Aditi Kulkarni
4. Nidhi Kumar
College– Sinhgad College of Engineering
Venue: SVERI’S College of Engineering , Pandharpur
(An Autonomous Institute)
Introduction
I. Global Concern
Water quality management is a significant global challenge due to pollution from urbanization,
industrial activities, and agriculture, which degrade aquatic ecosystems and harm human health.
II. Project Overview
Our project focuses on classifying, detecting and mapping the growth of aquatic plants (e.g. water
hyacinth) using datasets and machine learning techniques.
III. Proposed Solution
Develop an AI-powered system that analyzes images of water bodies to detect and classify invasive
species like water hyacinth. The system will leverage machine learning algorithms to automate the
identification process, enabling efficient and scalable monitoring of aquatic ecosystems.
13th -14th December 2024 Techno-Societal 2024 2
Scope of research work
Exploration of AI/ML 2 Policy Development Support
1
Techniques:
Apply the developed models to
Investigates the application of
other regions affected by
machine learning models such as
eutrophication to assess their
Neural Networks, Support Vector
effectiveness.
Machines, Random Forests for
water quality analysis.
3 Real-Time Monitoring 4 Policy Development Support
System
Utilize model insights to inform and
Develop a platform for real-time support the creation of environmental
monitoring and prediction of policies aimed at nutrient
eutrophication events. management.
13th -14th December 2024 Techno-Societal 2024 3
Methodology
Data Collection
We are collecting datasets from various public and private sources to ensure the data is accurate,
consistent, and aligned with the project goals.
Preprocessing
Clean and preprocess acquired dataset for enhanced accuracy and clarity. Normalize and prepare
environmental datasets for analysis.
Feature Extraction
We are extracting relevant features from the preprocessed dataset to enable accurate identification of
invasive water species.
13th -14th December 2024 Techno-Societal 2024 4
Methodology
Model Development
We are applying Convolutional Neural Networks (CNN) to classify and predict the presence of water
4 hyacinth. Future work may involve using Random Forest for additional classification tasks.
Validation and Testing
5 Validate model accuracy using test data and refine the model to improve predictions.
Visualization
6 To Provide user with a Clear visual representation of models output showing growth of water hyacinth
13th -14th December 2024 Techno-Societal 2024 5
Experimentation
13th -14th December 2024 Techno-Societal 2024 6
Experimentation
13th -14th December 2024 Techno-Societal 2024 7
Experimentation
Specific Class Performance
•Water Lettuce (Pistia stratiotes): The
model correctly classified 302 instances
of this class, with 6 misclassifications.
•Heartleaf False Pickerelweed
(Monochoria korsakowii): The model
correctly classified 284 instances, with 1
misclassification.
•Common Duckweeds (Lemna minor):
The model correctly classified 268
instances, with 1 misclassification.
•Common Water Hyacinth (Eichornia
crassipes): The model correctly
classified 299 instances, with 8
misclassifications.
13th -14th December 2024 Techno-Societal 2024 8
Experimentation
13th -14th December 2024 Techno-Societal 2024 9
Result Discussion
Key Findings: Impact of AI/ML:
1. AI/ML models demonstrated high •Enhanced anomaly detection led to early
accuracy in predicting water quality intervention for pollution events.
parameters (e.g., pH, turbidity, •Optimized treatment processes reduced costs
dissolved oxygen). and improved efficiency in water management.
•Geolocation techniques accurately pinpointed
2. Random Forests effectively contamination sources, guiding remediation
identified primary drivers of efforts.
eutrophication, such as nutrient
runoff and land use.
Challenges Observed:
3. Neural Networks excelled in •Data limitations (noise, inconsistencies)
processing satellite imagery for impacted some model predictions.
detecting chlorophyll concentrations •Scalability issues in handling large datasets
and anomalies. and real-time applications.
•Ethical concerns regarding transparency and
data privacy.
13th -14th December 2024 Techno-Societal 2024 10
Conclusion
• AI and ML have proven to be transformative tools in water quality analysis, offering
solutions for predictive modeling, anomaly detection, process optimization, and
geolocation.
• Despite advancements, challenges like data quality, model scalability, and ethical
concerns must be addressed for broader adoption.
• Further integration with technologies like big data, IoT, and remote sensing can
enable more robust and scalable solutions for global water quality challenges.
• AI and ML hold immense potential to ensure sustainable, safe, and efficient water
management practices for a healthier planet.
13th -14th December 2024 Techno-Societal 2024 11
References
M. A. Pirbasti, V. Akbari, D. Bhowmik, M. Savitri, and A. Marino,
“Detection and Mapping of Water Hyacinth Growth Cycle in Anzali International Wetland Using
Sentinel-2 Time Series,” IEEE J Sel Top Appl Earth Obs Remote Sens, vol. 17, pp.
13346–13357, 2024, doi: 10.1109/JSTARS.2024.3427002.
M. S. Hernández, “Beliefs and attitudes of canarians towards the chilean linguistic
variety,” Lenguas Modernas, no. 62, pp. 183–209, 2023, doi: 10.13039/501100011033.
E. C. Rodriguez-Garlito, A. Paz-Gallardo, and A. Plaza, “Monitoring the
Spatiotemporal Distribution of Invasive Aquatic Plants in the Guadiana River, Spain,”
IEEE J Sel Top Appl Earth Obs Remote Sens, vol. 16, pp. 228–241, 2023, doi:
10.1109/JSTARS.2022.3225201.
E. Salcedo, Y. Uchani, M. Mamani, and M. Fernandez, “Towards Continuous Floating
Invasive Plant Removal Using Unmanned Surface Vehicles and Computer Vision,”
IEEE Access, vol. 12, pp. 6649–6662, 2024, doi: 10.1109/ACCESS.2024.3351764.
• SRS: https://drive.google.com/file/d/1F6tmPoZ9YJ0qzX_7dGJRYtPqrkp7ZTuW/view?usp=sharing
13th -14th December 2024 Techno-Societal 2024 12
Thank You
• Name of Presenters-
1. Prof. Sabnaz S.Peerzade
3. Mahima Kukadapwar
2. Aditi Kulkarni
3. Nidhi Kumar
• Mail Id-
1. [email protected]
2. [email protected]
3. [email protected]
4. [email protected]
• Contact :+91 80871 24860
• College Name-Sinhgad College of Engineering
13th -14th December 2024 Techno-Societal 2024 13
