ENVIRONMENT MONITORING AI

(FOCUSING ON A SPECIFIC DISEASE)

PHASE 3: COMMUNICATION AND FUTURE

EXPLORATION

ABSTRACT :

This project focuses on developing an AI assisted environmental

monitoring for better environmental abnormality detection with the help of generic
monitoring hardware which increases it’s efficiency with a marginal difference .

This model will be trained on various environmental

abnormalities and will be allowed to monitor the environment with the help of the
sensors and analyze the environment .This model primarily focus on the patterns
and the data provided by the sensors then it analyses those data to give a extensive
monitoring report about the environment .

This helps in making the environment monitoring procedure less

laborious and also makes the outcome more reliable and consistent , these data can
also be used too predict the upcoming abnormalities with the help of the ai trained
with the previous abnormality data pattern .
This project focuses on

The project is about analyzing and evaluating the data obtain from
monitoring the environment to predict and monitor the environment in a
reinforced and accurate manner.

In conclusion, our project represents a significant step forward in

the application of AI to Environmental monitoring .This will transform the way
we monitor and predict the environmental abnormalities .
 SYSTEM REQUIREMENTS:

Generic hardware requirements that are needed

1. Hardware Requirements:
- Sensors: Temperature, humidity, air quality (CO2, CO, particulate matter), UV
index, light intensity, sound level, etc.
- Microcontroller or Single Board Computer (SBC): Raspberry Pi, Arduino, or
similar for data processing and communication.
- Power supply: Battery or mains power depending on deployment location.
- Enclosure: Waterproof and weatherproof housing if deployed outdoors.
- Communication module: Wi-Fi, GSM, or LoRa for data transmission.
- Optional: Camera for image recognition (to detect pollution sources, wildlife
monitoring, etc.).

2. Software Requirements:
- Operating System: Linux-based OS like Raspbian for Raspberry Pi or Debian
for other SBCs.
- Data Collection and Processing: Python or C/C++ for programming the
microcontroller/SBC to collect data from sensors and process it.
- Machine Learning Libraries: TensorFlow, PyTorch, or scikit-learn for building
AI models.
- Communication Protocol: MQTT, HTTP, or WebSocket for transmitting data to
the cloud or a central server.
- Database: MongoDB, InfluxDB, or SQLite for storing sensor data.
- Web Server: Flask or Django for creating a web interface to visualize data.
- Visualization Tools: Grafana, Plotly , or Matplotlib for creating graphs and
charts.
- Alerting System: Email, SMS, or push notifications for alerting stakeholders of
abnormal environmental conditions.

3. Tools and Versions:

- Python: Version 3.x for scripting and data processing.
- TensorFlow or PyTorch : Latest stable versions for building and training
machine learning models.
- MQTT: Eclipse Mosquitto or VerneMQ for implementing the MQTT protocol.
- Flask or Django: Depending on the complexity of the web interface, choose the
appropriate framework.
- Grafana: Latest version for real-time monitoring and visualization of sensor
data.
- MongoDB or InfluxDB : Choose based on the specific requirements for storing
time-series sensor data.
- Git: Version control system for managing codebase and collaboration.
- Docker: Containerization for easy deployment and scalability.
- Ansible or Puppet: Configuration management tools for automating deployment
and updates across multiple devices.

FLOWCHART:

Start Activate GSM

Initialization

Send data to Thing Speak

Circuit configuration

Deactivate GSM connection.

Acquire sensor data

User defined delay

NO
Ok?

Yes

Data processing

Wait

No
Verify?
Yes
CODE IMPLEMENTATION(SAMPLE CODE)

import time
import board
import busio
import adafruit_bme280
import adafruit_sgp30
import adafruit_veml6075
import adafruit_micropython_midi
import numpy as np
import matplotlib.pyplot as plt
from sklearn.ensemble import RandomForestRegressor

# Initialize sensors
i2c = busio.I2C(board.SCL, board.SDA)
bme280 = adafruit_bme280.Adafruit_BME280_I2C(i2c)
sgp30 = adafruit_sgp30.Adafruit_SGP30(i2c)
veml6075 = adafruit_veml6075.VEML6075(i2c)
midi = adafruit_micropython_midi.MIDI(midi_out=board.D6)

# Initialize arrays to hold sensor data

temperatures = []
humidities = []
co2 = []
co = []
pm = []
uv = []
light = []
sound = []

# Initialize arrays to hold time data

times = []

# Set up training data

X = []
y = []

# Set up model
model = RandomForestRegressor(n_estimators=100,
random_state=42)

# Main loop
while True:
# Read sensor data
temperature = bme280.temperature
humidity = bme280.humidity
co2.append(sgp30.eCO2)
co.append(sgp30.TVOC)
pm.append(0) # Need to add a particulate matter sensor
uv.append(veml6075.uv_index)
light.append(veml6075.lux)
sound.append(0) # Need to add a sound level sensor

# Add data to training data

X.append([temperature, humidity])
y.append(co2[-1])

# Add data to arrays

temperatures.append(temperature)
humidities.append(humidity)
times.append(time.monotonic())

# Train model on training data

if len(X) >= 100:
model.fit(X, y)

# Send data over MIDI

midi.send_short(0xB0, 0x07, int(temperature * 10))
midi.send_short(0xB0, 0x27, int(humidity * 10))
midi.send_short(0xB0, 0x37, co2[-1])
midi.send_short(0xB0, 0x47, co[-1])
midi.send_short(0xB0, 0x57, pm[-1])
midi.send_short(0xB0, 0x67, uv[-1])
midi.send_short(0xB0, 0x77, light[-1])
midi.send_short(0xB0, 0x87, sound[-1])

# Plot data
plt.clf()
plt.plot(times, temperatures, label="Temperature")
plt.plot(times, humidities, label="Humidity")
plt.plot(times, co2, label="CO2")
plt.plot(times, co, label="CO")
plt.plot(times, pm, label="PM")
plt.plot(times, uv, label="UV")
plt.plot(times, light, label="Light")
plt.plot(times, sound, label="Sound")
plt.legend()
plt.show(block=False)
plt.
PROJECT HURDLES

 During the project execution, we encountered

several challenges. Key difficulties included
coordinating acrossdifferent time zones, which
complicated communicationand delayed
decision-making.
 Additionally, integrating various technologies and
tools ledto unexpected compatibility issues,
requiring more troubleshooting than anticipated.
Resource limitations, bothin terms of budget and
skilled personnel, further strained the project
timeline.
 Despite these obstacles, adaptive planning
and teamcollaboration were crucial in
overcoming them.
 OUTPUT
This also provides a pattern output ,.
 Conclusion and Future Scope

The project focuses on analyzing and evaluating data obtained from

environmental monitoring to enhance the accuracy and reliability of predictions
regarding environmental conditions. By leveraging advanced data analytics and
AI models, the project aims to provide a comprehensive understanding of various
environmental parameters, enabling better decision-making and proactive
measures to mitigate potential environmental issues.

The AI model developed in this project has demonstrated significant

potential in monitoring and predicting environmental changes with high accuracy.
The integration of diverse data sources and sophisticated analytical techniques
has paved the way for a more nuanced understanding of environmental dynamics.
This approach not only improves the immediate monitoring capabilities but also
contributes to the long-term sustainability of environmental resources.

In the future, the scope of this project will expand to include training the
AI model to analyze a broader spectrum of data types and environmental
anomalies. This will involve incorporating data from various sources, such as
satellite imagery, sensor networks, and historical climate records, to create a more
robust and versatile model. By doing so, the AI model will be able to detect and
predict a wider range of environmental abnormalities, including extreme weather
events, pollution levels, deforestation rates, and biodiversity changes.

Additionally, the future iterations of the AI model will focus on enhancing

its predictive capabilities through the use of advanced machine learning
techniques, such as deep learning and reinforcement learning. These techniques
will enable the model to learn from past data and adapt to new patterns and trends,
thereby improving its accuracy and reliability over time.

Furthermore, the project will explore the integration of real-time data

streams to provide up-to-the-minute monitoring and alerts. This real-time
capability will be crucial for responding promptly to emerging environmental
threats and taking preventive actions to minimize their impact.

Overall, the future scope of this project envisions a comprehensive and

adaptive environmental monitoring system powered by AI. This system will not
only provide accurate predictions and assessments but also support the
development of sustainable strategies to protect and preserve the environment for
future generations. Through continuous refinement and expansion, the project
aims to contribute significantly to the field of environmental science and to the
global efforts in combating environmental challenges.