.png)
SOUMYA YADAV
Intro
During my Foreign winter internship at SUSTech, China, I had the opportunity to work in a Chinese government project for river water pollution early warning and traceability information monitoring system using Machine learning.
I worked on this project for six weeks at the lab.
Early warning and traceability information monitoring system
November - December 2019
About My Project
How I did It
This is my first experience in foreign research. With this project, I not only applied my knowledge in the field of Machine learning but also learnt the field of IoT very well. I worked in this project closely for about 45 days. I loved this project so much that many times I found myself working on Sundays.
The project can be broken down into three major parts:
Arduino and Sensors configuration
IoT and Backend Web development
Pattern recognition with OpenCV (ML)
This research was conducted by the school of environment, Southern university of science and technology, Shenzhen, China, which later got accepted as a research paper in the conference of 'AOGS'
Skills and Tools
-
MATLAB
-
Arduino
-
Sensors (MQ7, DHT22, BH1750 etc.)
-
Thingspeak (IoT based server)
-
PHP, Javascript, HTML, CSS
-
OpenCV
-
Web hosting
Arduino and Sensor configuration
-
For phase-I, I worked on the testing of different sensors like:
MQ7, MQ135, DHT22, DHT11, ORP meter, BH1750, MS1100
-
After calibrating them with the help of Arduino Uno we used NodeMCU (ESP8266) to send the readings to the cloud.
Given below are some of the images of the sensors I used for recording the data.
 ORP Meter |  Arduino Uno and ESP8266 |  ORP Meter with Arduino |
|---|---|---|
 DHT 11 |  MQ7. |  DHT22 Sensor |
IoT and Backend web development
-
For phase-II, I worked on storing the data to the clouds and accessing it. For reading data from the sensors I used Arduino Uno and from there I used ESP8266 for transmission of data to the Thingspeak servers using I2C protocol. I achieved it by using the circuit shown below.
-
Next for running a python script online, I used PHP and Javascript to run python script remotely from the system.
 |  |  |
|---|---|---|
 |  | |
Pattern recognition with OpenCV (ML)
-
For phase-III, I worked on integrating pattern recognition algorithm from OpenCV for matching the water sample pattern with the ground image.
 |  |  |
|---|---|---|
 |
Result
I presented the whole module to the Government officials with the help of my Professor Jiping Jiang. Further, this research has been accepted as a paper in the AOGS conference. Here is the copy of the acceptance letter.