Water Quality Prediction and Monitoring System Using IOT Sensors and Machine Learning
K. Prashanth Yadav1, K. Kishore2, N. Manish Kumar3, S. Kishor4, S. Vamshi Krishna5. Dr. S .Srinivas6
1Student, BTech CSE(DS) 4th Year, Holy Mary Inst. Of Tech. And Science, Hyderabad, TG, India,
kotaprashanthyadav@gmail.com
2Student, BTech CSE(DS) 4th Year, Holy Mary Inst. Of Tech. And Science, Hyderabad, TG, India,
Konakanchikishore24@gmail.com
3Student, BTech CSE(DS) 4th Year, Holy Mary Inst. Of Tech. And Science, Hyderabad, TG, India,
narsayollamanish1175@gmail.com
4Student, BTech CSE(DS) 4th Year, Holy Mary Inst. Of Tech. And Science, Hyderabad, TG, India,
kishorsheela95@gmail.com
5Student, BTech CSE(DS) 4th Year, Holy Mary Inst. Of Tech. And Science, Hyderabad, TG, India
Vamshipatel326@gmail.com
6Assoc. prof, CSE(DS), Holy Mary Inst. Of Tech. And Science, Hyderabad, TG, India
prof.srinivas26@gmail.com
ABSTRACT
Small farmers usually have to guess when to water or harvest, mostly because professional sensors cost too much and are way too complicated. That’s where my project steps in. I built a Smart IoT Irrigation and Prediction System that’s affordable, tough enough for the field, and actually smart. Instead of filling a farm with expensive devices, I just used a single 30cm pipe, an ESP32, and an ultrasonic sensor to keep tabs on water levels all the time. Rural areas like Peddapally have another problem—spotty internet. So, I gave the system a backup plan. It saves every bit of data to an SD card, right there on the farm. If the Wi-Fi goes out for days, nothing gets lost. Once the internet’s back, everything syncs up automatically to Google Sheets. But the real magic? That’s in the system’s brain. I went way past basic calculations. I trained Deep Learning models—LSTM and 1D-CNN. With these two models working together, the system tells farmers what they really need: how many days until harvest, and how many tons they can expect. And I know most farmers don’t want to stare at a screen all day, so I added a bilingual voice interface. The system actually talks, giving updates in both English and Telugu. With 96% accuracy, this project proves world-class tech doesn’t have to come with a huge price tag.
