Oasis: A computer vision approach to self watering system for green air purifier

Tan, Wei Jun (2025) Oasis: A computer vision approach to self watering system for green air purifier. Final Year Project, UTAR.

Preview

PDF Download (7Mb) | Preview

Abstract

This project presents an integrated smart system that enhances indoor plant care by combining real-time object detection with automated environmental monitoring and control. At its core, the system leverages the YOLOv10 deep learning model, renowned for its high accuracy in real-time object detection, to identify and classify various indoor plant species from continuous video feeds captured by a webcam. This facilitates real-time visual monitoring and enables a responsive, automated plant care process. To extend functionality, the project incorporates a self-sustaining air purification system that utilizes natural plants as biofilters. Recognized plants are analysed for their suitability in air purification, with the system adjusting care accordingly. The integration of environmental and moisture sensors allows for dynamic adjustments in watering schedules to maintain optimal plant health. A key component of the system is the use of Node-RED as a middleware platform to facilitate communication between the plant recognition module and an Arduino-based automated watering system. Detected plants are mapped to corresponding moisture sensors, and each plant receives water based on its specific moisture requirements. This precise matching ensures resource efficiency and tailored care, reducing human intervention while promoting sustainable indoor plant management. Area of Study: Internet of Things, Computer Vision Keywords: Data Collection in IoT, Monitoring Application, User-friendly Application, IoT Security, Smart Irrigation, YOLOv10, Real-Time Plant Detection, Node-RED Integration, Arduino-Based Automation, Environmental Sensors.

Actions (login required)