EverGreen AI Plant Care System

This project presents an AI-Powered Autonomous Smart Agriculture System designed to optimize plant care and prevent water waste. Built inside a structural cardboard enclosure, the system uses an integrated approach: a hardware layer managed by an Arduino Uno for soil substrate tracking, and an advanced software framework powered by Python and OpenCV. By transforming visual RGB camera streams into HSV color spaces, a computer vision model detects foliage anomalies (yellowing/chlorosis) and cross-references it with soil moisture percentages via a localized Digital Twin interface to initiate autonomous irrigation.

Here is the complete component list used to build this autonomous asset:

● Arduino Uno Microcontroller

● Capacitive Soil Moisture Sensor (Analog Input)

● 5V Single-Channel Relay Module (Isolated Actuation)

● Low-Voltage DC Water Submersible Pump & Vinyl Tubing

● External DC Battery Pack (Power Supply for Pump)

● Structural Cardboard and Dividers (Industrial Enclosure)

● Jumper Wires & Breadboard

Before physical implementation, the circuit logic was thoroughly simulated on Autodesk Tinkercad to validate pin mapping and avoid boundary errors.

● Soil Sensor: Connected to Analog Input Pin A0.

● Relay Control: Mapped to Digital Pin 11 (or Pin 4) to act as a binary switch (HIGH/LOW).

● LEDs: Integrated for real-time visual logic validation on the breadboard.

The system maintains multi-rail isolation, ensuring the microcontroller logic rails are safely separated from the pump's current requirements through the relay switch.

Since a standard Arduino board lacks the processing power for heavy image processing, a "Digital Twin" application was developed in Python.

Using OpenCV, the software processes real-time video feeds, converts frames into HSV color-space, and filters out chlorotic (yellow) foliage wavelengths. When the soil criteria drops below 35% or the vision model flags a yellow anomaly, an autonomous decision loop overrides the system to trigger irrigation, tracked on a localized English English HUD (On-Screen Display).

To meet safety requirements and maintain a professional aesthetic on the presentation bench, a custom cardboard enclosure was engineered. The chassis utilizes a zoned architecture, splitting the interior into a secure dry section for the sensitive electronic processing components (Arduino, breadboard, relay) and an isolated wet zone dedicated to the water reservoir and the organic parsley plant. This structure completely prevents short-circuit risks.