Farmaid: Autonomous Robot for Precise Plant Disease Detection & Marking

Farmaid is a compact, battery‑powered robot that autonomously traverses greenhouse environments, leveraging advanced computer‑vision models to identify diseased plants and mark them with a non‑toxic, environmentally safe dye. By replacing manual inspections, it drastically reduces labor time and eliminates inconsistencies that arise from disparate mobile devices used by farm workers.

Inspired by the research of PlantVillage and IITA, the team adapted the open‑source DonkeyCar platform to create a robot that moves without damaging crops or soil. The system uses dual Raspberry Pi cameras—one focused on the foliage for disease classification, the other directed forward for obstacle‑free navigation. Neural‑network inference runs locally on the Pis, enabling real‑time decision making without cloud dependence.

Key benefits include:

• Uniform, repeatable disease marking across entire farms.
• Elimination of manual phone‑based tagging, cutting inspection time by up to 70%.
• Low operational footprint, allowing deployment on multiple farms without costly equipment swaps.

Core engineering challenges were:

• Designing a chassis small enough to avoid crushing tender seedlings.
• Ensuring smooth navigation on varied greenhouse substrates without causing soil compaction.
• Implementing a safe, reversible marking mechanism that uses biodegradable dyes.
• Assembling a reliable test environment and obtaining a labeled dataset for validation.

The team—members of the Detroit and Ann Arbor Autonomous Vehicle Meetup groups—brought together expertise in robotics, computer vision, and mechanical design. With no budget and each member balancing full‑time commitments, the group established a robust workflow: weekly video conferences, shared GitHub repositories, and a clear milestone calendar. Despite initial setbacks with prototype chassis and uneven greenhouse terrain, iterative refinements yielded a robust, metal‑reinforced frame that supports high‑current motors and lightweight wheels, preventing soil damage and maintaining agility.

Camera Pole and Illumination

A lightweight carbon‑fiber pole, sourced from a garage sale, was 3D‑printed with dual camera mounts. A 1.2 V solar panel supplies illumination, while a custom 12 V RGB LED system—built from repurposed speaker lights and a pill bottle diffuser—provides real‑time status indicators. The Arduino controller drives the RGB lights via a relay, enabling color coding for navigation status and disease markers.

Overall, the Farmaid robot represents a scalable, low‑cost solution for modern precision agriculture, demonstrating that volunteer‑driven, open‑source hardware can deliver real, measurable impact on crop health monitoring.