At HackMerced, innovation meets opportunity as students from across the Central Valley gather for a weekend of creativity, coding, and collaboration. This annual hackathon challenges participants to build tech-driven solutions to real-world problems, fostering a spirit of innovation and problem-solving.
This year, VISTA F3 proudly sponsored the AgTech track, encouraging students to tackle pressing challenges in agriculture with technology. The results were nothing short of impressive — from AI-powered plant health monitors to smart livestock management systems, the winning projects showcased the incredible potential of merging technology with agriculture. Let’s take a closer look at the standout AgTech solutions that emerged from HackMerced and the brilliant minds behind them.
MooGuard revolutionizes livestock management by automating cattle tracking, health monitoring, and digital tagging. It reduces labor costs by keeping workloads constant, even as herd sizes grow, allowing farmers to focus on sick and at-risk animals. Using a Google Coral Dev Board with built-in tensor processing units, MooGuard runs parallel machine learning models to track cattle from a bird’s-eye view, processing data through Google Gemini for insights on physical health, movement patterns, and irregularities. The data is stored in a MongoDB backend and displayed on a secure online dashboard. Despite hardware challenges and limited cow data, the team trained their own model—bounding boxes and all. With more time and data, MooGuard aims to expand, covering more animals and deeper health insights.
PlantDoctor is an AI-powered plant disease diagnosis tool designed to support both small-scale gardeners and large agricultural producers. It uses a Convolutional Neural Network (CNN) to classify plant diseases, then passes the results to Google Gemini AI, which generates refined diagnoses and treatment plans. The user-friendly Streamlit interface presents detailed disease information and actionable recommendations to help farmers manage plant health. Built with a robust ML pipeline and backed by MongoDB Atlas, PlantDoctor efficiently processes images and structures AI insights. Despite challenges with fine-tuning the CNN and managing AI responses, the project successfully delivers automated, accurate plant care advice. Future plans include deploying the app, expanding the dataset, and integrating GCP for scalable data analysis.
PlantPulse is an AI-powered plant health tracking app designed for both seasoned gardeners and beginners. Using Google Gemini’s AI API, users can upload plant photos, and the app analyzes the images to provide a health score and tailored feedback on plant conditions. Beyond individual plant care, PlantPulse also tracks the overall health of users’ gardens. Built with HTML, CSS, JavaScript, Python Flask, and NodeJS, with Firebase handling user authentication and data storage, the app processes plant images through AI prompts to deliver actionable insights. Despite challenges with GitHub collaboration and frontend design, the team successfully integrated AI feedback and a functional backend. Future plans include adding data visualization, growth tracking, and automated reminders for plant care.