Research @Missouri S&T

FarmVates

Research-driven agriculture technology at Missouri S&T

Year

2024 - 2025

Role

Research Intern

Team

Missouri S&T Research Lab

Duration

6 months

Overview

FarmVates is an AgriTech research project at Missouri University of Science & Technology focusing on using computer vision and machine-learning models to determine active pest infestations and suggesting an appropriate intervention.

The Challenge

Pest infestations are a major problem in agriculture, with many farmers being unaware of the specific bugs that are active in the crop fields. To resolve this, they often indiscriminately use pesticides against a wider range of bugs that are not the specific ones causing the problem, leading to wasted resources and harm to the environment.

The Solution

FarmVates solves this problem by using computer vision and machine-learning models to identifiy specific pests and suggest appropriate interventions that are tailored to the specific pest and crop. FarmVates is currently undergoing testing and validation by farmers in the Missouri region.

01

Deployable Device

The first component of the FarmVates system is a deployable, field-ready device designed to attract insects, trap them, and capture images for analysis. The device is 3D-printed and powered by a Raspberry Pi, which serves as the primary processor. I advised the system’s design and technical implementation of the hardware and software.

02

Image Processing

Once images are captured, they are sent over a cellular network to the backend for processing and storage. I helped design and implement this data pipeline using MQTT for device-to-server communication and a Python-based backend, work I completed during my first summer internship at FarmVates.

03

DevOps Pipeline

During my second summer at FarmVates, I focused on infrastructure and deployment. I helped build the CI/CD pipeline for the backend services and assisted in deploying and managing the Kubernetes cluster responsible for running the production backend.

Technologies

PythonTensorFlowRaspberry PiMQTTAWS S3KubernetesDockerArgoCDJenkins