TomateRitmo

The problem

Diagnosing the health of a tomato crop from photos means combining a solid backend with a vision model.

My role

I built the Spring Boot backend that orchestrates the vision model and exposes the results.

Features

• A Spring Boot backend that orchestrates the vision model.
• A Python vision model with Flask and TensorFlow.
• Image classification with diagnosis, anomaly and confidence.
• An Angular interface.
• Service integration with Docker.

Architecture

The Spring Boot backend orchestrates a Python vision model (Flask and TensorFlow) that classifies each image and returns diagnosis, anomaly and confidence level. The interface is Angular and everything is integrated with Docker.

Challenges

Coordinating services across different languages while keeping response times reasonable.

Outcome

A working AgroTech platform, delivered as an academic project at UPC.