Development of an Educational Application for Visual Soil Analysis Using Artificial Intelligence

Abstract

Objective – to present the development and application of an educational prototype based on artificial intelligence for visual soil analysis, focused on interactive learning and sustainability.
Methodology – the application was developed in Python (Streamlit), integrating modules for color, texture, structure, moisture, and root analysis, supported by Munsell visual references and soil science literature. The process included support from generative AI tools (ChatGPT, OpenAI) and was tested with Environmental Engineering students at UNESP – Sorocaba.

Originality/Relevance – the proposal introduces a bilingual interface (Portuguese and Spanish) combining technology and environmental education. It highlights how AI can empower non-programmers to create innovative sustainability tools.

Results – the tests revealed high student engagement and didactic potential. The app proved effective for guided interpretation of soil characteristics, though color variation under different lighting and camera conditions suggests future technical refinements.

Theoretical/Methodological Contributions – presents a replicable approach to developing educational tools with AI support, emphasizing human–machine collaboration in environmental learning.

Social and Environmental Contributions – promotes knowledge democratization and awareness about soil conservation among students and small farmers, contributing to SDGs 4, 11, and 15.