A Dataset for Nitrate Test-Strip Analysis via Smartphone Camera for Agricultural Soil Testing

Minimizing the usage of fertilizer in agriculture requires knowledge of the nitrogen already available in the soil. Traditional methods for measuring soil nitrate levels involve sending samples to laboratories, which can be time-consuming and costly. To address this, nitrate test-strips provide a quicker alternative but often suffer from inaccuracies when compared visually against reference scales. This white paper presents a novel approach leveraging smartphone cameras to analyze nitrate test-strips, offering a more accessible and precise solution for farmers. We introduce a comprehensive dataset of images capturing test-strips under various conditions, alongside reference colors. The dataset includes images taken with different smartphones, both indoors under controlled lighting and outdoors in natural light. Additionally, we propose a benchmark algorithm for estimating nitrate levels from these images, demonstrating its potential for improving measurement accuracy. This dataset and methodology aim to facilitate the development of advanced nitrate analysis applications, ultimately simplifying the process for farmers and enhancing nitrogen management practices