Cotton is a critical crop for more than 80 countries, with an annual output of 113 million bales. Yet viral, bacterial, and fungal diseases still take a heavy toll, accounting for 30-60% of yield losses. These diseases often go unnoticed until manual inspection is no longer effective, making early detection a key challenge for growers.

A new preprint on arXiv proposes a solution that pairs unmanned aerial vehicle (UAV) imagery with a custom YOLO deep learning framework. The approach is designed to improve the detection of cotton diseases from above, offering an alternative to traditional manual scouting.

The paper, listed as arXiv:2610.24182, is a preprint and has not yet been peer-reviewed. It does not provide specific accuracy figures in the abstract, but the authors position the framework as a step toward more precise, automated disease management in cotton farming.