Damage caused by frost on coffee plants can impact significantly in the reduction of crop quality and productivity. Remote sensing can be used to evaluate the damage caused by frost, providing precise and timely agricultural information to producers, assisting in decision making, and consequently minimizing production losses. In this context, this study aimed to evaluate the potential use of multispectral images obtained by unmanned aerial vehicle (UAV) to analyze and identify damage caused by frost in coffee plants in different climatic favorability zones. Visual evaluations of frost damage and chlorophyll content quantification were carried out in a commercial coffee plantation in Southern Minas Gerais, Brazil. The images were obtained from a multispectral camera coupled to a UAV with rotating wings. The results obtained demonstrated that the vegetation indices had a strong relationship and high accuracy with the frost damage. Among the indices studied the normalized difference vegetation index (NDVI) was the one that had better performances (r = − 0.89, R² = 0.79, MAE = 10.87 e RMSE = 14.35). In a simple way, this study demonstrated that multispectral images, obtained from UAV, can provide a fast, continuous, and accessible method to identify and evaluate frost damage in coffee plants. This information is essential for the coffee producer for decision-making and adequate crop management.

霜冻对咖啡植物造成的损害会显着降低作物质量和生产力。遥感可用于评估霜冻造成的损害，为生产者提供准确及时的农业信息，协助决策，从而最大限度地减少生产损失。在此背景下，本研究旨在评估无人机 (UAV) 获得的多光谱图像在分析和识别不同气候适宜区咖啡植物霜冻造成的损害方面的潜在用途。在巴西米纳斯吉拉斯州南部的一个商业咖啡种植园中对霜冻损害和叶绿素含量定量进行了目视评估。这些图像是从多光谱相机中获得的，该相机与带有旋转机翼的无人机相连。得到的结果表明，植被指数与冻害有很强的相关性和较高的准确性。在所研究的指数中，归一化差异植被指数 (NDVI) 是性能更好的指数 (r = − 0.89, R²  = 0.79，MAE = 10.87 e RMSE = 14.35）。本研究以一种简单的方式证明，从无人机获得的多光谱图像可以提供一种快速、连续且易于访问的方法来识别和评估咖啡植物的霜冻损害。这些信息对于咖啡生产商的决策和适当的作物管理至关重要。