In subtropical coastal waters, the explosive growth of phytoplankton under favorable conditions can lead to water discolouration and massive fish kills. Manual field sampling and laboratory analysis of chlorophyll-a concentration (Chl-a) as an indicator to algal biomass, is resources intensive and time consuming, delaying responses to disastrous harmful algal blooms. Cloudy weather often precludes the use of satellite images for water quality and algal bloom monitoring. This study aims at developing an estimator algorithm for quantitative mapping of surface Chl-a for coastal waters, based on surface reflectance measurement from an Unmanned Aerial Vehicle (UAV) with a five-band multispectral camera. The surface reflectance is obtained from calibrated multispectral images which are radiometric-corrected against incoming solar radiation. It is found that Chl-a has an inverse correlation with the Normalized Green-Red Difference Index (NGRDI). A regression estimator model for Chl-a from NGRDI is developed, showing excellent performance for fish farms in coastal waters with different characteristics. The technology is demonstrated for mapping the spatial and temporal variation of Chl-a during an algal bloom, offering a useful complement to traditional field monitoring for fisheries management and emergency response.

在亚热带沿海水域，浮游植物在有利条件下的爆炸式生长会导致水体变色和大量鱼类死亡。作为藻类生物量指标的叶绿素a浓度 ( Chl- a ) 的人工现场采样和实验室分析是资源密集型和耗时的，延迟了对灾难性有害藻类大量繁殖的响应。多云天气通常会妨碍使用卫星图像进行水质和藻华监测。本研究旨在开发一种用于表面叶绿素定量映射的估计算法。用于沿海水域，基于使用五波段多光谱相机的无人驾驶飞行器 (UAV) 的表面反射率测量。表面反射率是从校准的多光谱图像中获得的，这些图像针对入射的太阳辐射进行辐射校正。发现叶绿素与归一化绿红差异指数（NGRDI）呈负相关。开发了来自 NGRDI 的 Chl- a 回归估计模型，显示了沿海水域养鱼场具有不同特征的出色表现。该技术被证明可用于绘制藻华期间叶绿素的空间和时间变化图，为渔业管理和应急响应的传统现场监测提供有用的补充。