在棉花生长周期中快速、无损伤、大批量地获取田间棉花的株高，并进行建模，为棉花株高的快速检测提供技术支撑。本研究基于大疆精灵4 RTK无人机搭载高清数码相机组成的低空遥感平台，获取棉田无人机超高分辨率的遥感影像(GSD=0.33 cm/pixel)，生成数字表面模型(DSM)，同时，人工测量了1 126株棉花花铃期的株高，并通过便携式Lite RTK对棉株顶端进行了厘米级高精度定位。经过对DSM和实测株高(H)之间的回归分析，发现二者存在显著的线性回归关系(R2=0.88，RMSE=8.49 cm)，表明基于DSM预测的棉花株高具有较高的精度。

Rapid, non-destructive, and high-throughput acquisition of cotton plant height in the field during its growth periods is critical for modeling and supports efficient detection technology. This study utilized a low-altitude remote sensing platform composed of a DJI Wizard 4 RTK UAV equipped with a high-definition digital camera to capture ultra-high-resolution remote sensing images of cotton fields (GSD=0.33 cm/pixel), from which a digital surface model(DSM) was generated. Simultaneously, the height of 1 126 cotton plants at the flowering and boll setting stage was measured and the tops of the cotton plants were positioned with centimetre-level high accuracy by portable Lite RTK. After regression analysis between DSM and measured plant height(H), it was found that there was a significant linear regression relationship between them(R2=0.88, RMSE=8.49 cm), indicating that cotton plant height predicted from DSM achieves high accuracy.

S127；S562

新疆维吾尔自治区重点研发计划项目(2022B02053-2)。