利用无人机多光谱数据监测玉米对不同灌溉模式的响应差异

[Objectives]Monitoring crop moisture conditions in real-time is critical for adopting water-saving irrigation and reducing China’s present water scarcity. This study explores the feasibility of using drone multispectral image data for real-time monitoring of maize drought stress and compares the sensitivity of drone data and field-measured agronomic indicators to crop drought stress. [Methods]Two maize cultivars, ‘Fumin 985’ and ‘Zhengdan 958’, were used as test materials in the field experiment. Border irrigation, drip irrigation, and rainfed irrigation were employed as treatments. The pigment content and specific leaf area (SLA) of the most recently unfolded maize leaves were determined at 60, 70, 76, 84, 90, and 95 days after sowing. Similarly, a UAV equipped with multi-spectral cameras was used to collect near-ground remote sensing data to extract five vegetation indexes: the normalized difference vegetation index (NDVI), green normalized difference vegetation index (GNDVI), normalized difference red-edge index (NDRE), leaf chlorophyll index (LCI), and optimum soil adjust vegetation index (OSAVI). [Results]The changes in vegetation indexes were earlier than in the leaf pigment content and the leaf area index. At 70 days after sowing (tasseling period), there was a significant difference in the NDRE and LCI of the three treatments. NDVI, GNDVI, and OSAVI indexes differed between the irrigation and rain-fed modes. No difference was observed in the treatments' pigment content and specific leaf area index at the same stages. At 90 days following sowing, there was a considerable difference in pigment concentration among the treatments (filling period). Furthermore, correlation analysis revealed that the relationship between vegetation indexes and pigment content varied depending on the growth stage. The correlations between pigment content and two vegetation indices (NDRE and LCI) were higher for 80 days after sowing (flowering phase) than for NDVI, GNDVI, and OSAVI; for 90 days after sowing (filling period), the correlation between the five vegetation indices and pigment content was high. [Conclusions]According to this study, the utilization of UAV multi-spectral data to monitor maize drought proved better than some measured agronomic indicators. However, further research should be conducted on the best spectral indicators and periods for drought monitoring in various situations and environments.