Field Crops Research 177 (2015) 148–160 Contents lists available at ScienceDirect Field Crops Research jou rn al hom ep age: www.elsevier.com/locate/fcr Comparing the performance of active and passive reflectance sensors to assess the normalized relative canopy temperature and grain yield of drought-stressed barley cultivars Salah Elsayed a,b,∗ , Pablo Rischbeck a , Urs Schmidhalter a a Department of Plant Sciences, Technische Universität München, Emil-Ramann-Str. 2, D-85350 Freising, Germany b Evaluation of Natural Resources Department, Environmental Studies and Research Institute, Sadat City University, Egypt a r t i c l e i n f o Article history: Received 22 August 2014 Received in revised form 17 January 2015 Accepted 18 March 2015 Keywords: Canopy temperature Phenotyping Proximal sensing Spectral reflectance Thermal infrared a b s t r a c t High-throughput precision phenotyping, using spectral reflectance measurements, has the potential to provide more information for making better-informed management decisions at the canopy scale in real time. Active and passive spectral reflectance sensors are available for ground-based remote sensing; however, they have not been compared in their performance for assessing the normalized relative canopy temperature (NRCT) and the grain yield of drought-stressed plants. In this study, five spectral passive and active reflectance sensors, including a hyperspectral passive sensor (HPS), a hyperspectral active sensor (HAS), an active flash sensor (AFS), the Crop Circle (CC) and the GreenSeeker (GS), were tested to assess the NRCT and grain yield of barley cultivars under mild and severe drought stress in 2012 and 2013. Simple linear regression and partial least squares regression models were used for analysing the spectral data. The results showed that the spectral indices of all sensors were more closely related to NRCT and grain yield under mild drought stress (R 2 up to 0.70, significant correlation at p ≤ 0.001) than under severe drought stress (R 2 up to 0.53, significant correlation at p ≤ 0.001). Closer relationships between three normalized water indices (NWI-1, NWI-3 and NWI-4) and NRCT and grain yield were obtained for the hyperspectral passive sensor compared to the same indices of the hyperspectral active sensor and the active flash sensor under both mild and severe drought stress. Multivariate analysis using partial least square regression improved the relationship (R 2 up to 0.77, significant correlation at p ≤ 0.001) compared to the individual spectral indices and the single reflectance bands for each sensor. In conclusion, both the selection of adapted measurement devices and advanced statistical methods can improve assessments of NRCT and grain yield. © 2015 Elsevier B.V. All rights reserved. 1. Introduction Spectral and thermal high-throughput technologies have the potential to provide quick and precise measurements of important physiological and agronomic traits for crop phenotyping in breed- ing nurseries (Hatfield et al., 2008; Mistele and Schmidhalter, 2010; Gutierrez et al., 2010; Elsayed et al., 2011). Proximal remote sensing systems for phenotyping on the field scale can be based on passive and active reflectance sensing. Passive sensor systems depend on sunlight as a source of light in contrast to active sensors, which are equipped with light-emitting components that provide radiation ∗ Corresponding author at: Evaluation of Natural Resources Department, Environ- mental Studies and Research Institute, Sadat City University, Egypt. Tel.: +20 1090305222. E-mail address: sala7emam@yahoo.com (S. Elsayed). in specific waveband regions. Therefore, active sensors are more independent of changing irradiation conditions (Kipp et al., 2014). Whereas passive sensors allow hyperspectral information to be obtained in the visible and near-infrared range currently, com- mercially available active sensors such as the GreenSeeker (NTech Industries Inc., Ukiah, California), the Crop Circle ACS-470 ® (Hol- land Scientific Inc., Lincoln, Nebraska) and an active flash sensor (AFS) (tec5 AG, Oberursel, Germany) are limited to comparatively few wavelengths according to the number and type of light sources and possible user-selectable filters (Erdle et al., 2011; Kipp et al., 2014). Active hyperspectral sensing allows for the evaluation and testing of the relationship of not yet identified wavelength combi- nations to relevant crop traits, which are nearly independent of the ambient environmental conditions (Erdle et al., 2011; Rischbeck et al., 2014), and has, therefore, been further tested in this study to investigate the relationship to the normalized relative canopy temperature and grain yield of drought stressed barley cultivars. http://dx.doi.org/10.1016/j.fcr.2015.03.010 0378-4290/© 2015 Elsevier B.V. All rights reserved.