Monitoring and Managing Rice Pest Infestation through Hyperspectral Remote Sensing Technology under Field Conditions

Timely assessment of infestation symptoms in crops is critical for pest control and precision farming. The use of non-contact, highly-efficient, and affordable methods such as hyperspectral data for detecting and monitoring plant pests could greatly facilitate plant protection management. A field experiment was carried out during season of 2018 at Al-Qurien city, Sharkia Governorate, Egypt. The exploratory objectives were: (i) to establish a monitoring method for the damages caused by rice leaf folder (RLF), Cnaphalocrocis medinalis (Guenée), and yellow stem borer (YSB), Scirpophaga incertulas (Walker), of rice (Oryza sativa L.) variety (Giza177) based on hyperspectral data, and (ii) to manage the RLF and YSB of rice eco-friendly using Azadirachtin 0.15% EC. Spectral reflectances from rice at different stages in various bands were recorded using field portable hyperspectral radiometer (FPHR). The vegetation indices (VIs) were calculated and correlated with pest damage, sensitivity analysis of spectral bands, and red edge position to estimate the extent of damage caused by each pest. The results showed that spectral reflectance of rice hills damaged by studied pests was different from that of the undamaged hills. In the damaged hills, there was a decrease in near infrared (NIR) reflectance (760 to 900 nm) while the green (520 to 600 nm) and red reflectance (630 to 690 nm) increased compared to undamaged hills treated with Azadirachtin 0.15% EC. The mean percent reflectance values (of all days of observation) in the red region in RLF and YSB damaged hills were 2.41±0.86 and 9.31±3.96, respectively, while the treated hills with Azadirachtin 0.15% EC recorded 1.53±0.40 and 5.80±2.09. The percent reflectance values in NIR region were 36.37±7.93 (RLF) and 37.80±10.22 (YSB) in untreated hills while treated ones recorded 39.70±7.80 for RLF and 43.74±8.74 for YSB. The red reflectance had significant positive correlation with both pests while green and NIR reflectance recorded non-significant changes due to pest damage on all days of observation. On the basis of a comprehensive analysis of the hyperspectral data, significant spectral indices such as the normalized difference vegetation index (NDVI), simple ratio (SR), and green red vegetation index (GRVI) values were explored to provide an accurate and robust assessment of rice infestation by studied pests. Among the vegetation indices, SR had the highest sensitivity compared to NDVI and GRVI. Linear regression equations were adopted for studied crop ages for estimating RLF and YSB damage based on NDVI and SR values. The results demonstrated the operational applicability of ground-based hyperspectral measurements for diagnostic mapping of pest symptoms. Considering the efficacy and eco-friendly nature of Azadirachtin it could be considered as effective botanicals in successful management of the pests RLF and YSB of rice. It has a great potential in detecting early pest infestation for precision farming.