Early detection of red palm weevil using distributed optical sensor

Islam Ashry,Mansour Al-Bagshi,Salman Al-Brahim,Yuan Mao,Yousef Al-Fehaid,Boon S Ooi,Tien Khee Ng,Abdulmoneim Al-Shawaf

doi:10.1038/s41598-020-60171-7

Islam Ashry, Mansour Al-Bagshi + Show 6 more

Open Access

https://doi.org/10.1038/s41598-020-60171-7

Copy DOI

Abstract

Red palm weevil (RPW) poses a serious threat to the cultivation of date palms. It is considered to be the most destructive epidemic pest of palms, responsible for massive economic losses worldwide. Curative methods for RPW are not difficult to apply; however, the early detection of the pest remains a great challenge. Although several detection techniques have been implemented for the early detection of RPW, none of these methods have been proven to be reliable. Here, we use an optical-fiber-distributed acoustic sensor (DAS) as a paradigm shift technology for the early detection of RPW. Our sensitive sensor shows a detection of feeding sound produced by larvae as young as 12 days, in an infested tree. In comparison with existing, commonly-used technologies, this novel sensing technique represents a cost-effective and non-invasive alternative that could provide 24-7, real-time monitoring of 1,000 palm trees or even more. It could also monitor the temperature, an essential feature to control farm fires, another major problem for the cultivation of palm trees around the world.

Highlights

Based on the results obtained with the microphone, we applied a band-pass filter of [200–800 Hz] range to the time-domain signals collected by the optical fiber DAS in order to discard most of the low-frequency environmental noise, such as that from tree swinging[20] and environmental thermal fluctuations[21], as well as the high-frequency electronic noise of our system
In some rare cases (Fig. 4j), the moving average of the healthy tree might include some ups and downs, and the corresponding signal-to-noise ratio (SNR) value exceeds the decision threshold to produce a false alarm of infestation
We developed an acoustic-sensing hardware and algorithm for monitoring and distinguishing infested from healthy trees, using an optical fiber DAS of 10 m spatial resolution and ~1.1 km fiber length

Summary

Introduction

Recognizing the sound spectrum’s signature of the larvae will facilitate discarding noises when using the optical fiber DAS to locate infested trees. It is worth mentioning that both the larvae sound and background noise were recorded exactly under the same setting parameters for the microphone; their signal strengths in the time and frequency domains shown in Fig. 2b–g are comparable.

Results

Conclusion