The fall armyworm (Spodoptera frugiperda) significantly impacts agriculture and the economy. Traditional management methods are often ineffective, as this pest grows rapidly and spreads widely. By the time farmers notice signs of crop damage, extensive harm may have already occurred. There is an urgent need for rapid and accurate detection technologies that are environmentally friendly to enhance agricultural sustainability. Thus, our aim was not to identify any individual VOC compounds; instead, we obtained readings from sensors representing the entire VOC composition released by plants when damaged by these pests. This study developed a real-time monitoring system for fall armyworm outbreaks, with alerts delivered via a smartphone application. We detected VOCs in various types of corn, including maize, sweet corn, and purple corn, under three conditions: healthy controls, infested by fall armyworms, and mechanically damaged. Results show that VOC responses from fall armyworm-infested plants significantly increased, while mechanically damaged plants spiked quickly but then decreased rapidly compared to healthy plants. The Linear SVM model achieved over 93.5% accuracy and an F1-score between 0.88 and 0.93. The system can process and classify corn plant statuses in real-time. This study demonstrates that VOCs may serve as biomarkers for detecting Spodoptera frugiperda. Findings support the development of an early warning system using e-nose technology and SVM algorithms for monitoring and managing pest infestations in their initial stages.

Keywords

Fall armyworm, Volatile organic compounds (VOCs), Real-time monitoring, E-nose, Pest detection