MONITORING INVASIVE INSECTS USING ARTIFICIAL INTELLIGENCE

The Asian citrus psyllid (ACP), Diaphorina citri Kuwayama (Hemiptera: Liviidae), is a significant pest for citrus growers in the U.S., due to its ability of transmitting the bacterium responsible for huanglongbing (HLB) disease among citrus trees. The traditional ACP monitoring and controlling methods are time-consuming and labor-intensive. Developing innovative methods to detect ACP and design intelligent traps could greatly enhance management of psyllid infestations and HLB transmission. Monitoring ACP calls between males and females provides an opportunity to detect ACP presence, lure males into traps, or disrupt communication by introducing synthetic calls that interfere with the male signals. This paper presents an artificial-intelligence-based method to detect and classify ACP calls leveraging the multi-resolution feature of the continuous wavelet transform (CWT) technique and the feature extraction and classification capability of deep learning neural networks (DLNN). This method successfully identifies calls with a high accuracy rate exceeding 93%.