Review On Multi Class Classification Of Coconut Pests And The Corresponding Natural Enemies (Predators And Parasitoids)

Abstract

Pests pose serious threats to the development of coconuts (Cocos nucifera L.) that impair both production volume and product quality. The sustainability of coconut farming relies on get- ting pest control right by performing exact pest identification as well as finding suitable natural enemies among diseases and parasitoids and predators. This review contains a detailed summary of multi-class identification techniques alongside an organized hierarchy connecting pests to their naturally occurring enemies. This paper examines recent developments in deep learning as well as machine learning through transfer learning and ensemble methods and vision trans- formers and CNNs and SVMs to accurately identify different pest species from visual images. The unbalanced distribution of pests within datasets creates difficulties because cost-sensitive learning suffers under sampling methodology. This work explores natural enemy ecological functions within pest management strategies for integrated pest control (IPC) planning purposes. The compilation of recent research findings through this review demonstrates the successful contribution of automated category systems to pest surveillance as well as control activities. The research presents a model to improve coconut pest management through biological control methods alongside advanced classification systems. The information acquired through this research will promote sustainable pest management practices which integrate natural enemies for efficient pest control systems to develop for improving coconut crop output and welfare.