Study on the determination of pest attack time of tea plant by gas sensor

Abstract

Pests cause much loss in tea (Camellia sinensis) production, but there is no appropriate method to detect them. In this work, gas sensors were employed to detect the attack time of tea plants having been attacked by Ectropis obliqua. The volatiles emitted by tea plants attacked by E. obliqua change during different periods of 1 day, and so detection results of the gas sensors are influenced by the detection time point. However, none of the previous studies about pest detection considered this time point. In this study, we determined the pest attack time of tea plants considering the detection time point. The classification performances of the gas sensors based on various detection time points were compared and the best one was determined. Besides, an extreme learning machine was employed for qualitative classification and quantitative regression analysis of tea plants with different pest attack times at the best detection time point. The results showed that the best detection time point of the gas sensors for tea plants was 12 noon, and the extreme learning machine for classification and prediction provided good results, which indicated the feasibility of the gas sensors for determining the pest attack time of tea plants.

Plant phenotyping relevance

茶樹が害虫被害を受けた状態と被害時期をガスセンサーで推定する手法が研究の中心であり、検出時点の比較と機械学習による分類・回帰も評価しているため、植物状態のセンサーフェノタイピングに該当する。

Code and data availability

公開状態または取得可能な本文経路を確認できませんでした。

No evidence-backed public reproduction asset is currently recorded.