Department of Engineering, School of Engineering of Orihuela (EPSO), Miguel Hernández University (UMH), Carretera de Beniel, km 3.2, 03312 Orihuela, Alicante, Spain.
Department of Agricultural Biotechnology, College of Agricultural Sciences and Technology, Palestine Technical University-Kadoorie (PTUK), P.O. Box 7, Tulkarm, Palestine.
Biosensors (Basel). 2020 Nov 23;10(11):188. doi: 10.3390/bios10110188.
Lethal Bronzing Disease (LB) is a disease of palms caused by the 16SrIV-D phytoplasma. A low-cost electronic nose (eNose) prototype was trialed for its detection. It includes an array of eight Taguchi-type (MQ) sensors (MQ135, MQ2, MQ3, MQ4, MQ5, MQ9, MQ7, and MQ8) controlled by an Arduino NANO microcontroller, using heater voltages that vary sinusoidally over a 2.5 min cycle. Samples of uninfected, early symptomatic, moderate symptomatic, and late symptomatic infected palm leaves of the cabbage palm were processed and analyzed. MQ sensor responses were subjected to a 256 element discrete Fourier transform (DFT), and harmonic component amplitudes were reviewed by principal component analysis (PCA). The experiment was repeated three times, each showing clear evidence of differences in sensor responses between the samples of uninfected leaves and those in the early stages of infection. Within each experiment, four groups of responses were identified, demonstrating the ability of the unit to repeatedly distinguish healthy leaves from diseased ones; however, detection of the severity of infection has not been demonstrated. By selecting appropriate coefficients (here demonstrated with plots of MQ5 Cos1 vs. MQ8 Sin3), it should be possible to build a ruleset classifier to identify healthy and unhealthy samples.
致死性青铜病(LB)是一种由 16SrIV-D 植原体引起的棕榈科植物病害。我们尝试使用低成本的电子鼻(eNose)原型机对其进行检测。该电子鼻原型机由八个 Taguchi 型(MQ)传感器(MQ135、MQ2、MQ3、MQ4、MQ5、MQ9、MQ7 和 MQ8)组成,由 Arduino NANO 微控制器控制,加热电压在 2.5 分钟的周期内呈正弦变化。对未感染、早期症状、中度症状和晚期症状感染的菝葜叶样本进行了处理和分析。对 MQ 传感器的响应进行了 256 个元素离散傅里叶变换(DFT),并通过主成分分析(PCA)对谐波分量的幅度进行了审查。该实验重复了三次,每次实验都清楚地表明,未感染叶片样本与早期感染叶片样本之间的传感器响应存在差异。在每次实验中,都确定了四组响应,表明该装置能够重复区分健康叶片和患病叶片;然而,尚未证明其能够检测感染的严重程度。通过选择适当的系数(此处通过展示 MQ5 Cos1 与 MQ8 Sin3 的关系图来演示),应该可以构建一个规则集分类器来识别健康和不健康的样本。