近红外光谱技术快速无损诊断黄瓜植株氮、镁元素亏缺

Rapid and non-destructive diagnostics of nitrogen and magnesium deficiencies in cucumber plants by near-infrared spectroscopy

摘要: 为了快速无损诊断黄瓜植株氮、镁元素亏缺，以K-最近邻域法（k-nearest neighbors, KNN）模型识别率作为遗传算法中染色体的适应度值，采用遗传算法（genetic algorithm, GA）优选近红外光谱特征子区间，交互验证优化主成分因子数和K值后建立了特征子区间对应的GA-KNN诊断模型。通过控制营养液中氮和镁的含量，在温室大棚内以无土栽培方式培育氮、镁元素亏缺植株，并用近红外光谱仪采集缺氮、缺镁植株老叶的光谱信息。结果表明：原始光谱经过标准正交变换预处理和子区间总数优化后，均匀划分为

Abstract: To diagnose nitrogen and magnesium deficiencies in cucumber plants, genetic algorithm (GA) combined with K-nearest neighbors (KNN) were proposed to select characteristic intervals and establish diagnostic model based on near infrared spectroscopy. Near-in