论文部分内容阅读
探讨番茄叶片在波长350~2500nm范围的光谱信息,分析不同施磷水平下番茄叶光谱信息的变化规律,建立基于水培番茄叶片光谱信息的施磷水平鉴别模型。研究结果表明:在可见光波段420~690nm,高磷和低磷水平番茄叶片的光谱反射率均高于正常养分的光谱反射率;在近红外波段780~1383nm和1469~1858nm,高磷和低磷水平番茄叶片的光谱反射率均低于正常养分的光谱反射率;在近红外波段1413~1465nm和1929~2500nm,叶片的光谱反射率随着施磷量的增加而增加。由BP神经网络、主成分回归和支持向量机建立的模型,都能很好地鉴别水培番茄的施磷水平,其中主成分回归模型潜力更大,可满足实际应用。
The spectral information of tomato leaves in the range of 350-2500 nm was explored. The changes of spectral information of tomato leaves under different phosphorus levels were analyzed. The phosphorus level identification model based on the spectral information of hydroponic tomato leaves was established. The results showed that the spectral reflectance of tomato leaves at high and low phosphorus levels was higher than that of normal nutrients at the wavelengths of 420-690 nm in the visible light range. The spectral reflectance of tomato was higher at 780 ~ 1383nm and 1469 ~ 1858nm in the near infrared band, The spectral reflectance of horizontal tomato leaves was lower than that of normal nutrients. The spectral reflectance of leaves increased with the increase of phosphorus application in the near infrared range from 1413 to 1465 nm and from 1929 to 2500 nm. The model established by BP neural network, principal component regression and support vector machine can well identify the phosphorus application level of hydroponic tomato. The principal component regression model has more potential and can meet the practical application.