Predicting the nutrition deficiency of fresh pear leaves with a miniature near-infrared spectrometer in the laboratory

Nutrient deficiencies often occur during the growth of pear trees; therefore, rapid, cost-effective monitoring of the nutritional deficiency status of pear leaves is of great value for effective cultivation management. The nitrogen, phosphorus and potassium contents of nutrient-deficient pear leaf samples were analysed with a handheld miniature near-infrared (NIR) spectrometer operating at a reflectance spectrum of 900–1700 nm. Combined with different pre-treatment and feature extraction methods, 42 recognition models were established by random forest (RF), support vector machine (SVM), gradient boosting decision tree (GBDT) and extreme gradient boosting (XGBoost). Finally, the best accuracy and F1-score of the SVM with the testing dataset, with standard normal variate (SNV) pre-processing and genetic algorithm (GA) feature extraction, were 82.06% and 80.25%, respectively. The proposed method using a miniature NIR spectrometer can quickly predict the nutrient deficiency status of pear leaves during the cultivation period.