[Measurement of Nitrogen Content in Lettuce Canopy Using Spectroscopy Combined with BiPLS-GA-SPA and ELM].

Nitrogen fertilizer is necessary to improve yield and quality of lettuce. Spectroscopy is one of the most effective techniques used to detect crop nitrogen content. In this study, canopy reflectance spectra were acquired under five levels of nitrogen, and then were Savitzky-Golay smoothed, the first-order derivative spectra were calculated from the smoothed spectra to eliminate noise effects. Backward interval partial least squares (BiPLS), genetic algorithm (GA) and successive projections algorithm (SPA) were combined to select the efficient wavelengths. The number of variables was decreased from 2,151 to 8. The optimal intervals or variables were used to build multivariable linear regression (MLR) model, radial basis function neural network (RBFNN) models and extreme learning machine (ELM) models. This work proved that the results of BiPLS-GA-SPA-ELM model was superior to others with RMSEC was 0.241 6%, Rc was 0.934 6, RMSEP was 0.284 2% and Rp was 0.921 8. Our research results may provide a foundation for nutrition regulation and developing instrument.

[Study on the polarized reflectance characteristics of single greenhouse tomato nutrient deficiency leaves].

In order to improve accuracy of quantitative analysis model for the greenhouse tomato nitrogen, phosphorus and potassium nutrient stress, and explore the advantages of polarization non-destructive detection in single-leaf plants scale, polarized reflectance characteristics of greenhouse nutrient deficiency tomato leaves in different growing seasons and different deficiency extents were both examined via means of polarized reflectance spectroscopy system, which was self-developed by the research group. The main factors with effects on the polarized reflectance characteristics of tomato leaves were discussed, such as incident zenith angle, azimuth angle, detection zenith angle, light source polarizer degree, and detector polarizer degree. Experiments were carried out to verify the optimum level of above five parameters by means of range analysis of orthogonal experiments, through that way we can know the best angle combination of five parameters. Based on the above analysis, the angle combination and sorting of detecting tomato nutrients deficiency leaves via means of polarization spectroscopy system were obtained as follows: incident zenith angle 60 degrees, light source polarizer degree 0 degrees, detection zenith angle 45 degrees, detector polarizer degree 45 degrees and azimuth angle 180 degrees. At the same time, both the spectra of nitrogen, phosphorus and potassium deficiency leaves in different growth stages and different deficiency extent leaves were compared with each other. Results show that there is a positive correlation between the greenhouse nutrient deficiency tomato leaves growth cycle and tomato leaves polarized reflectance spectra. Nutrient excess or nutrient deficiency can both lead to polarized reflectance decline and polarized reflectance decline extent of greenhouse tomato leaves is more obvious during the fruiting and harvest period. This paper has a certain theoretical and practical significance in the research on nutrition rapid detection on the plant single leaf scale by means of polarized reflectance spectrum.

[Study on the polarized reflectance-hyperspectral information fusion technology of tomato leaves nutrient diagnoses].

With 25%, 50%, 75%, 100% and 150%, five levels of, nitrogen (N), phosphorus (P) and potassium (K) nutrition stress samples cultivated in Venlo type greenhouse soilless cultivation mode as the research object, polarized reflectance spectra and hyperspectral images of different nutrient deficiency greenhouse tomato leaves were acquired by using polarized reflectance spectroscopy system developed by our own research group and hyperspectral imaging system respectively. The relationship between a certain number of changes in the bump and texture of non-smooth surface of the nutrient stress leaf and the level of polarization reflected radiation was clarified by scanning electron microscopy (SEM). On the one hand, the polarization spectrum was converted into the degree of polarization through Stokes equation, and the four polarization characteristics between the polarization spectroscopy and reference measurement values of N, P and K respectively were extracted. On the other hand, the four characteristic wavelengths of N, P, K hyperspectral image data were determined respectively through the principal component analysis, followed by eight hyperspectral texture features extracted corresponding to the four characteristic wavelengths through correlation analysis. Polarization characteristics and hyperspectral texture features combined with each characteristics of N, P, K were extracted. These 12 characteristic variables were normalized by maximum-minimum value method. N, P, K nutrient levels quantitative diagnostic models were established by SVR. Results of models are as follows: the correlation coefficient of nitrogen r = 0.961 8, root mean square error RMSE= 0.451; correlation coefficient of phosphorus r = 0.916 3, root mean square error RMSE = 0.620; correlation coefficient of potassium r = 0.940 6, root mean square error RMSE = 0.494. The results show that high precision tomato leaves nutrition prediction model could be built by using polarized reflectance spectroscopy combined with high spectral information fusion technology and achieve good diagnoses effect. It has a great significance for the improvement of model accuracy and the development of special instruments. The research provides a new idea for the rapid detection of tomato nutrient content.

[Identification of lettuce leaf nitrogen level based on adaboost and hyperspectrum].

In order to facilitate lettuce fertilization in an economically rational way, an intelligent method to identify lettuce leaf nitrogen levels was studied. Lettuce samples of different nitrogen levels were cultivated in greenhouse with soilless cultivation method. In a particular growth period, the lettuce samples in various nitrogen levels were collected, then the FieldSpech3 spectrometer was used to acquire the hyperspectral data of the cultivated lettuce leaves. As there were much noise and redundant information in original hyperspectral data, standard normal variate transformation (SNV) was used to reduce the noise of the original hyperspectral data in this paper, then the principal component waves were extracted by principal component analysis (PCA). While K nearest neighbor (KNN) and support vector machine (SVM) were used for classification studies on the processed hyperspectra data respectively, adaptive boosting (Adaboost) was introduced into the two classifiers as it could improve the classification performance of weak classifiers, then Adaboost-KNN and Adaboost-SVM, the two integrated classification algorithms, were proposed. At last, the four classification algorithms were used for classification and identification of the same test sample data respectively, with the results showing that the classification accuracies of KNN, SVM, Adaboost-KNN and Adaboost-SVM were high up to 74.68%, 87.34%, 100% and 100%, among which the classification accuracies of Adaboost-KNN and Adaboost-SVM proposed in this paper were both good, and the stability of Adaboost-SVM was the best. Therefore, Adaboost-SVM used as a modeling method is suitable for the identification of lettuce leaf nitrogen level based on hyperspectrum, and it can also be used for reference to identify the nutrient elements of other crops in nondestructive testing methods.

[Rapid diagnostics of early phosphorus deficiency in mini-cucumber plants under protected cultivation by near infrared spectroscopy].

The morphological symptom of phosphorus deficiency at early stage is similar to the appearance of leaf aging process in preliminary phase, so that visual diagnostics of phosphorus deficiency in mini-cucumber plants at early stage is practically impossible. Near infrared reflectance spectra contain information about differences in compositions of leaf tissues between phosphorus-deficient plants and healthy plants. In the present paper, near infrared reflectance spectroscopy was used to provide diagnostic information on phosphorus deficiency of mini-cucumber plants grown under non-soil conditions. Near infrared spectra was collected from 90 leaves of mini-cucumber plants. Raw cucumber spectra was preprocessed by SNV and divided into 27 intervals. The top 10 principal components (PCs) were extracted as the input of BP-ANN classifiers by principal component analysis (PCA) while the values of nutrient deficient were used as the output variables of BP-ANN and three layers BP-ANN discrimination model was built. The best experiment results were based on the top 3 principal components of No. 7 interval when the spectra was divided into 27 intervals and identification rates of the ANN model are 100% in both training set and the prediction set. The overall results show that NIR spectroscopy combined with BP-ANN can be efficiently utilized for rapid and early diagnostics of phosphorus deficiency in mini-cucumber plants.

[Research on distinguishing weed from crop using spectrum analysis technology].

Automatic detection of weeds is necessary for site--specific application of herbicides or precise physical weed control. Leaf reflectance is mainly determined by photosynthetic pigments, leaf structural properties and water content, so spectral reflectance characteristics can be used for weed discrimination. The spectral reflectance of cotton, rice and weeds was determined in the range from 350 to 2 500 nm using the Analytical Spectral Device Full Range FieldSpec Pro (ASD) in laboratory. The discrimination analysis was done using the statistical software package SAS. The characteristic wavelengths were selected by using STEPDISC procedure. With the selected characteristic wavelengths, discriminant models were developed using the DISCRIM procedure in SAS. For distinguishing spine-greens from cotton, three characteristic wavelengths, 385, 415, and 435 nm, were selected, and good classification performance (100% accuracy) was achieved. The combination of characteristic wavelengths 415 and 435 nm has the biggest contribution to discrimination model. For distinguishing barnyard-grass from rice, five characteristic wavelengths, 375, 465, 585, 705, and 1 035 nm, were selected, and also good classification performance (100% accuracy) was obtained. The transition point from yellow to orange wavelength (585 nm) and the wavelength 705 nm in the red edge contributed more to discrimination model.

[Determination of total nitrogen content in fresh tea leaf using visible-near infrared spectroscopy].

To monitor tea tree growth and nitrogen nutrition in tea leaves, visible-near infrared spectroscopy was used to determine total nitrogen content. One hundred eleven fresh tea leaves of different nitrogen levels were sampled according to different tea type, plant age, leaf age, leaf position and soil nutrients, which covered a wide range of nitrogen content. Visible-near infrared reflectance spectra were scanned under the sunlight with a portable spectroradiometer (ASD FieldSpec 3) in field. The software of NIRSA developed by Jiangsu University was used to establish the calibration models and prediction models, which included spectra data editing, preprocessing, sample analysis, spectrogram comparison, calibration model and prediction model, analysis reporting and system configuration Eighty six samples were used to establish the calibration model with the preprocessing of first/second-order derivative plus moving average filter and the algorithm of PLS regression, stepwise regression, principal component regression, PLS regression plus artificial neural network and so on The result shows that the PLS regression calibration model with 7 principal component factors after the preprocessing of first-order derivative plus moving average filter is the best and correspondingly the root mean square error of calibration is 0. 973. Twenty five unknown samples were used to establish the prediction model and the correlation coefficient between predicted values and real values is 0.8881, while the root mean square error of prediction is 0. 130 4 with the mean relative error of 4.339%. Therefore, visible-near infrared spectroscopy has a huge potential for the determination of total nitrogen content in fresh tea leaves in a rapid and nondestructive way. Consequently, the technique can be significant to monitoring the tea tree growth and fertilization management.