[Near-infrared spectroscopy quantitative analysis model based on inverse regression].

In the present paper, an inverse regression method is used in near infrared (NIR) spectroscopy analysis to reduce dimension of predictor at first, then estimate linear regression function using the new derived low dimensional data. A real data set of 103 corn samples was used for analysis with this new inverse regression method. Taking 103 corn samples as experiment materials, seventy samples were chosen randomly to establish predicting model, the remaining thirty-three corn samples were viewed as prediction set. The new derived model is used to the prediction set. The coefficient is 0.986 and the average relative error is 2.1% between the model predication results and Kjeldahl's value for the protein content, and the resulis of using partial least square regression are 0.978 and 2.5%, respectively. The results demonstrate that the inverse regression method is feasible and has good property in near-infrared spectroscopy quantitative analysis, and also provides a new idea for chemometrics quantitative analysis.

Utilization of thermal infrared image for inversion of winter wheat yield and biomass.

The present paper utilizes thermal infrared image for inversion of winter wheat yield and biomass with different technology of irrigation (drip irrigation, sprinkler irrigation, flood irrigation). It is the first time that thermal infrared image is used for predicting the winter wheat yield and biomass. The temperature of crop and background was measured by thermal infrared image. It is necessary to get the crop background separation index (CBSI(L), CBSI(H)), which can be used for distinguishing the crop value from the image. CBSI(L) and CBSI(H) (the temperature when the leaves are wet adequately; the temperature when the stomata of leaf is closed completely) are the threshold values. The temperature of crop ranged from CBSI(L) to CBSI(H). Then the ICWSI was calculated based on relevant theoretical method. The value of stomata leaf has strong negative correlation with ICWSI proving the reliable value of ICWSI. In order to construct the high accuracy simulation model, the samples were divided into two parts. One was used for constructing the simulation model, the other for checking the accuracy of the model. Such result of the model was concluded as: (1) As for the simulation model of soil moisture, the correlation coefficient (R2) is larger than 0.887 6, the average of relative error (Er) ranges from 13.33% to 16.88%; (2) As for the simulation model of winter wheat yield, drip irrigation (0.887 6, 16.89%, -0.12), sprinkler irrigation (0.970 0, 14.85%, - 0.12), flood irrigation (0.969 0, 18.87%, -0.18), with the values of R2, Er and CRM listed in the parentheses followed by the individual term. (3) As for winter wheat biomass, drip irrigation (0.980 0, 13.70%, -0.13), sprinkler irrigation (0.95, 13.15%, -0.14), flood irrigation (0.970 0, 14.48%, -0.13), and the values in the parentheses are demonstrated the same as above. Both the CRM and Er are shown to be very low values, which points to the accuracy and reliability of the model investigated. The accuracy of model is high and reliable. The results indicated that thermal infrared image can be used potentially for inversion of winter wheat yield and biomass.

[Study on spectral detection of green plant target].

Weeds grow scatteredly in fields, where many insentient objects exist, for example, withered grasses, dry twig and barriers. In order to improve the precision level of spraying, it is important to study green plant detecting technology. The present paper discussed detecting method of green plant by using spectral recognizing technology, because of the real-time feature of spectral recognition. By analyzing the reflectivity difference between each of the two sides of the "red edge" of the spectrum from plants and surrounding environment, green plant discriminat index (GPDI) is defined as the value which equals the reflectivity ratio at the wavelength of 850 nm divided by the reflectivity ratio at the wavelength of 650 nm. The original spectral data of green plants and the background were measured by using the handhold FieldSpec 3 Spectroradiometer manufactured by ASD Inc. in USA. The spectral data were processed to get the reflectivity of each measured objects and to work out the GPDI thereof as well. The classification model of green plant and its background was built up using decision tree method in order to obtain the threshold of GPDI to distinguish green plants and the background. The threshold of GPDI was chosen as 5.54. The detected object was recognized as green plant when it is GPDI>GPDITH, and vice versa. Through another test, the accuracy rate was verified which was 100% by using the threshold. The authors designed and developed the green plant detector based on single chip microcomputer (SCM) "AT89S51" and photodiode "OPT101" to realize detecting green plants from the background. After passing through two optical filters, the center wavelengths of which are 650 and 850 nm respectively, the reflected light from measured targets was detected by two photodiodes and converted into electrical signals. These analog signals were then converted to digital signals via an analog-to-digital converter (ADS7813) after being amplified by a signal amplifier (OP400). The converted digital signal of reflected light was eventually sent to the SCM (AT89S51) and was calculated and processed there. The processing results and the control signals were given out to actuate executive device to open or close the solenoid valve. The test results show that the level of detectivity of the designed detector was affected by the species, size, and density of weeds. The detectivity of broad-leaf species is higher than that of narrow-leaf species. Broad-leaf species are more easily detected than those narrow-leaf ones; the bigger the plants and the denser the leaves are, the higher the level of detectivity is.

[Study on the application of ridge regression to near-infrared spectroscopy quantitative analysis and optimum wavelength selection].

In the present paper, taking 66 wheat samples for testing materials, ridge regression technology in near-infrared (NIR) spectroscopy quantitative analysis was researched. The NIR-ridge regression model for determination of protein content was established by NIR spectral data of 44 wheat samples to predict the protein content of the other 22 samples. The average relative error was 0.015 18 between the predictive results and Kjeldahl's values (chemical analysis values). And the predictive results were compared with those values derived through partial least squares (PLS) method, showing that ridge regression method was deserved to be chosen for NIR spectroscopy quantitative analysis. Furthermore, in order to reduce the disturbance to predictive capacity of the quantitative analysis model resulting from irrelevant information, one effective way is to screen the wavelength information. In order to select the spectral information with more content information and stronger relativity with the composition or the nature of the samples to improve the model's predictive accuracy, ridge regression was used to select wavelength information in this paper. The NIR-ridge regression model was established with the spectral information at 4 wavelength points, which were selected from 1 297 wavelength points, to predict the protein content of the 22 samples. The average relative error was 0.013 7 and the correlation coefficient reached 0.981 7 between the predictive results and Kjeldahl's values. The results showed that ridge regression was able to screen the essential wavelength information from a large amount of spectral information. It not only can simplify the model and effectively reduce the disturbance resulting from collinearity information, but also has practical significance for designing special NIR analysis instrument for analyzing specific component in some special samples.

[Near-infrared spectrum quantitative analysis model based on principal components selected by elastic net].

Elastic net is an improvement of the least-squares method by introducing in L1 and L2 penalties, and it has the advantages of the variable selection. The quantitative analysis model build by Elastic net can improve the prediction accuracy. Using 89 wheat samples as the experiment material, the spectrum principal components of the samples were selected by Elastic net. The analysis model was established for the near-infrared spectrum and the wheat's protein content, and the feasibility of using Elastic net to establish the quantitative analysis model was confirmed. In experiment, the 89 wheat samples were randomly divided into two groups, with 60 samples being the model set and 29 samples being the prediction set. The 60 samples were used to build analysis model to predict the protein contents of the 29 samples, and correlation coefficient (R) of the predicted value and chemistry observed value was 0. 984 9, with the mean relative error being 2.48%. To further investigate the feasibility and stability of the model, the 89 samples were randomly selected five times, with 60 samples to be model set and 29 samples to be prediction set. The five groups of principal components which were selected by Elastic net for building model were basically consistent, and compared with the PCR and PLS method, the model prediction accuracies were all better than PCR and similar with PLS. In view of the fact that Elastic net can realize the variable selection and the model has good prediction, it was shown that Elastic net is suitable method for building chemometrics quantitative analysis model.

[Study on a method of selecting calibration samples in NIR spectral analysis].

In the present paper, a simple but novel method based on maximum linearly independent group was introduced into near-infrared (NIR) spectral analysis for selecting representative calibration samples. The experiment materials contained 2,652 tobacco powder samples, with 1,001 samples randomly selected as prediction set, and the others as representative sample candidate set from which calibration sample set was selected. The method of locating maximum linearly independent vectors was used to select representative samples from the spectral vectors of representative samples candidate set. The arithmetic was accomplished by function rref(X,q) in Matlab. The maximum linearly independent spectral vectors were treated as calibration samples set. When different calculating precision q was given, different amount of representative samples were acquired. The selected calibration sample set was used to build regression model to predict the total sugar of tobacco powder samples by PLS. The model was used to analyze 1001 samples in the prediction set. When selecting 32 representative samples, the model presented a good predictive veracity, whose predictive mean relative error was 3.6210%, and correlation coefficient was 0.9643. By paired-samples t-test, we found that the difference between the predicting result of model obtained by 32 samples and that obtained by 146 samples was not significant (alpha=0.05). Also, we compared the methods of randomly selecting calibration samples and maximum linearly independent selection by their predicting effects of models. In the experiment, correspondingly, six calibration sample sets were selected, one of which included 28 samples, while the others included 32, 41, 76, 146 and 163 samples respectively. The method of maximum linearly independent selecting samples turned out to be obviously better than that of randomly selecting. The result indicated that the proposed method can not only effectively enhance the cost-effectiveness of NIR spectral analysis by reducing the number of samples required for cockamamie and expensive chemical measurement, but also improve the analysis accuracy. In conclusion, this method can be applied to select representative samples in near-infrared spectral analysis.

[SVM-based spectral recognition of corn and weeds at seedling stage in fields].

A handheld FieldSpec 3 Spectroradiometer manufactured by ASD Incorporated Company in USA was used to measure the spectroscopic data of canopies of seedling corns, Dchinochloa crasgalli, and Echinochloa crusgalli weeds within the 350-2 500 nm wavelength range in the field. Each canopy was measured five times continuously. The five original spectroscopic data were averaged over the whole wavelength range in order to eliminate random noise. Then the averaged original data were converted into reflectance data, and the unsmooth parts of reflectance spectral curves with large noise were removed. The effective wavelength range for spectral data process was selected as 350-1 300 and 1 400-1 800 nm. Support vector machine (SVM) was chosen as a method of pattern recognition in this paper. SVM has the advantages of solving the problem of small sample size, being able to reach a global optimization, minimization of structure risk, and having higher generalization capability. Two classes of classifier SVM models were built up respectively using "linear", "polynomial", "RBF"(radial basis function), and "mlp (multilayer perception)" kernels. Comparison of different kernel functions for SVM shows that higher precision can be obtained by using "polynomial" kernel function with 3 orders. The accuracy can be above 80%, but the SV ratio is relatively low. On the basis of two-class classification model, taking use of voting procedure, a model based on one-against-one-algorithm multi-class classification SVM was set up. The accuracy reaches 80%. Although the recognition accuracy of the model based on SVM algorithm is not above 90%, the authors still think that the research on weeds recognition using spectrum technology combining SVM method discussed in this paper is tremendously significant. Because the data used in this study were measured over plant canopies outdoor in the field, the measurement is affected by illumination intensity, soil background, atmosphere temperature and instrument accuracy. This method proposes a kind of research ideology and application foundation for weeds recognition in the field.

(2S,8Z)-2-butyroxy-8-heptadecene: major component of the sex pheromone of chrysanthemum gall midge, Rhopalomyia longicauda.

The sex pheromone of the chrysanthemum gall midge, Rhopalomyia longicauda (Diptera: Cecidomyiidae), the most important insect pest in commercial plantations of chrysanthemum, Dendranthema morifolium (Ramat.) Tzvel., in China, was identified, synthesized, and field-tested. Volatile chemicals from virgin females and males were collected on Porapak in China and sent to the United Kingdom for analysis. Coupled gas chromatographic-electroantennographic detection (GC-EAG) analysis of volatile collections from females revealed two compounds that elicited responses from antennae of males. These compounds were not present in collections from males. The major EAG-active compound was identified as 2-butyroxy-8-heptadecene by gas chromatographic (GC) retention indices, mass spectra, in both electron impact and chemical ionization modes, hydrogenation, epoxidation, and derivatization with dimethyldisulfide. The lesser EAG-active compound was identified as the corresponding alcohol. The ratio of butyrate to alcohol in the collections was 1:0.26. Racemic (Z)-8-heptadecen-2-ol and the corresponding butyrate ester were synthesized from (Z)-7-hexadecenyl acetate, and the synthetic compounds found to have identical GC retention indices and mass spectra to those of the natural, female-specific components. Analysis of the volatile collections on an enantioselective cyclodextrin GC column showed the natural pheromone contained (2S,8Z)-2-butyroxy-8-heptadecene. Field tests showed that rubber septa containing racemic (Z)-2-butyroxy-8-heptadecene were attractive to R. longicauda males. The (naturally occurring) S-enantiomer was equally as attractive as the racemate, while the R-enantiomer was not attractive to males, and did not inhibit the activity of the S-enantiomer. The attractiveness of the butyrate was significantly reduced by the presence of even small amounts of the corresponding alcohol.

[Study on the application of supervised principal component regression procedure to near-infrared spectroscopy quantitative analysis].

The present paper introduces the principle of a new modeling method, called supervised principal component regression, with which the model of the near-infrared (NIR) spectroscopy quantitative analysis was established. Usually, there are many difficulties such as collinearity when establishing the quantitative analysis model for the high dimension of the spectral data. Using this new method, firstly according to some criterion, the wavelength information is selected in order to reduce the dimension of spectral data. Then the selected lower dimensional spectral data set is used to establish the principal component regression model. Taking sixty-six wheat samples as experiment materials, forty samples were chosen randomly to establish the predicting model, while the remaining twenty-sixth wheat samples were viewed as prediction set. In this example, 4 wavelengths, 4 632, 4 636, 5 994 and 5 997 cm(-1), were selected at first according to the coefficients between the response variable and each spectral data. Then two principal components of the spectral data at those four wavelengths were extracted to establish the principal component regression model. The model was used to the prediction set. The coefficient was 0.991 and the average relative error was 1.5% between the model predication results and Kjeldahl's value for the protein content. It is very important to select the most significant part of wavelength information from a large number of spectral data, not only because this procedure can alleviate the influence of collinearity in modeling, but also because it can be used to guide the design of special NIR analysis instrument for analyzing specific component in some samples.

[Target infrared detection in target spray].

Crops in agriculture and forestry are normally planted discretely. The chemical sprayed between crops would cause great waste and serious environment pollution. Therefore realization of the precision spray has great significance. This research discussed the method to realize automatic target detection using infrared detect technology. The infrared can avoid the interference of the visible light effectively and the response speed is very fast. Therefore it can be used to implement non-tough detection. Photoelectric detection systems based on infrared detect technology are normally stable, reliable, low cost, simple structure, and easy to be practically utilized. Therefore it is widely used in the on-line real time detection field. Its key point is to determinate the characteristic wavelength or wave band. The infrared lights emitted from the infrared light emitting diode were irradiated to the detected objects. The reflected infrared lights could be received by the photoelectric device. Then control signal was triggered and automatic target spray was realized. Code-division infrared detection circuit was used in the system. Modulated pulse infrared signals using different coding were used in different photodetector units in the built system so as to eliminate the light path interference between different detector units and other light signal interferences. Therefore the interference capacity of the system is high. The test results showed that the automatic target spray equipment set up in the study could detect crop targets automatically. The light wavelength used in the test is 850 nm. The detection range was tunable within 0.1-0.5 m. The least targets detectable distance was less than 0.3 m.