Differenct Response Of Electronic Nose Before And After Rice Infested By Brown Rice Planthoppers

Brown Rice Plant-hopper（BRPH）（Nilaparvata lugens,Sogatella furcifera,and Laodelphax striatellus）is one of the most serious disasters of rice production in China.Timely and accurate acquisition of insect BRPH information is a key aspect of pest control.Electronic nose is a new method in the research stage.It has been studied in the detection and identification of BRPH,and many problems have not been explored.In this paper,rice is used as the main research object before and after being harmed by brown planthopper.The electronic nose（PEN 3）is used to study the differences of rice volatiles from the following three aspects.The main research work and achievements are as follows:（1）Differences in electronic nose response before and after germination of different rice varietiesUsing the electronic nose,the odor data of six different rice varieties,such as Huahang31,Mexiangzhan 2,TN1,Teyou 3813,Wuyou 1179 and Nipponbare NIP,were sampled in the rice,budding and seedling stages,and the maximum（Max）value,mean differential value（Kave）,stable value（Stable）and mean（Zmean）were extracted.,characteristic parameters of four electronic nasal response curves,respectively,using principal component analysis（PCA）,linear discriminant analysis（LDA）and support vector machine（SVM）analysis methods were used to study the difference of electronic nose response before and after of the budding above six different rice varieties.The test results show that the electronic nose sensors are different in different rice varieties.The response curves of different growth periods were significantly different（P=0.000）.The variance（ANOVA）was used to optimize the rice,the germination stage and the seedling stage.The average,average and average differential values were used as the optimal characteristic parameters.Load analysis is used to optimize the sensor array based on the above preferred characteristic parameters.As a result,sensors R8,R5,R2,R7,R4 and R10 contain most of the useful rice odor information,R5,R8,R2.R6,R4 and R10 contain most of the useful budding odor information.R9,R3,R4,R7,R8 and R6 contain most of the useful seedling odor information.Different varieties of rice and budding volatiles have certain The similarity of the main components of the volatiles in the seedling stage is quite different from that in the rice and germination stages.The recognition rate of PCA and LDA for different rice varieties was rice>budding period>seedling stage.The recognition accuracy of using SVM to identify rice,sprouting and seedling stage was 88.89%,70%and 57.14%,respectively.Different rice varieties were identified by different identification methods,and the recognition rate was rice>budding period>seedling stage.That is,from the beginning of rice development,the difference of gas volatiles in different varieties of rice was gradually reduced,indicating that the varieties were in the late stage of rice development.The impact on odor is small,which provides a theoretical basis for the monitoring of rice planthoppers in different rice varieties in the future.（2）Differences in electronic nose response between Brown Rice Plant Hoppers and Rice StemElectronic nose was used to sample electronic nose response data of rice stem,nymph of brown planthopper（O3IN）and nymph of brown planthopper（U3IN）.The results show that there are some similarities among rice stalk volatiles,O3IN and U3IN.The results of principal component analysis（PCA）showed that the three categories were overlapped,leading to incomplete pca,which also showed that rice stalk volatilites,O3IN and U3IN were similar to some extent.The overlap between rice stem and O3IN is greater than that between rice stem and U3IN,which further proves that the similarity between O3IN and rice stem is stronger than that between U3IN and rice stem.In addition,the first four principal components of the total contribution contains information most useful samples,the samples information can be used as an analysis of the characteristic value of next time,further reduce redundant information,improve the efficiency of computing KNN,PNN,SVM method to the training set of straw,O3IN,U3IN recognition accuracy is 100%,KNN,PNN,SVM method of validation set identification accuracy were 93.33%,93.33%and 90%respectively.However,the identification accuracy of KNN and PNN of rice stem and brown planthopper is 100%,and KNN and PNN are more suitable than SVM to solve the identification problem of rice stem and brown planthopper.（3）Difference of electronic nose response among different rice varieties after Brown PlanthopperUsing electronic nose(PEN（3）to 3813（in）resistance,excellent Wu You（resistance）in 1179,China airlines 31（susceptible）,mei xiang accounted for no.2（susceptible）,TN1（Gao Gan）,the NIP（Japan shine,Gao Gan）the resistance of six different rice varieties was BPH damage process of sampling.Max,K_ave,Stable and Z_mean of the sensors were extracted from the samples and analyzed by ANOVA.Variation of time series of each sensor.The results showed that there was no difference in the response of electronic nose during the infection of different resistant rice varieties by brown planthopper.The response time series curves of optimized sensors R1,R6,R3 and R8 were plotted to observe the response changes of different rice varieties to brown planthopper.The results showed that the aromatic components（selective volatiles of R1 and R3 sensors）first increased and then decreased,reached the maximum and minimum values at D4 and D17 respectively,methane（selective volatiles of R6 sensors）reached a small peak at D4 and a peak at D17 respectively,and ethanol（selective volatiles of R8 sensors）reached the minimum and maximum values at D4 and D17 respectively.