Font Size: a A A

Study On Biochemical Parmeters And Disease Of Gannan Navel Orange Leaves By Spectroscopy

Posted on:2016-09-08Degree:MasterType:Thesis
Country:ChinaCandidate:Y H ZhouFull Text:PDF
GTID:2283330452468859Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
The study on navel orange leaves as the research object. There are three biochemicalparameters, leaevs nitrogen content and leaevs water content and leaevs chlorophyll content,predicted by different models, compared with the result of different models, to realize thedetection of three parameters at the same time. The main results of this study are shown asfollows:1With near infrared spectroscopy on navel orange leaves, in order to realize thedetection of leaevs biochemical content, it is compared the prediction effect of PLS model,PCR model and MLR. The results show that the best prediction model of three kinds of navelorange leaves biochemical parameter index was GA-PLS model, the correlation coefficientswere0.844,0.969,0.959, the root mean square error of prediction of were0.267,0.009,2.847.2It was set up a portable spectroscopic system device by the basic of spectrometerQE65000as the core device.Compared the prediction effect of PLS model and PCR model,the results show that the best prediction model of three kinds of navel orange leavesbiochemical parameter index was GA-PLS model, the correlation coefficients were0.712,0.916and0.729, the root mean square error of prediction of was0.360,0.017,2.676.3With near infrared spectroscopy on navel orange leaves, it was detected three kinds ofnavel orange leaves disease types, healthy leaves, leaves, leaves ulcer anthrax, classification.Comparison of chemometrics common discriminant method of two kinds of SIMCA methodand PLS-DA method, the study results were involved:the accuracy rate of SIMCA modelidentification of health class, ulcer type and anthrax class were85.5%,82%and71%, theerror rate is respectively14.5%,18%and29%using SIMCA differential method.Identifyingthe PLS-DA model of the three categories of accurate own types of other categories were100%, and the three recognition error rate is0%, using the PLS-DA method to the sample.
Keywords/Search Tags:navel orange leaves, near infrared, portable, nitrogen, chlorophyll, moisture, anthracnose, ulcer disease
PDF Full Text Request
Related items