Detection Of Vegetative Parameters And Classification Of Cucumber Plants In Greenhouse Based On Chlorophyll Fluorescence Technique

Nowadays,the technologies for monitoring plant growth processes and detecting diseases of plants are very important for achieving fine management of greenhouse crops,which has great significance for improving crop quality and maintaining the safety of agricultural products.Chlorophyll fluorescence technology has the characteristics of high efficiency,high precision and non-destructive.It is mainly used to monitor the fluorescence intensity of plant photosynthesis,which can accurately reflect the photosynthetic status of plants and provide a reliable basis for the judgment of plant growth.In this paper,the chlorophyll fluorescence imaging system was designed based on the chlorophyll fluorescence phenomenon.On the one hand,this system was used to explore the changes of various nutritional parameters in the growth process of cucumber plants.On the other hand,the early identification and classification of cucumber brown spot and anthrax with similar symptoms were carried out with this system.The main research contents of this paper include:(1)According to the design,the hardware equipment platform of the fluorescence imaging system was composed of mini host,touch screen,CCD camera,programmable DC power supply,etc..And the software system was composed of Lab View and Matlab software to realize the control of the system and the processing of data.Combining the software and the hardware,a chlorophyll fluorescence imaging system for cucumber plants was developed.(2)The chlorophyll fluorescence imaging system was used to monitor the growth process of cucumber plants.After the image processing,chlorophyll fluorescence parameters(a total of 19)reflecting the different photosynthetic physiological conditions of plants were extracted and calculated,and the pseudo-color images of the photosynthetic efficiency of the entire plant was produced through the relationship between the fluorescence pictures and photosynthetic physiology of the plant,which reflects the plant photosynthetic efficiency distribution.(3)The linear regression model carried out with support vector machine was built between the average fluorescence parameters of leaves(a total of 19)with the nitrogen and chlorophyll contents of plants.The correlation coefficient of the model about chlorophyll fluorescence and nitrogen content was 0.839,and the correlation coefficient of the model about chlorophyll fluorescence and chlorophyll content was 0.890.According to the above model,the nitrogen and chlorophyll contents corresponding to each pixel of the image were predicted,then convert them to the corresponding grayscale.Finally,the pseudo-color images were used to make the nitrogen and chlorophyll contents of plant leaves correspond to different colors from high to low.This method provides a simple observation routine to monitor the nutrient content of cucumber plants.(4)The chlorophyll fluorescence imaging system was used to identify and distinguish cucumber leaf diseases.first,the fluorescence images of diseased cucumber plant were collected.Then,the Matlab was applied to image processing and to obtain the fluorescence value of the region of interest,and the SVM and XGBoost were used to identify and classify the diseases.Finally,the recognition rate of brown spot group and control group was 94.6%and the recognition rate for anthrax group and control group was 89.2%.To further distinguish different diseases,the classification accuracy of the control group,anthrax group and brown spot group was 88.6%.The classification accuracy of the control group,the early brown spot group,the late brown spot group,the early anthrax group and the late anthrax group was 78.9%.This provides an idea for the early detection of similar diseases of greenhouse plants.