Early Warning Study On Powdery Mildew Of Reticulated Melon Based On Air-sensitive Sensing Technology Detection

Powdery mildew is a common disease in the production of muskmelon.After being infected by this pathogen,the surface or back of plants can be infected by the pathogen.Although the symptoms of powdery mildew infection can be observed with the naked eye,it cannot be detected timely based on the experience of the grower alone.This study utilizes gas sensitive sensor(electronic nose)technology to achieve early monitoring and early warning of powdery mildew in muskmelon by analyzing the occurrence,development,and physiological and biochemical changes of powdery mildew,providing technical support for scientific prevention and control of powdery mildew in muskmelon production.In this experiment,the melon was inoculated with two different concentrations of powdery mildew spore suspensions of powdery mildew in melon,and normal plants were used as control for the test under different treatment conditions using West Island honey reticulated melon as test material.This study was intended to be carried out in two aspects:(1)Analysis of changes in physiological indicators of melon leaves after powdery mildew fungus infection.(2)Early warning of powdery mildew disease in melon by electronic nose detection technique."Dynamic analysis of physiological and biochemical indicators in melon leaves inoculated with powdery mildew pathogen".The indexes of peroxidase(POD)activity,superoxide dismutase(SOD)activity,soluble protein content,soluble sugar content,and plasma membrane permeability in leaves of different treatments were measured and analyzed.It was found that these indexes did not significantly change on the first day after inoculation;On the the fourth day after inoculation,the activity index of POD and SOD in the leaves of the test group was significantly higher than that of the control group(P<0.05),while the content of soluble protein and soluble sugar in the leaves was increased,but the test group did not reach a significant level compared with the control group(P>0.05);After 7 days,the activities of POD and SOD in the leaves of the experimental group were slightly lower than those of the fourth day,but much higher than those of the control group.However,the content of soluble protein and soluble sugar in the leaves of different concentrations of treatment was relatively stable;In short,after being infected by pathogenic bacteria,the physiological metabolism of melon leaves has undergone significant changes,and the changing rules of their physiological and biochemical indicators are very complex.It is difficult to accurately determine them based on physiological and biochemical indicators alone.Moreover,due to the heavy workload of its measurement,it has brought great inconvenience to practice.Early warning study of melon powdery mildew based on electronic nose detection technology.In this paper,we conducted real-time monitoring of healthy melon and foliage at 1,4,7 and 10 days after inoculation with powdery mildew disease,and screened the best disease characteristic values(mean values)by random forest(RF),support vector machine(SVM)and neural network(BPNN)models and PCA and LDA techniques to comprehensively evaluate the key factors(mean values)in the process of disease development.After two linear model algorithms(PCA,LDA)were used for identification,it was found that the results were not very satisfactory,and even on the 7th day after inoculation,there was a large overlap between the leaf samples of the control and test groups in the identification results,which made it difficult to achieve accurate differentiation.However,the recognition results of the three nonlinear pattern algorithms(RF,SVM,and BPNN)used with the mean value of the electronic nose response data extracted as the characteristic quantity showed that the recognition accuracy of the nonlinear pattern algorithms was higher than that of the linear algorithms(PCA,LDA),and they also had greater advantages in terms of the time to achieve accurate recognition,among which,SVM had the best recognition effect from the 4th day after inoculation.After the 4th day after inoculation,the healthy leaves and leaves infected with powdery mildew fungus of reticulated melon could be identified more accurately with an accuracy of 97%,and the identification accuracy was higher thereafter.In summary,this paper proposes that by using electronic nose detection technology with the mean value as the characteristic quantity and combining with SVM pattern algorithm,it can achieve the accurate identification of healthy nettle melon leaves and those infected with powdery mildew fungus,which has the advantages of fast,non-destructive and efficient.This study provides a new idea for early warning detection of powdery mildew in reticulated melon,which is important for disease control in reticulated melon production.