mainlyfocusedonparticularlythesingleplantwhichisthe tomatoplant.Forclassificationpurposes,theCNNalgorithm
2Purva Gholap, Dept of Information Technology Engineering, Atharva College of Engineering
1.1 Proposed approach
1.2 Motivation
Aspertheinformation,thetraditionaltoolsandtechnology arenotverybeneficialbecauseitistime consumingandtake plenty of human effort. The plan behind our project is to develop an Android application with a simple GUI so that users with less experience can easily discover the disease fromtheleafandcangetadviceformanagingthedisease.
2. LITERATURE SURVEY
Kirti Kangane1, Purva Gholap2, Tejal Kadam3 , Prof. Anuradha Kapoor4
Abstract Entirehumangrowth isencircledandreliant on plant life but it is very troublesome to take care of them. As these plants have numerous types of diseases merged into them that are hard to recognize. Every time scientists or doctors can't travel from one place to another to examine them, for this kind of problem we have identified the use of Technology like mobile application and deep learning to examine the affected plant and to reduce the human efforts This mobile Android application is built with the latest technologywhereallkindsofuserscanexaminetheirplant to understand the disease embedded into them by simply capturing the image with a camera or by importing from the gallery and can get diagnosis reports with some useful management tips. The user can also get some helpful information about the plants and can contact the adviser or the pesticides retailer to clear their doubts and to proper advice about the diseased plant.
Agroforestry is an occupation where the practice of growing trees and crop farming on an equivalent area of land.Itisaveryimportantandcrucialoccupationinhuman life.Butsometimesduetoinsects,pesticides,globalclimate change,oranotherfactoroftheenvironment,itcanaffectthe plantandmaycauseeconomicandsocialloss.Butmanya timefarmersarenotabletounderstandtheprobleminthe plantwiththeirnakedeyes.So,plantdiagnosisischallenging workintheagroforestryfield.Toovercomethisproblemthe farmerstillusesthetraditionaltechniqueswhicharenotso usefulanddonotgiveproperresults.
Key Words: Leaf Detection, Mobile Application, CNN, Firebase,Authentication.
4Prof. Anuradha Kapoor, Dept of Information Technology Engineering, Atharva College of Engineering, Maharashtra, India
Plant Disease Doctor App
1Kirti Kangane, Dept of Information Technology Engineering, Atharva College of Engineering
Plant disease is one kind of natural disaster that can influence our economic progress as our complete human existenceisdependentonitsowehavetoconserveandtake careofthemfromallthenaturalandmanmadecalamities. Tosearchouttheproblemfromtheaffectedplantleafwe haveproposedanddevelopedamobileAndroidapplication with TFlite, Deep Learning, and Firebase; in which deep learningalgorithmlikeconvolutionneuralnetworkisthere totrainthemodelwithnumeroustypesofplantwiththeir different type of disease. We have gathered 15 different typesofplantswhere5differenttypesofdiseasesofeach plant are there to train the model with TFLite. We have developedanAndroidapplicationsothatuserscanusethe systemanywhereandanytime.Aone timeloginprocessis theretokeeptrackofalltheusers.
3Tejal Kadam, Dept of Information Technology Engineering, Atharva College of Engineering
International Research Journal of Engineering and Technology (IRJET) e ISSN: 2395 0056 Volume: 09 Issue: 03 | Mar 2022 www.irjet.net p ISSN: 2395 0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page1502
1. INTRODUCTION
***
Accordingto[1],10 95%oftheproductivityofagriculture isbeenaffectedbythediseaseintheplant.So,toincreasethe Shampa,ofdetection.accuratetheyNAssociationtoadoptedandproductivityproblemthediseaseshouldbeproperlydetectedbetreated.Forthedetectionofthepurpose,theyhavethemachinelearningalgorithmandcomparedthemgetthehighestaccuracy.Theycomparedalgorithmslikerule,SupportVectorMachine,Clustering,KearestNeighbor,Geneticalgorithm.Fromthisalgorithm,gettheresultthattheSVMalgorithmgivesmoreandbetterclassificationpredictionresultsfortheTheyalsodescribeallthebenefitsanddrawbacksthealgorithmTahminaTashrifMim,HelalSheikh,RoksanaAkterShamimReza,andSanzidulIslamin2019, [2] has
Thisapplicationcannotonlybeusedinagroforestrybut alsoinourday to daylifeaswehaveplantsinourhouseand surround ding nature. To get a more accurate and better result, we are using technology like deep learning i.e., Convolutional Neural Network with Mobile android developmentfordetectingthediseasefromplant'sleafand torecommendthemanagementremediesfortheparticular plantdisease.
Chunnu Khawas, Pritam Shah in [6] , stated the importance of Android mobile development with firebase over the website. The Android application is the fastestas comparedtotheRDBMS.Thepaperalsogavetheimportant ideaoftheAndroiddevelopmentwithfeatureslikeFirebase Analytics, Notification, Crash Reporting, etc. They also comparethefirebasewithSQLandMSSQLanddisplayhow firebaseispreferabletoawebsite.
Forthecomparisonanalysis,wehavecompareddifferent machinelearningalgorithmslikeSVM,LogisticRegression, KNN,DecisionTree,CNN.Theoverallresultofclassification gavetheaccuracyof78%usinglogisticregression,73%for SVM,93%forCNN,83%forKNN.Fromthecomparison,we get to know that CNN is preferable as it gives better classificationandhigherpredictionresultsforthedetection ofleafdiseases.
ShimaRamesh,RamachandraHebbar,NivedithaM,Pooja R, Prasad Bhat N, Shashank N, and P. V. Vinod. in 2018, [3] focused particularly on the Papaya leaf for the proposed system. The algorithms of the machine learning were compared.ThealgorithmslikeLogisticregression,Support VectorMachine,K NearestNeighbor,CART,RandomForest werecompared.Fromthisalgorithm,RandomForestgave thehighestaccuracyofclassificationwhichisapprox.70%, themodelisbuiltfromtherandomforestalgorithmwiththe trainingdatasetof160images.Thepaperalsostatesthatthe accuracyofthesystemcanbeincreasedbyusingafeature like Scale Invariant Feather Transform, Speed Of Robust Feature,andDENSEwithaBagOfVisualWord
3. COMPARISON ANALYSIS
International Research Journal of Engineering and Technology (IRJET) e ISSN: 2395 0056 Volume: 09 Issue: 03 | Mar 2022 www.irjet.net p ISSN: 2395 0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page1503
Shuangjie Huang, Guoxiong Zhou, Mingfang He, Aibin Chen, Wenzhuo Zhang, And Yahui Hu in [4] , aimed at the background interface and the noise problem with image processingonthepeachdiseaseplants.Thispaperstatesthe detectionmethodforthepeachdiseasewiththeAsymptotic Non Local
ChunquanaccuracyoftheConvolutionMeansImagealgorithmandfusedwiththeParallelNeuralNetwork.TheANLMmethodistoreduceinterferenceofbackgroundfromtheimages.Themodelthesystemisbuiltbyusing25513imagesandhasanofapprox.86%withaparticulardiseaseleafPengJiang,YuehanChen,BinLiu,DongjianHe,andLiangin [5] ,focusedonreal timedetectionofthe images of the diseased plant. Disease like Alternaria Leaf, Mosaic,BrownSpot,GreySpot,andRustoftheAppleplant hasbeenimplementedintheproposedsolution.Thepaperis basedonreal timedetectionusingtheCNNalgorithmwith GooglenetInceptionandRainbowconcentration.Themodel ofthesystemiscomposedofadatasetof26,377imagesof theappleplantwhichgaveanaccuracyofapprox.78%
4.1:SystemArchitecture
Chart 3.1:AccuracyofCNNAlgorithm 4.
ofdeeplearningisused.Thesystemisbuiltwithonesingle plantwhichincludes5differenttypesofdiseasewithhealthy plantdiseaselikeBacterialspot,Septorialeaf,Yellowcurve, Tomatomosaic,andLateblightaretaken.Forclassification image processing and CNN algorithm gave the highest accuracy that the other system, the implemented system givestheaccuracyofapprox.89%
ARCHITECTUREChart
Aboveisthearchitectureofthesystemthatrepresents overalltheworkingofthesystem.Firstly,userscansimply loginthroughtheirphonenumberthentheusercancapture thediseasedleaforcanselecttheleafimagefromthegallery. Once the image is selected, the CNN model built with the datasetcomparestheimage withthedatasetanddisplays theresult.Thepredicteddiseaseresultisthensearchedinto thedatabaseforfetchingthesymptomsandmanagementof the disease's leaves and then the fetched information is displayedwith boththeimagesofleavesi.e.,thecaptured imageandcomparedimage
FullyCNN.
7. SYSTEM DESIGN
The best way to focus on users and their activities is throughFirebaseAuthentication.Firebaseisdeliberateasa webapplicationplatform.Ithelpsdevelopersbuildexcellent apps.ItstorestheinformationinJavaScriptObjectNotation (JSON) format that doesn’t use a query for inserting, amending,deletingdatatoit.Itisthebackendofasystemi.e., usedasadatabaseforstoringdata.Toauthenticateourusers, all weneed is to take the verification credentials from the consumerandthenpassthesereferences/credentialstothe FirebaseAuthenticationSDK. Thesereferencescanbeemail password or mobile number or any token from identity providers like Facebook, Google, Twitter, GitHub, etc. The choiceisgiventotheuserabouthowtopresentlogintothe user.Theusercanbuildtheinterfaceorcantakeadvantage of open source UI, which is fully customizable. No matter which one to use, once a user authenticates information about the user is returned to the device via callback. This allows personalizing the application user experience for those specific users. Firebase will also manage the user sessionsothatuserswillremainloggedinafterthebrowser orapplicationrestart.
Connectedlayer:Thislayerformsthelastfewlayersin the network. The input to the fully connected layer is the outputfromthelastPoolinglayerorfromtheConvolutional layer CNNModelBlockDiagram
Plantdiseasedoctorapplicationisanend to endmobile Androidapplicationwhereanalgorithm likeneuralnetwork isused.Theconvolutionalneuralnetworkmodeliscreated thatisaccurateandlightweight.Alightweightmodelmeans the exported model is low in size which is favorable for building a mobile application. We used the plant village datasetandtrainedandtestedmorethan20000plantleaf images.Thisalgorithmisthesimplestchoiceinautomatic imageprocessing.Theapplicationalsoconsistsofabackend sectioncreatedwithPHPTechnologyforthemanagementof blogs and remedies. Firebase is a Google mobile platform that helps in developing high quality applications. This firebase is used for login purposes. With firebase OTP authentication the user can authenticate themselves and entersintothehomescreenoftheapplicationwhenOTPis verified, else the users will get verification failed error message.Wehavealsousedretrofit,i.e.,aRestclientforJava and Android with the help of this we established a client serverconnectionviaHTTPprotocolsuchastogetandpost whichsendsarequesttoawebserver,andreturnaresponse
This system is developed by using the Deep Learning algorithmi.e.,CNN.TheConvolutionalNeuralnetworkshave been used to create and train the system model with the differenttypesofplantleafdatasetsasinput.Thesystemis builtintheandroidstudiousingJavaandKotlinlanguage,for the remedy management wecreateda backendusing PHP Language,foruserloginauthenticationfirebaseisused.
Fig 5.1:
5.2 Login and Authentication in Firebase
Convolutionallayer.Layer:
5.1 Convolutional Neural Network
ForimageprocessingConvolutionNeuralnetworkisthe most effective and reliable algorithm. For automatic processingoftheimage,CNNisthemostreliablealgorithm. AlltheimagesareacombinationofRGB(Red,GreenBlue). Imagesaredifficultforthecomputertoreadascomputers onlyunderstandthenumbers,sotheCNNalgorithmmakes the images simpler and easier so that computer can be processedandunderstood.Allthecoloredimagesarestored in 3 dimensional arrays where the first two dimensions pertain to the height and width of the image and the last dimension corresponds to the red, green, and blue colors presentineachsmallestelementofanimage.Convolutional Neural Network consists of mainly 3 layers i.e., first Convolutional layer, then pooling layer, and then fully connected
5. METHODOLOGY
6. IMPLEMENTATION
International Research Journal of Engineering and Technology (IRJET) e ISSN: 2395 0056 Volume: 09 Issue: 03 | Mar 2022 www.irjet.net p ISSN: 2395 0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page1504
Inausualinterconnectedsystemwhere eachinputneuralisrelatedtothenexthiddenlayer.InCNN, onlyasmalldomainoftheinputlayerneuralcombineswith theotherhiddenlayer.
Pooling Layer: This layer is used to decrease the dimensionality of the feature map. There will be various activationandpoolinglayersinsidethehiddenlayerofthe
The system design comprises the UML Diagram and SequenceDiagram.
9. FUTURE SCOPE
Fig 7.2.2:AdminSequenceDiagram
The UML i.e., Unified Modelling Language is a use case diagram that summarizes the analyses of our system’s consumer and their communication with the system. To build one, we have used a set of specific characters and connectors. An effective use case diagram has helped our teamtoexplainandrepresentourplan.
To increase the number of images, present in the predefined databaseand to modify thearchitecture by thedatasetforachievingthebestaccuracy.
5.Providethebloggingsystem
4.Providesellerandconsultant
Fig 7.2.1:UserSequenceDiagram
7.2 Sequence Diagram
2.prediction.Lesstimeandinexpensivetogetdiagnosisresult
Theblogsareveryuseful.Asnowadaysisusefultospread awarenessofmanyusefulthings.Throughtheblogs,theuser can get to know the various latest information and more usefultipsregardingtheplansmanymore.
3.Recommendationofpesticides
8. ADVANTAGE
Fig 7.1.1:UMLDiagram
A sequence diagram shows object relation ordered in a timesequence.Itpresentsthecommunicationbetweenthe objectsbypassingmessagesfromoneobjecttoanotherin respectoftimeinasystem
7.1 UML Diagram
Nowadayseveryoneofushasasmartphone,byinstalling theapplicationintothemobilephoneanyusercanquickly scan the image with their phone and get the disease
1.Easytodetectdisease
Withthehelpofthelatesttechnologylikedeeplearning,the CNN model is built. The result of the diagnosis will be predicted within a minute at free of cost as installing and usingtheapplicationisfree.
As the system is not only for scanning and the result purpose.Theusercanalsogetaccesstocallandtocontact the pesticides seller or with the consultant to get their problemssolved
With the PHP language, the backend is created for recommendationpurposes.Theusercangettheremedytips andpesticidesrecommendationfordisease.
More enhanced recommendations of pesticides and diagnoses.
Integratingdronesforcapturingimagesfromlargefields.
International Research Journal of Engineering and Technology (IRJET) e ISSN: 2395 0056 Volume: 09 Issue: 03 | Mar 2022 www.irjet.net p ISSN: 2395 0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page1505
subject. We are sincerely grateful for Prof. Anuradha Kapoor as our guide and Prof. Deepali Maste, Head of Information Technology Department, and our project coordinator Prof. Renuka Nagpure for providing help during our research, which would have seemed difficult without their motivation, constant support, and valuable suggestion. Moreover, the complication of this research paperwouldhavebeenimpossiblewithouttheco operation, suggestion,andhelpofourfriendsandfamily
[3] Shima Ramesh; Ramachandra Hebbar; Niveditha M; PoojaR;PrasadBhatN;ShashankN;P.V.Vinod,"Plant disease detection using machine learning" 2018, Internationalconferenceondesigninnovationsfor3Cs computecommunicatecontrol(ICDI3C),IEEE.
ACKNOWLEDGEMENT
Weowesincerethankstoourcollege Atharva College of Engineering forgivingusaplatformtoprepareaprojecton thetopic"PlantDiseaseDoctorApp"andwouldliketothank our principal Dr. Shrikant Kallurkar for giving us the opportunities and time to conduct and research on the
[2] Tahmina Tashrif Mim; Helal Sheikh; Roksana Akter Shampa;ShamimReza;SanzidulIslam,”LeavesDiseases DetectionofTomatoUsingImageProcessing”,20198th International Conference System Modeling and AdvancementinResearchTrends.
[6] ChunnuKhawas;PritamShah,“ApplicationofFirebase in Android App Development”2018, International JournalofComputerApplications,0975 8887.
[8] HarshithaPoojary,ShabariShedthiB,"ASurveyonPlant DiseaseDetectionusingSupportVectorMachine",2018 International Conference on Control, Power, CommunicationandComputingTechnologies.
[7] Taohidul Islam,Manish Sah,SudiptoBaral,Rudra Roy Choudhury,”Afastertechniqueonricediseasedetection using image processing of affected area in agro field”, 2018 International Conference on Inventive CommunicationandComputationalTechnologies.
[1] JyotiShirahatti;RutujaPatil;PoojaAkulwar,"ASurvey Paper on Plant Disease Identification Using Machine LearningApproach",20183rdInternationalConference onCommunicationandElectronicsSystems(ICCES).
[4] Huang,Shuangjie;Zhou,Guoxiong;HeMingfang;Chen, Aibin;Zhang,Wenzhuo;Hu,Yahui,“DetectionofPeach DiseaseImageBasedonAsymptotic Non Local Means and PCNN IPELM” 2020, IEEE Access, 8(), 136421 136433.
[5] Peng Jiang; Yuehan Chen; Bin Liu; Dongjian He; Chunquan Liang, “Real Time Detection of Apple Leaf Diseases Using Deep Learning Approach Based on ImprovedConvolutionalNeuralNetworks.”2019,IEEE Access,7(),59069 59080.
10 1: Implementedsystem
11. CONCLUSION
Addingane commercesitefororderingpesticides.
10. RESULT
Anandroidapplicationfordetectingplantdiseaseandfor providingremedyrecommendationsforaparticulardisease hasbeenimplementedwithtechnologieslikedeeplearning, androidapplication,andfirebase.The proposedsystem is implementedon15typesofplantslikeapple,pepperbell, corn,blueberry,etc.Thissystemcanrecognizethediseaseat anearlystageandcanprovidehelpingtipsforit.Afuture enhancement of the project is to develop hardware technologyforautomaticdetectionwithlesshumaneffort
REFERENCES
Integrationofachatbotformoreadviceandintegration withusers.
International Research Journal of Engineering and Technology (IRJET) e ISSN: 2395 0056 Volume: 09 Issue: 03 | Mar 2022 www.irjet.net p ISSN: 2395 0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page1506
CNNwiththemobileapplicationcanpermittheuserto getthepredictionofdiseasethatemergedintotheplantat anearlystage.Also,itrecommendsusefultipsandremedies formanagingtheaffectedplant.Theusercanalsogetchat accessandcallthesellerorconsultanttoaskforanykindof doubts. Theapplicationalsohasabloggingsystemwhere theusercangettoknowthelatestnewsandtipsaboutthe plants for better economic growth. Following as some imagesofthesystem.
Fig
e
Certified Journal | Page1507
© 2022, IRJET | Impact Factor value: 7.529 |
[9] SachinD.Khirade,A.B.Patil,"PlantDiseaseDetection UsingImageProcessing",2015InternationalConference onComputingCommunicationControlandAutomation.
[10] Mrunmayee Dhakate, Ingole A. B, "Diagnosis of Pomegranate Plant Diseases using Neural Network",2015FifthnationalconferenceonComputer Vision, Pattern Recognition, Image processing, and Graphics.
International Research Journal of Engineering and Technology (IRJET) ISSN: 2395 0056
p
Volume: 09 Issue: 03 | Mar 2022 www.irjet.net ISSN: 2395 0072 ISO 9001:2008