Abstract:
India, thecountrywherethemainsourceofincomeisfromagriculture.
Farmersgrowavarietyofcropsbasedontheirrequirement.Sincetheplants
su er fromthedisease,theproductionofcropdecreasesduetoinfections
caused byseveraltypesofdiseasesonitsleaf,fruit,andstem.Leafdiseases
are mainlycausedbybacteria,fungi,virusetc.Diseasesareoftendi cult
to control.Diagnosisofthediseaseshouldbedoneaccuratelyandproper
actions shouldbetakenattheappropriatetime.ImageProcessingisthe
trending techniqueindetectionandclassi cationofplantleafdisease.This
workdescribeshowtoautomaticallydetectleafdiseases.Thegivensystem
will provideafast,spontaneous,preciseandveryeconomicalmethodin
detecting andclassifyingleafdiseases.Thispaperisenvisionedtoassistin
the detectingandclassifyingleafdiseasesusingMulticlassSVMclassi cation
technique.First,thea ectedregionisdiscoveredusingsegmentationbyK-
means clustering,thenfeatures(colorandtexture)areextracted.Lastly,
classi cation techniqueisappliedindetectingthetypeofleafdisease.
Keywords: Image Processing,Leafdiseasesdetection,K-meansclustering,
featureextraction,MulticlassSVMClassi cation.
