特约专稿：作物病害智能诊断与处方推荐技术研究进展——张领先 等

DOI:10.6041/j.issn.1000-1298.2023.06.001。

Abstract:The"植物诊所" hasprovided newideas fortherecommendation ofcrop diseasetreatments.However, itis stillan urgentandchallenging tasktoeffectively extractand assisttherecommendation ofcrop diseasetreatmentsbased onelectronicmedical records(EMR). Inthispaper, wesystematicallyanalyzed anddiscussed thekeytechnologies,including imageanalysis ofplantpathogenic bacteriausingmicroscopy,spectroscopy,andEMR,and therecommendationof cropdiseasetreatments basedon theEMR. Wealso proposeda researchdirection thatfocuses onthe analysisandutilization ofEMRbased onthecharacteristics ofplantpathogenic bacteriaand theneeds ofprecisionrecommendation incropdiseases. Theresults showedthat theresearch onthe analysisandutilization ofEMRbased onthecharacteristics ofplantpathogenic bacteriaand theneeds ofprecisionrecommendation incrop diseasescan bedivided intothreecategories:(1) theanalysis andutilizationofEMR basedon thecharacteristicsof plantpathogenicbacteria andthe needsof precisionrecommendationin cropdiseases,(2) theanalysis andutilizationofEMR basedon thespectroscopiccharacteristicsof plantpathogenicbacteria andthe needsof precisionrecommendationin cropdiseases,and(3)the analysisandutilization ofEMRbased onthecombination ofthe abovetwocategories.The researchon theabove threecategoriescan provideascientific andsystematicbasis fortherecommendation ofcrop diseasesbased onEMR.

Research ProgressinIntelligentDiagnosis andPrescriptionRecommendation ofCropDiseases

Abstract:The"plantclinic" hasled tothe creationof plantelectronicmedical recordsthat providenewpossibilities forthe diagnosisandrecommendation ofcropdiseases.However, theefficient miningandutilization ofthese dataremains achallenge,and thereis anurgent needfor researchto beconducted bothat homeandabroad. Based ona reviewof theexistingliterature,we haveidentifiedthe keytechnologiesinvolved incrop diseasediagnosis andrecommendation, includingsporesrecognitionbased onmicroscopicimages, cropdisease diagnosisbased onspectra, andcrop diseaseprescriptionrecommendationbased onelectronicmedicalrecords. Wehave alsoanalyzed anddiscussed theexisting researchon thesetopics,as wellas thechallengesthat mustbe addressedin ordertosuccessfully implementthem. Inthe contextof cropdisease diagnosisandrecommendation,it isclear thatthe analysisofelectronic medicalrecord datamining willbecome acentralfocus, withthe goalof assistingin theidentificationandmanagement ofcroppathogens. Thiswill beachieved throughtheapplication ofintelligentprescriptionrecommendationalgorithmsthat takeinto accountboth thecharacteristicsandcomplexities ofcroppathogenesis. In orderto achievethisgoal, itwill benecessary toconduct researchon theanalysis ofcrop diseasepathogenesis,diagnosticreasoning,andintelligentprescriptionrecommendation.This willinvolve theuse ofa rangeoftechnologies,including knowledgegraphanalysis, bigdatamining, andmachine learningalgorithms. Byutilizing thesetechnologies, itwill bepossible toanalyze andvisualizethepathogenicmechanisms ofcropdiseases, aswell astheircorrelations withcharacteristicsfrom aregional macroperspective. Thisresearch willalso aimto providea basisfor thedevelopmentof singleand multiplecrop diseaseprescriptionrecommendationalgorithms, basedonsemantic matchingand knowledgegraphanalysis. The resultsof thisresearch havethe potentialto greatlyimprove thediagnosis andmanagementof cropdiseases,and willbe ofgreat practicalsignificancein thefield ofplanthealth.

Keywords: cropdisease,pathogensporerecognition, diseasedetection,diseasediagnosis,prescriptionrecommendation,plantelectronic medicalrecords。

请点击阅读原文查看全文。