2020 · 12 citations · 9 references
AgriculturePrecision AgricultureWeed ScienceComputer VisionCrop ManagementWeed IdentificationAgricultural Machine VisionCrop ClassificationDeep LearningPrecision Crop ManagementWeed DetectionWeed ManagementConvolutional Neural NetworksCrop YieldsField Weed ClassificationField-based Weed RecognitionPrecision Weed ControlPrecision WeedingSupport Vector ClassificationSupport Vector Machine Classification
Weeding is an effective method to improve crop yields. In precision agriculture, accurate and reliable weed identification methods provide the basis for achieving high-precision weed position control. This paper proposes a field weed and crop classification algorithm based on deep convolutional neural network and uses support vector machine (SVM) to optimize the algorithm. The model using VGG16+SVM for classification tasks has achieved an accuracy rate of 96,4%, which has practical significance for modern agricultural precision weeding.
9