An Algorithm for Various Crop Diseases Detection and Classification using Leaves Images

Economic graph of the most of the developing countries depends on the agricultural production. To maintain the economical figure the crop damaging factors that affects yield as various diseases needs to be diagnosed and cured appropriately. To determine the crop disease based on the leaf image analysis the proposed technique uses local image information. The local information of leaf image can be extracted using the leaf color feature for seed region extraction and region growing along with the contouring through the level-set and sample consensus to fit the leaf boundary. The classification is performed using the Fisher vectors of the localized images. The effectiveness of this proposal is analyzed with PlantVillage dataset using the vector support machine and multi-layer perceptron model. The analysis shows that the proposed technique is better as compared to the state of art methods with average classification accuracy and area under curve of 91.22 % and 92 % respectively.