Hybrid Decision Support System Framework for Leaf Image Analysis to Improve Crop Productivity

Crop disease is one of the major problems with agriculture in India. Identifying and classifying type most important which can be made possible using deep learning technique. To perform this verified dataset required consists healthy leaf images all crops. The proposed model uses a hybrid approach integrates VGG16 classifier an attention mechanism, transfer dropout operation. rice it achieves train accuracy 96.45 percent loss 0.09 validation 0.44. collected from plant village project for 4955 include Brown Spot, Healthy, Hipsa, Leaf Blast images. use mechanism that focuses mainly on part image rather than whole glimpse ratio 3:1. traditional method detecting crop diseases needs high experience knowledge experts field time consuming, ineffective, cost. In study, Deep Convolutional Neural Networks (DCNN) Transfer Learning Attention models are used to detect associated plants without overfitting model.