Convolutional Neural Networks in Detection of Plant Diseases

الملخص

Pests and illnesses that affect plants can drastically lower crop yields and quality, endangering consumers' health. Some plant diseases can be so bad that they totally ruin grain harvests. Therefore, there is a great need in the agricultural data sector for systems that can detect and diagnose plant illnesses automatically. Color features, vein features, Gray-Level Matrix techniques, and Fourier descriptors are few of the feature extraction-based image processing approaches used here. The results concluded here support the utilization of the features obtained from divided leaves instead of the entire leaf. Combining  CNN with SVM and KNN  was suggested in this study. It evaluated and collated the suggested method's accuracy with approaches from other studies after using 10-fold cross-validation to assess its accuracy.