Hybrid Feature Extraction Method for Efficient Leaf Disease Detection and Grading

Efficient feature Extraction Technique is a ablaze topic that how-to take-out prominent image structures with a sturdy representation from the processing image using brilliant technology. Since every disease has new symptoms, it is very difficult to process and train the algorithm. The proposed research steps consist of features extraction and classification. Histogram of gradient (HOG) method is used for colour feature extraction, Shape features are extracted using Edge histogram descriptor (EHD). In texture-based features, we extracted features using gray level co-occurrence matrix (GLCM). After combining all features it results in the high dimensional feature vector. Finally, the classification is performed using the artificial neural network (ANN) classifier and Support vector machine (SVM) classifier, MATLAB is software for simulation. Proposed approach achieved the significant improvement in performances with 95.12 and 93.30% accuracy with different training testing ratio with hybrid feature extraction method and ANN classifier.