An Adaptive Image Processing Model of Plant Disease Diagnosis and Quantification Based on Color and Texture Histogram

In this paper, a new approach for the detection and classification of potato plant disease is implemented using computer vision techniques. Most of the existing algorithms based on plant disease detection and classification are limited to common types of feature extraction methods. However, feature extraction is an important area as the classification of diseases of any leaf. The proposed method is based on color and texture features. The implemented method processed in four steps- In the preprocessing and segmentation, LAB color space and Delta E color difference method are applied. Later, features are extracted based on RGB, HSV and Local Binary Patterns (LBP). The extracted patterns are finally classified by Multi Support Vector Machine (SVM). Moreover, we compare the results of feature subsets of RGB and HSV color features with the addition of LBP texture features and found a classification difference of 3.6% between RGB and HSV color feature extractors. The overall results show our method outperforms as compared to existing techniques.