Categories Leaf Healthiness using RGB Spectrum and Fuzzy Logic

In this paper, a general approach is to classify of the green leaf healthiness. Fuzzy logic tool (FuzzyLite 3.2 software) and color features (RGB Spectrum) are used in this experiment. Mean values of primary colors (Red, Green and Blue) channels as input to FIS (Fuzzy Inference System). FIS gives decision whether this part of leaf is healthy, unhealthy or dying. Experimentation is conducted on our own dataset for determining knowledge base, consisting of 40 images of leaves for each category; 20 for training and 20 for testing. The experiment has 4 phases which were data preparation, features extraction, features selection and classification. The experimental results indicate that proposed model achieves a good average classification accuracy which are 85% healthy, 95% unhealthy and 100% dying.