Leaf morphological feature extraction based on K-Nearest Neighbor

Type a plant can not only be seen with the constituent structure of the body alone. The species can be classified based on each constituent part of the body such as bone and classification based on color of the leaves of a plant. Processes required by way of feature extraction using the concept of Image Processing. Firstly, Feature Extraction has been done using methods Distance Angle-Signature, the classification is done leaf by leaf bone forms, named: pinnate, palmate, curved and parallel. Secondly, Feature Extraction using the Histogram of Oriented Gradient, performed classifying the leaves by color pigments contained in the leaves, such as: Chlorophyll-containing pigment green, carotenoid-containing pigment in red, orange and yellow, anthocyanins pigment containing blue and purple. Our proposed method using K-Nearest Neighbor with K = 1 to K = 9 for each feature extraction. On Feature Extraction Distance Angle Signature obtained the best percentage with the value of K = 1 and K = 4 at 100% and 90%, while in Feature Extraction Histogram of Oriented Gradients with K = 1 and K = 4 at 100% and 73.33%.