Herb Leaves Recognition using Gray Level Co-occurrence Matrix and Five Distance-based Similarity Measures

R. Rizal Isnanto (Diponegoro University)
Munawar Agus Riyadi (Diponegoro University)
Muhammad Fahmi Awaj (Diponegoro University)

International Journal of Electrical and Computer Engineering
Vol 8, No 3: June 2018

Herb medicinal products derived from plants have long been considered as an alternative option for treating various diseases.  In this paper, the feature extraction method used is Gray Level Co-occurrence Matrix (GLCM), while for its recognition using the metric calculations of Chebyshev, Cityblock, Minkowski, Canberra, and Euclidean distances. The method of determining the GLCM Analysis based on the texture analysis resulting from the extraction of this feature is Angular Second Moment, Contrast, Inverse Different Moment, Entropy as well as its Correlation.  The recognition system used 10 leaf test images with GLCM method and Canberra distance resulted in the highest accuracy of 92.00%. While the use of 20 and 30 test data resulted in a recognition rate of 50.67% and 60.00%.