Identifikasi Tumbuhan Obat Herbal Berdasarkan Citra Daun Menggunakan Algoritma Gray Level Co-occurence Matrix dan K-Nearest Neighbor

Fittria Shofrotun Ni'mah (Department of Informatics, Universitas Dian Nuswantoro)
T Sutojo (Department of Informatics, Universitas Dian Nuswantoro)
De Rosal Ignatius Moses Setiadi (Department of Informatics, Universitas Dian Nuswantoro)

Jurnal Teknologi dan Sistem Komputer
Volume 6, Issue 2, Year 2018 (April 2018)

Medicinal plants can be used as an alternative natural treatment, instead of chemical drugs. But because of too many types of plants and lack of knowledge, it will be difficult to identify these herbs. Computer assistance can be used to facilitate the identification of these herbs. This research proposes the identification of herbal plants based on leaf image using texture analysis. There are 10 types of herbal medicinal plants used in this study. The texture analysis used was GLCM by extracting contrast, correlation, energy, and homogeneity. Classification is done by KNN. The result of the experiment showed that the accuracy of identification using 9-fold cross-cross validation method was 83.33% using 9 subsets.