<![CDATA[Citrus as a major agricultural commodity in Indonesia, plays a crucial role in the industry and farmers' income. Identification of citrus seedling types is a major challenge, due to the lack of knowledge and experience of farmers, causing potential financial and time losses. This study compares the Artificial Neural Network Backpropagation (JST-PB) method and Gray Level Co-occurrence Matrix (GLCM) features in orange seedling type identification through leaf vein images. Data was collected using a macro camera with Samsung ISOCELL GM2 sensor, with various cropping sizes on a total dataset of 1250 training images and 625 test images. The JST-BP method and GLCM features provided an accuracy rate of 91.2% at a cropping size of 200x200 piksels, 87.2% at a cropping size of 250x250 piksels, 90.4% at a cropping size of 300x300 piksels, 95.2% at a cropping size of 350x350 piksels, and the highest accuracy rate at a cropping size of 400x400 piksels, reaching 98.4%. The results of this study make an important contribution to the understanding of the identification of citrus seedling types through leaf vein images, highlighting the comparison between the JST-PB method and GLCM features at various image cropping sizes.]]>
Copyrights © 2025
