Glaucoma is a chronic eye disease causing permanent damage due to optic disc deterioration, leading to vision reduction and potential blindness if untreated. Early detection is crucial as symptoms mainly appear in advanced stages. Retinography, a low-cost examination, is typically used for eye disease detection by imaging the optic nerve. This research introduces a method combining texture and structural feature analysis on fundus images to classify them as glaucomatous or healthy. The method comprises seven stages: (1) image acquisition; (2) image pre-processing, based on the application of histogram specification and CLAHE, prior to texture feature extraction; (3) segmentation of the optic disc and excavation using the deep neural network DC-GNet; (4) texture feature extraction using diversity indices; (5) extraction of structural features including CDR and ISNT; (6) feature combination and selection, and (7) classification. Experiments conducted on 660 images from three public retinography datasets showed the method achieved 90% accuracy, 90% sensitivity, 89% specificity, and a 92% F1-score, proving its effectiveness and feasibility for use on datasets with limited images.