The increasing global demand for electricity and the environmental impacts caused by thermal power generation have led to investments in renewable energy sources, such as photovoltaic solar energy (PV). However, the growth of installed capacity in photovoltaic generation systems requires intelligent systems to detect faults in their components, particularly in PV cells. As converters of solar energy into electricity, the latter operate exposed to adverse environmental conditions such as wind, rain, salinity, and dust, which degrade and compromise the efficiency and reliability of the generation system. In this context, an intelligent system is proposed in this study for the detection of defects in PV cells using electroluminescence (EL) images. This system is based on a supervised classification model, in which image features are represented by a texture descriptor constructed from the Gray-Level Co-occurrence Matrix. The model was evaluated through cross-validation on a public dataset of 2624 PV cell images, demonstrating sensitivity to defects such as shading, hotspots, cracks, and microfissures. Its best performance was achieved with the Random Forest (RF) classifier, yielding competitive results compared to those found in the literature. Thus, it presents a viable alternative for automated inspection of PV cells.