Exporting mangos generally requires accurate assessment of the mango quality based on different criteria to ensure the fruit’s value and prestigious brand-name. Automating the mango classification process based on surface features is usually performed by a machine vision solution. Because mangoes have heterogeneous surface curvature, quantitative evaluation of some features such as the area of defects on the mango skin requires that images of this feature area be taken from the front view to ensure accurate assessment results. Due to the oblong shape of the mango fruit, the quantitative indicators on the mango skin can be effectively determined from the front-view images of the mango surface on both side of its seed, on the dorsal and ventral surface. Therefore, this study aimed to develop an image acquisition system to capture images of the whole mango surface from which the front-view images of the four major mango faces could be identified for the benefit of developing an automatic mango grading system based on surface criteria. Experimental results on 104 Cat Hoa Loc mangoes showed that the images of the mango sides and edges were identified with the accuracy of 94.7% and 92.8%, respectively. These preliminary results show the potential for developing an automatic mango classification system using machine vision based on various quality features, especially quantitative ones.