Asphalt Road Quality Classification Based on CNN Architecture VGG16

Damaged road conditions often cause accidents, even resulting in casualties and loss of property as a result of being mired or struck or hit by another vehicle while avoiding the damaged road. Therefore, it is necessary to survey the condition of asphalt roads for the transportation agency, not only to ensure road safety but also to properly implement road maintenance and rehabilitation. In this study, the authors conducted research to classify the good and bad quality of asphalt highways using the convolution neural network method with the VGG16 architecture. In this study, two tests were carried out, namely the first test with 224x224 images and the second test with 64x64 images. The results of this test obtain a good accuracy value for the 224x224 image, which is 92.47%. In an image measuring 64x64 the accuracy level beats in research [12] which is only 87.80%.