Abstract: |
Asphalt pavement distress, the various defects such as holes and cracks, represent a significant engineering and economic concern. It is estimated that pavement defects cause damage costing $10 billion each year in the United States alone [10]. One important step in managing this problem is accurately assessing the pavement condition and its change over time. In this paper we compare three methods for automatically classifying pavement cracks, genetic algorithms, multilayer perceptrons, and self-organizing maps. We also discuss the impact of feature representation on the resulting classification. Our best classifiers demonstrated accuracies between 86 and 98%. |
