Kimberley Road Degradation Detection Using Digital Image Processing

The city of Kimberley in the Northern Cape province of South Africa faces significant road degradation due to heavy traffic, poor road design, extreme weather, and substandard construction materials, posing safety hazards and increasing maintenance demands. Traditional monitoring methods are costly and labor-intensive; thus, this chapter explores a cost-effective solution using smartphone technology to collect data and develop a model for detecting and classifying road defects. A Kimberley-specific dataset was created with six classes of road defects, annotated using the Roboflow platform, and aligned with TRH22 guidelines for pavement management systems. The YOLOv3-CNN developed on the Kimberley dataset achieved an 83.38% accuracy. Future work will focus on improving limitations and integrating the solution into municipal road maintenance systems. This chapter provides a framework for similar applications in other regions and aligns with industry standards for civil engineering practices.