Loading...

Crack Detection of Asphalt Concrete Pavements Based on Deep Learning

Sepidbar, Alireza | 2023

64 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56266 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Sabouri, Mohammad Reza
  7. Abstract:
  8. The health of the pavement ensures the safety and convenience of drivers and passengers. In the past few decades, pavement management systems have encountered challenges that often have produced solutions with excessive demand for resources, but low-accuracy results. New approaches must be developed in order to quickly and economically identify pavement failure, especially cracks. In recent years, researchers have focused on identifying pavement failures, but previous methods only worked on images that solely included pavements and cracks. However, when foreign objects such as cars and vegetation were present, these methods were not as effective. To improve upon these methods, semantic segmentation was used to separate the pavement surface from the images. Also, two methods were presented for identifying pavement cracks. In the first method, using image-to-image translation, the input image is converted into the structure of the ground truth image, and cracks are identified in this way. In the second method, U-Net architecture and the grasshopper optimization algorithm for optimization were used. This method produced better and more accurate results than the first method. On average, the presented method has a precision of 0.9484, a recall of 0.8803, and an F1-score of 0.9123. Additionally, a method for pavement management based on crack dimensions was presented, which involved measuring cracks in images and displaying them on a map. Urban areas were compared based on three characteristics: average crack length, average crack width, and number of cracks. The analytic hierarchy process was used to give weight to each characteristic based on its importance, and each area was given a score according to its pavement crack condition. This method allows for urban areas to be compared and critical areas to be identified.
  9. Keywords:
  10. Pavement Distress ; Crack Detection ; Semantic Segmentation ; U-Net Model ; Crack Measurement ; Image-to-Image Translation ; Deep Learning

 Digital Object List

 Bookmark

No TOC