Abstract
Pothole is one of the problems that can cause harm to a person or a lot of people and can even cost lives. So a lot of research has been done to detect potholes, especially image-based. This research uses Unmanned Aerial Vehicle (UAV) to get aerial video dataset and train Convolutional Neural Network (CNN) with the dataset. However, instead of doing learning from the beginning, transfer learning can be used to train CNN to recognize the object of a pothole and measure the value of its performance and what the optimal frame rate is. Then the results of this study indicate that the CNN model, ssd_resnet_50_fpn_coco gets an average performance value of 48.90 mAP. And the optimal frame rate with the average highest performance value at a frame rate of 30FPS with a value of 49.43 mAP, followed by 1FPS with a value of 48.36 mAP .
 Keywords: Performance, Transfer Learning, Convolutional Neural Network, Pothole Detection, Aerial Video.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
