Connie Lee Wai Yan creator author. text monographic eng

print

xiv, 81 pages : illustrations ; 1 CD-ROM Road maintenance is a challenging task in Malaysia. Malaysia has seen an increase in the number of fatal transport accidents, such as road accidents caused by potholes, year after year. Aside from potholes, speed bumps increase the number of injuries in road traffic accidents, even if no fatalities occur. Rather of waiting for the government to remedy the problem, there is an increasing need for a low-cost automatic detection of speed bumps and potholes. Many recent techniques have demonstrated promising results when using deep learning to various object identification tasks. Convolutional Neural Networks (CNNs) able to learn how to extract significant characteristics from an image. Many recent techniques have demonstrated promising results when using deep learning to various object identification tasks. In this paper, two image datasets has been download from Kaggle. The datasets are annotated by LabelImg and trained YOLO (You Only Look Once) version 4. The result is evaluated based on the confusion matrix, precision, recall, f1-score and mAP. The model is tested on different speed bumps and pothole images, and it detects with a reasonable accuracy. The expected outcome is to build a model that can effectively detect speed bumps and potholes to help drivers and car manufacturers. Connie Lee Wai Yan Centre for Mathematical Sciences Bachelor of Applied Science in Data Analytics with Honours -- Universiti Malaysia Pahang – 2023 Includes bibliographical references Centre for Mathematical Sciences Dissertations Universities and colleges Dissertations Final Year Project Dissertations THE0010020 (Local) UMPSA 250502 20251125111037.0 eng