Automated Road Condition Analysis from Video Footage and Accelerometer Data in Developing Countries
The objective of this study was to compare road quality monitoring methods and determine a suitable method for Zanzibar, Tanzania. Road quality monitoring is required for timely road maintenance, enabling uninterrupted and safe travel for citizens and businesses. With growing economies and infrastru...
| Main Author: | |
|---|---|
| Format: | Dissertation (University of Nottingham only) |
| Language: | English |
| Published: |
2020
|
| Online Access: | https://eprints.nottingham.ac.uk/62038/ |
| _version_ | 1848799928112906240 |
|---|---|
| author | Shah, Maryam |
| author_facet | Shah, Maryam |
| author_sort | Shah, Maryam |
| building | Nottingham Research Data Repository |
| collection | Online Access |
| description | The objective of this study was to compare road quality monitoring methods and determine a suitable method for Zanzibar, Tanzania. Road quality monitoring is required for timely road maintenance, enabling uninterrupted and safe travel for citizens and businesses. With growing economies and infrastructure, the network of roads in developing countries is expanding, and monitoring them through traditional ways is becoming cumbersome and ineffective. This requires advanced methods that are easily replicable, incur fewer costs and enable quick decision making. Existing methods have certain advantages and limitations, which makes matching the context of the problem with the method’s capabilities very important. This study compares two frequently used road monitoring methods, vibration-based and vision-based, in Zanzibar’s context. Data collected in collaboration with the Zanzibar Department of Roads, including accelerometer data and video footage of the roads, was used to train three different models. Vibration-based models were established to be more suitable for Zanzibar, specifically the random forest classifier with feature extraction, as they provided higher accuracy and had ease of replication using fewer resources. Additionally, within the accelerometer data used for vibration-based methods, it was found that each axis’ data could predict road quality independently with equally good accuracy. |
| first_indexed | 2025-11-14T20:43:27Z |
| format | Dissertation (University of Nottingham only) |
| id | nottingham-62038 |
| institution | University of Nottingham Malaysia Campus |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-14T20:43:27Z |
| publishDate | 2020 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | nottingham-620382022-12-21T14:32:07Z https://eprints.nottingham.ac.uk/62038/ Automated Road Condition Analysis from Video Footage and Accelerometer Data in Developing Countries Shah, Maryam The objective of this study was to compare road quality monitoring methods and determine a suitable method for Zanzibar, Tanzania. Road quality monitoring is required for timely road maintenance, enabling uninterrupted and safe travel for citizens and businesses. With growing economies and infrastructure, the network of roads in developing countries is expanding, and monitoring them through traditional ways is becoming cumbersome and ineffective. This requires advanced methods that are easily replicable, incur fewer costs and enable quick decision making. Existing methods have certain advantages and limitations, which makes matching the context of the problem with the method’s capabilities very important. This study compares two frequently used road monitoring methods, vibration-based and vision-based, in Zanzibar’s context. Data collected in collaboration with the Zanzibar Department of Roads, including accelerometer data and video footage of the roads, was used to train three different models. Vibration-based models were established to be more suitable for Zanzibar, specifically the random forest classifier with feature extraction, as they provided higher accuracy and had ease of replication using fewer resources. Additionally, within the accelerometer data used for vibration-based methods, it was found that each axis’ data could predict road quality independently with equally good accuracy. 2020-12-01 Dissertation (University of Nottingham only) NonPeerReviewed application/pdf en https://eprints.nottingham.ac.uk/62038/1/Lixms71-BUSI4374-FinalDissertation.pdf Shah, Maryam (2020) Automated Road Condition Analysis from Video Footage and Accelerometer Data in Developing Countries. [Dissertation (University of Nottingham only)] |
| spellingShingle | Shah, Maryam Automated Road Condition Analysis from Video Footage and Accelerometer Data in Developing Countries |
| title | Automated Road Condition Analysis from Video Footage and Accelerometer Data in Developing Countries |
| title_full | Automated Road Condition Analysis from Video Footage and Accelerometer Data in Developing Countries |
| title_fullStr | Automated Road Condition Analysis from Video Footage and Accelerometer Data in Developing Countries |
| title_full_unstemmed | Automated Road Condition Analysis from Video Footage and Accelerometer Data in Developing Countries |
| title_short | Automated Road Condition Analysis from Video Footage and Accelerometer Data in Developing Countries |
| title_sort | automated road condition analysis from video footage and accelerometer data in developing countries |
| url | https://eprints.nottingham.ac.uk/62038/ |