Road condition automatic identification and analysis system
碩士 === 東吳大學 === 資訊管理學系 === 106 === According to the statistics of the Ministry of Communications, drivers are the ones to blame for most of the car accidents. This paper aims to use software to identify the approaching vehicle and help the driver to avoid collisons using the computer vision techniqu...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2018
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/529843 |
| Summary: | 碩士 === 東吳大學 === 資訊管理學系 === 106 === According to the statistics of the Ministry of Communications, drivers are the ones to blame for most of the car accidents. This paper aims to use software to identify the approaching vehicle and help the driver to avoid collisons using the computer vision techniques. The vehicle detection technique used in this dissertation is based on the Haar Cascade method. The eigenvalues are calculated from Haar-like features, and the values are then iterated into many weak classifiers and paralleled into strong classifiers via AdaBoost, and eventually the trained classifiers are used. The matching of our experiment successfully finds features from the vehicle image to achieve rapid vehicle identification. In the future, it can be combined with a driving recorder or modified into a real-time vehicle detection system.
|
|---|