A comparison of HOG and SURF methods for vehicles type classification | |
| Author | Boonthuang Yodon |
| Subject(s) | Vehicles--Classification Image processing |
| Note | The Degree of Master of Science in Mechatronics, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Abstract | The detection and classification system by image processing are quite popular and attractive method to many researchers because those method can be apply to use in many areas. For example it can be apply in the traffic control purpose or medical purpose. This research has presented the comparing between HOG and SURF method to see that which method show a better result. This research divide the vehicle in to six types : motorcycle,car,van,suv,pick up and tuk tuk. In the algorithm part was starting by cascade training and detection algorithm. This part required a large amount of pictures to train the system in order to detect the vehicles. HOG and SURF method are the process to find the extract feature image to train SVM for recognize and memorize vehicle and finally can be classify vehicle. After tested for those system, HOG has the accuracy result on average at 73.17 % and SURF show the accuracy result at 60.54 % . Eventually, HOG produce the better result than SURF method. |
| Year | 2014 |
| School | School of Engineering and Technology (SET) |
| Department | Department of Industrial Systems Engineering (DISE) |
| Academic Program/FoS | Microelectronics (ME) |
| Chairperson(s) | Mongkol Ekpanyapong; |
| Examination Committee(s) | Manukid Parnichkun;Dailey, Matthew N.; |
| Scholarship Donor(s) | Western Digital (Thailand) Co., Ltd.;National Electronics and Computer Technology Center,;AIT Fellowship; |