Somphop Limsoonthrakul creator Dailey, Mathew N. Chairperson Manukid Parnichkun Examination Committee Kiyoshi, Honda Examination committee Royal Thai Government Scholarship donor text series technical report th Pathum Thani, Thailand Asian Institute of Technology 2009 continuing eng 40 p. : ill. + 1 online resource In term of intelligent vehicle localization, GPS have many advantages to overcome other robot{u2019}s perception techniques. It requires less data processing and is independent from distance traveled and noises in vehicle{u2019}s surrounding environment. However, GPS data also has a limitation that its accuracy depends on the number of available satellites and the quality of signal from them. This affects the uncertainty of GPS and quality of localization in urban area. Therefore, other kinds of perception techniques are required to fuse with GPS in order to increase accuracy of localization. This research proposes a technique to use machine vision with GPS and compass to improve the precision of localization in the situation that GPS is unreliable. The technique is to classify road regions from the scene in front of the vehicle. The information of road regions is used as an addition measurement or observation data for localization process. A filtering technique is require to fuse data from multi sensors and in this thesis, the observation data from GPS, compass, and vision technique are fused together with particle filter method. A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Computer Science (Software Engineering), School of Engineering and Technology Thesis (M.Eng.) - Asian Institute of Technology, 2009 Global Positioning System Computer vision Electronics in navigation Asian Institute of Technology. http://203.159.5.9/ait-thesis/detail.php?q=B02097 http://203.159.5.9/ait-thesis/detail.php?q=B02097 091111 20260818143640.0