A light point algorithm of classification of vehicle at night

[[abstract]]It is years to apply the image process to the traffic. However, no matter domestically or internationally, there is rare study, which is made for the night condition. Here, by judging the single motionless image of traffic flow that is on the basis of the amount of light points showing on the image and the relation of light points, we could classify the vehicles as large vehicles, small vehicles and motorcycles. Thus, with the statistical data on rear light templates, we conclude the light points height, width, distances between two rear lights and angle, which is formed. Then, comparing above information with the traffic flow image, different sorts of vehicles would be distinguished.[[sponsorship]]視訊媒體新世紀CVGIP2001研討會=2001 14th IPPR Conference on Computer Vision Graphics and Image Processing[[sponsorship]]中山大學資訊工程系; 中華民國影像處理與圖識別學會; 第一科技大學電腦與通訊工程系; [特刊名中文名稱] National Kaohsiung First University of Science and Tachnology[[notice]]補正完畢[[conferencetype]]國際[[conferencedate]]20010819~20010819[[booktype]]紙本[[iscallforpapers]]Y[[conferencelocation]]屏東縣, 臺