Character Recognition In Computer Vision And Software Development

Auto-driving robot is a complex system composed of entironment detect, layout decision and auto-driving. It also is one of important parts of ITS(InteIIigent Transportation System).This thesis mainly studies character locating and recognition in vision systems of auto-driving robot.Character locating and recognition is a kind of intelligent technology which must be solved. It is not only be applied in driving robots and auto-moving androids,but has lots of potential application values.In this thesis, first apply different preprogress in images shot in different entironments. Linear transform extends grey range , moving-dark image recover can effectively conquers moving-dark phenomena. Second several algorithms presented formerly are contrasted in thie thesis, common points are summarized A new locating algorithm is then designed based on these research results.Third apply binary , segment and thinning in character image area ,make character images become satisfying. Finally combine character recognition based on characteristics of characters and the method of artifical neural nerwork A high recognition is reached At the same time lots of programmings. All algorithms refered in this thesis are programmed and compiled into .dll files/They can be transferred by other systems, modularized and easy to maintain.Lots of experimentations proved .the algorithms are doable,the rate of recognition is high,and the practicability is good.This system is promising.