Font Size: a A A

Based On Regional Characteristics And Relative Space Relations Tie Pattern Recognition

Posted on:2006-04-24Degree:MasterType:Thesis
Country:ChinaCandidate:H J ZhuFull Text:PDF
GTID:2208360152990757Subject:Systems analysis and integration
Abstract/Summary:PDF Full Text Request
Textile CAD are applied so widely that people can design tie very quickly and a large quantity of tie pattern are produced. So it is a problem that people can find the tie pattern designed or similar in the condition that the tie sample is provided. If it can do, time waste will be avoided and efficiency will be increased. The recognition of tie pattern is that the tie sample is scanned to image which will be pretreated and abstracted to character vectors and will be matched by computer so that the designed or similar tie pattern in database will be found.The writer studys the character of tie pattern widly and deeply, and then introduces a kind of applied tie recognition based on region character and location relations. The recognition is efficient .The character of tie textile pattern is that the category of color is little and color spreads by region obviously and the tie pattern designed is different from tie pattern weaved in color of same regions. The method avoids analyzing the structure and style of tie pattern and reduces time of running and so cost is lower. All of these are tested by practice and the result is good.A tie pattern recognition system is designed according to the method. It applies the recogntion function through three modules:1) Pretreatment part. It includes the translation from 24 bit real color image to 256 bit color index image, the translation from color image to lightness image, the smoothness treatment and collecting color to color block.2) Character abstracting part. In this part , a 8 direction operator is introduced, which puts number to color blocks. The character vector of color block comes from the searching path among color blocks. The chatacter vectors are composed to a character matrix.3) Match and recognition part. In this part, distinguish function is the least square error rule function. Tie pattern characters are put into tie pattern data structure in tie pattern character database management and so tie pattern assistant information can be moved conveniently.
Keywords/Search Tags:region character, location relation, pretreatment, character abstract, character match
PDF Full Text Request
Related items