The Study On Chain Code And Implementation On Chess-image Recognition Based On Chain Code

The research about image processing and pattern recognition is a focus in the field of computer science.The digital image processing is the precondition for image recognition.In 1962,Freeman created the theory about chain code that is called Freeman chain code.The technology of chain code is a practical way for processing ditital image and capturing features from it.It makes capturing features much easier through changing this two-dimensional matrix of pixels to a one-dimensional chain code.The research about how to get the chain code efficiently from an image is a worthy field.This article introduces and researches the theories about boudary-tracking automation in details,and the algorithms about how to get the chain code via this boundary-tracking automation.There are already many published papers about filling region upon getting the chain code of digital image.This paper makes a simple comparision among normal existing region-filling algorithoms based on chain code and focuses on REN algorithm that is the most effieciect one so far.At the same time,this paper points out some shortcomings about REN algorithm and proposes an optimized region-filling alogithm.According to the result of experiments,it's more effiecient when filling a multi-connectivity region.The I-go game has long history in China and attracts lots of players.This paper researches and creates the algorithm series about chess-image processing and recognition based on the technology of chain code in order to recognize static chess documents into interactivedigital chesses.The preprocessing and recognition of chess-image can be divided into seveval procedures,including the detection of skew angle for the chess image,the capturing of recognition target from image chess,the recognition of cell,white chess and black Chess from the target region.