Research And Application Of Block Image Segmentation And Recognition

The recognition and comprehension of images is an important research direction in the field of machine vision.It has important academic value and broad application field to automatically recognize user's operation behaviors and contents.This paper,aiming at the need of automatic recognition of user's behaviors and contents in the Internet cafes billing system,carries out the research and the preliminary application of image shots on factor segmentation and automatic identification of information.Firstly,a measure based on the similarity of color histograms of two adjacent rows or columns of a screenshot defines the concept of a screenshot image stabilization block(a stable row or column block).And a model of automatic segmentation of the three functional areas of the screenshot image(that is,image grid area,menu bar area and dialog area)based on the stable row or column block of screenshot is proposed.Moreover,an automatic segmentation algorithm is designed.Secondly,further combining with the characteristics of each functional areas to carry out the research on automatic segmentation of elements in the three areas and automatic recognition of segmentation smallest sub-images of text information,an automatic segmentation method of human head in dialog area is proposed,and an algorithm is proposed to segment text line automatically and divide each text line into smallest sub-images in the menu bar areas and dialog areas without human head.Thirdly,the OCR recognition software is used to recognize the segmentation smallest sub-images of text information.These sub-images and recognition results are saved in the prescribed format.Fourthly,this paper uses MFC technology to develop a demonstration software,which is designed and implemented according to the software engineering method,to segment and recognize the image shots of Internet cafes billing system based on Visual Studio 2013 development environment.The main functions of the system include image preprocessing,segmentation of the three functional areas of screenshot image,each functional area within the text line segmentation and recognition,image segmentation,mouse screenshot segmentation,text segmentation and recognition information etc.The effectiveness of the proposed segmentation image segmentation model and related algorithms are verified by using this system..At the end of this paper,the research and application work are summarized,the existing problems and shortcomings are given,and the future research and improvement direction are expected.