Research And Design Of Online Marking System Based On Image Engineering

Artificial marking method adopts centralized water way, lacking of effective supervision mechanisms and can't guarantee the marking quality.Besides,the marking medium is paper, which can't be easily replicated and transported, and it is difficult to store and transport.Online marking, an effective marking way, sweeps away the defects of artificial marking way and has many other advantages, and becomes an alternative to artificial marking.Online marking system has three parts,the data source subsystem, the marking subsystem, and the management subsystem.It is vital to research method to handle the images according to the need of onling marking.It is possible to find new image processing methods with the development of image engineering.This paper firstly introduces working principle of online marking system,and states its superiority to artificial marking,its current development and application.It analyzes the defects of optical recognition system and proposes effective method directing at data source subsystem.It obtains the digital image of paper to store through high speed scanner,and it is convenient to process,replicate and transmit,which is a big advantage compared to OMR.It realizes segmenting by two steps,and recognizes the markers using proposed method.It analyzes pattern recognition and OCR,and discusses the current character recognition method,and gets good recognition results by combining some features to verify accuracy of candidate's information.The marking subsystem can be completed within LAN, beginning shifting from C/S structure to B/S structure with the development of web technology.Web service,emerging in recent years, are web development technology,using URI to identify the application or components.Its resources and binded information can be described in a standard format, and it can be allowed to directly interact with other applications or components through popular internet protocols, providing basis for better development of online marking.The paper dwells the importance and feasibility of this research,es respondents to common quality papers,and continues the topic in terms of promotional work.It proposes reasonable image processing flow from some points including image preprocessing, image segmentation, image information identification and analysis, information storage in this paper.Image preprocessing involves image enhancement and tilt correction.The former enhances image quality using space filting,and the latter uses the improved hough transformation,preparing for the follow-up image segmentation and recognition.It obtains target images though two steps and employs dynamic threshold to binarize image, well saving the good features of original image, and avoiding the defects of using global or local threshold.It proposes some methods to analyze the object areas in this paper,including improvement on average gray value feature extraction,new recognition method based on SIFT and so on,and they all have good recognition results.OpenCV, as an open computer vision library, contains many image processing algorithms and brings some convenience to our research.In this paper,we design a simple scanning and marking system with the help of hardware and software.The system not only avoids some shortcomings of OMR, but also recognizes different type of paper, and has more flexiblity to meet practical needs.It has friendly interface, simple procession, high accuracy, timely storement and so on.All these have extended the application scope effectively, and prepared for the emergence of low-cost, effective, and intelligent online marking technology. The final chapter briefly summarizes the major work of this paper and makes the further work plan for online marking.