An Intelligent Marking System Combining DOM And Image Engineering

For traditional artificial marking method lacking of effective supervision of the marking process, it is difficult to ensure a fair and impartial examination requirements. Besides, for the marking medium is paper, the work of transportation, store and inspection requires a lot of manpower resources. With the rapid development of computer technology, network technology, image processing and pattern recognition technology and scanner hardware technology, online marking system is an effective solution to solve the traditional marking process factors that affect the impartial. Examination of marking process can be carried out effective supervision, and data analysis and statistical achievements can be done automatically by the computer, which greatly reducing the workload of teachers. For marking medium are electronic papers image, the papers of storing, inspection and transportation are more convenient and efficient.Online marking system has been widely used in various exams. As the existing online marking system which the marking information is cured in the program only read dedicated answer sheets, when the answer sheet style changes, marking system cannot work properly and cannot meet the individual requirements of the examination. Aiming at these problems existing online marking system, for papers with question and answer card intelligent marking system is proposed based on the combination of DOM and image engineering. It focuses on the main supporting method of universal marking system technology and image processing technology. The main work includes:(1)Put forward a filtering algorithm based on gradient correlation. By introduction of adaptive noise judgment mechanism, for the area of the image which is not infected will not be treated. The efficiency of the filtering algorithm is improved;(2)Tilt correction algorithm based on image texture is applied in complex situations of papers with text, images, and information of handwritten. Paper textures in the image can show the direction of the overall characteristics of the image, and the image texture is not affected by light, color and the impact of the brightness, therefore, a skew correction algorithm based on image texture correction is applied in the intelligent marking system. The correction effect is better than the others;(3) Put forward a method of making full use of the location of the area of identify when designing paper. Firstly create the schema tree of papers structure, then automatically generate papers XML description file based on the schema tree. The location information described in the XML file is mapped to image space by coordinate conversion, and acquires the identified and split area of paper; (4) Template matching algorithm based on direction vector is applied in locating the position of option box, which is not impacted of blocked, disconnection, light and other factors.OpenCV library contains more than500image processing algorithms which is widely used to process image. EmguCV wrapper the functions of the OpenCV which makes these versatile image processing algorithms be ported to C#platform to improve the efficiency of systems development and reusability of software. With these generic image processing algorithms, an universal marking system that papers with question and answer card is developed based on image recognition. Experiments and data analysis show that the proposed method can solve the technical problems which exist online marking system that makes papers design more flexible based on users’demand. Based on the above technical support, Intelligent marking system that papers with question and answer card is completed that allows users to define the format and content of the papers, expanding the apply scope of the system. At the end of the article, the research work is summarized, and the inadequacies and vision for the future development of the subject is proposed.