Online Marking Information Extraction And Recognition Technology

In recent years, the pace of modernization of China's education is speeding up. Standard tests have been applied to several of social talent selection processes. The traditional online test system has not been adopted by schools and other small-scale field applications because of the high cost and technical constraints. This thesis takes auto-marking amending system to replace the OMR machines,which uses digital image processing technology of image recognition software to complete the automatic marking. This method can not only reduce the cost of scoring but also can deep into the subtle field to make up the deficiencies of large-scale online marking system because of high degree of flexibility.The summary of automatic marking is given at the beginning of this thesis, one of the main marking methods is OMR machine, and the other is computer software. Deficiencies and the restriction of application of the two methods are proposed under analysis of principle and process of marking finally. The content of study in this thesis also belongs to computer software marking, but it focuses on how to cope with the task of large quantities handling.Two core elements of this thesis are OMR information card and barcode identification. Barcode image preprocessing technology is discussed in the CODE128 barcode processing system design and implementation. At last a more applicable of image preprocessing flow is given out. Compared to the self-adaptive threshold, the operational efficiency of system is improved with a new method of using a threshold model which builds on the stored threshold and then gets value from the model for image binarization. Finally the thesis introduces two methods of the barcode'width measurement and decoding in the current application: the average distance method and similar-edge distance measurement. All of them have been applied in the barcode processing system. The actual processing effect of barcode processing system is shown lastly.In the design and the implementation of OMR information card processing system, image pre-processing and target point positioning and some other critical technologies are studied in this thesis. Mathematical morphology is used in the image pre-processing to extract and locate the feature of target points. There are also existed two ways for target point positioning: gray-scale projection and the reference point. Advantages, disadvantages and preconditions of the two methods are discussed. Effects of practical work are shown finally, and the thesis analyzes the performance of a variety of corrective methods.In the end of this thesis, a few of areas for improvement and some deficiencies of the two systems are summarized, includes the prospects of future work.