Research On Partial Table Recognition System Based On Data Color Specificity

At present,many of the handwritten form information needs to be entered and stored into the computer.In the education system,during the process of dealing with the table,there are some problems as follows:(1)the improper use of human resources resulted in the waste of human resource.There was a large part of the highly educated and high-quality researchers who are doing this repeated work in the education system,which no doubt caused a lot of waste of human resources on the community and the country.The research of this thesis will help to reduce the waste of human resources.(2)A large number of small points score to process will cause the calculation of fatigue,calculation of errors.Long time to revise the papers have produced mental and physical fatigue,in which case the extra sore calculation was prone to errors and low efficiency.In order to solve the above problems,this thesis considers the recognition of table images from two aspects including the detection of table line cell and identification of the data in the table.Firstly,current situation about the table recognition related technology research from our country to foreign is analyzed and reviewed after reading a lot of literature and studying the related technology of the table recognition.Secondly,the score statistics table is used to study as object for recognition.The characteristics of the table are analyzed,and the basic idea of table recognition is determined according to the particularity of data color.Thirdly,the method of the image predictly preprocessing to obtain the clear picture and the method of straight line detection to obtain the position of the table cell were applied in this thesis.Fourthly,according to the particularity of the color,the data in the table is extracted directly and by using the technology of segmentation,to achieve the data picture and get a single digital picture.Due to the particularity of the data extraction,the boundary between the data is blurred and the distinction is not clear.In this thesis,we put forward algorithm of digital region diffusion boundary definition to conform the distinction of each cell data.The two-digit data character is segmented by the drip falling get a single digital picture,and the OCR technique is used to identify the single digit and calculate the cell data.Finally,according to the location of the cell,the identification data and the cell sort association algorithm are came up with to store or output the identified table data.In this thesis,I used the Microsoft Visual Studio development platform and Emgu CV to achieve the score plus point table recognition system prototype development,and calculate the total score.It is proved that the related methods proposed in this thesis have certain feasibility for the identification of special color forms.The development of this system has some value to practice.In this thesis,I used the Microsoft Visual Studio development platform and Emgu CV to achieve the score plus point table recognition system prototype development,and calculate the total score.It is proved that the related methods proposed in this thesis have certain feasibility for the identification of special color forms.The development of this system has some value to practice.