A Method of Form Image Detection Based on Projection Feature and Structure Feature
A technology of structural features and images, applied in the directions of instruments, calculations, character and pattern recognition, etc., can solve the problems that cannot be effectively solved and the accuracy of table detection is reduced, so as to speed up the detection speed, avoid image preprocessing and other operations, increase adaptive effect
Active Publication Date: 2019-02-22
COMMUNICATION UNIVERSITY OF CHINA
View PDF4 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
The existing table image detection method based on projection and structural features does not ignore the loss of the image during the acquisition process, but recovers this part of the loss by performing preprocessing operations such as tilt correction and breakpoint repair on the image, which cannot Effectively solve the problem of reduced accuracy of table detection caused by slightly tilted images, slightly blurred images, slight breakpoints, etc.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0027] The present invention will be further described below in conjunction with accompanying drawing:
[0028] The technical scheme of this embodiment is as figure 1 Shown:
[0029] The image database used in this implementation example is a document image database scanned by a scanner, and the text image to be detected can be an image in bmp format (or other formats), which is stored on a local disk. The table image detection method in this implementation example is implemented by VS2012 programming.
[0030] (1) First, the user inputs the text image to be detected, performs grayscale transformation and global binarization on the input text image to be detected, and sets the global binarization threshold to 220. Horizontally project the binarized image to obtain a horizontal projection histogram A = { a i},in: a i for A First i value, i The values are 0,1,2,…, L -1, L for A The length of is also the height of the text image to be detected.
[0031] The wirele...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention is a method for detecting table images based on projection features and structural features. This method is to perform grayscale transformation, binarization, and horizontal projection on any text image, and calculate the number of straight lines that are close to the actual width of the image and the distribution of straight lines that are close to the maximum peak value in the image through the horizontal projection histogram, and judge whether it is wired or not. Table; if it is not a wired table, perform horizontal segmentation, and judge whether it is a wireless table through the height of the word string, the height between the word strings, and the width between adjacent word strings obtained by vertical projection. The present invention only performs simple preprocessing on the image, and when problems such as slight inclination and slight break point occur in the image, the angle of image projection is the same, and the loss generated during the projection process is also the same. Discarding this part of the loss will not interfere with the table detection results, so as to ignore such problems and avoid complex preprocessing operations. The invention can improve the efficiency of form image detection and simplify the operation steps of form image detection.
Description
technical field [0001] The invention relates to a method for detecting table images based on projection features and structural features, belonging to the fields of image processing and image recognition. Background technique [0002] Table image detection based on projection features and structural features is a relatively common method. When performing image detection based on projection features and structural features, the image must first be preprocessed. Preprocessing generally includes binarization, tilt correction, image refinement, etc.; Or detect whether the row intervals between word strings are distributed at equal intervals to determine whether the image is a table image. [0003] Various digital images stored in the computer, due to the noise introduced during the scanning or transmission process (such as scanned e-books), will cause problems such as slight tilting of the image, slight blurring of the image, and slight breakpoints in the image. Also, table im...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/34
CPCG06V30/412G06V10/273G06V10/267
Inventor 黄祥林宋格格杨丽芳陶竹林刘守训
Owner COMMUNICATION UNIVERSITY OF CHINA