WO2006032184A1

WO2006032184A1 - A scanning method for image photographing

Info

Publication number: WO2006032184A1
Application number: PCT/CN2005/001270
Authority: WO
Inventors: Fuchen Wang; Yan Lu; Huaguo Wang
Original assignee: Fuchen Wang
Priority date: 2004-09-20
Filing date: 2005-08-16
Publication date: 2006-03-30
Also published as: CN1635777A; CN100553281C

Abstract

The invention discloses a photograph scanning method for scanning texts and pictures and the device thereof. According to the invention, images of the texts or pictures are obtained through digital cameras and distortions of these images are calibrated according to a distortion rule, then these calibrated images are tiled based on their corresponding positions. These digital cameras are advanced digital cameras or special bare digital cameras designed for scanning which include a set of lens, a CCD, a circuit, a casing and some simple necessary components. The invention also relates to a standard pattern designed for obtaining the distortion rule of images, a method for calibrating these distorted images, a method for tiling these calibrated images, which has advantages of convenient usage, high scanning definition, multiple functions low maintenance cost, and can be upgraded.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a graphic digitization-scanning method for text image scanning, and more particularly to a text image scanning method using digital photographing technology. It belongs to the international patent classification G06K9 I 20" image capture" technology field. BACKGROUND OF THE INVENTION Current scanners (collectively referred to herein as ordinary scanners) use a one-dimensional array of image sensors (CCDs) for pattern scanning, that is, the sensors are arranged in a row, one line at a time (width about Ιθ μ πι ), as the sensor and the position of the scanned image change, the scanner scans the line in one line, and the entire graphic is scanned. There are two ways to change the position of the sensor and the graphic. One is that the drawing does not move and the sensor moves—the flatbed scanner, the other is the sensor does not move and the drawing moves—the drum scanner.

1 Due to the relative movement of the drawing and the sensor, to ensure the accuracy of the scanning pattern, the processing and installation of the scanner's transmission (including stepper motor, belt, pulley, gear, guide rail, etc.), the sensor arrangement must be very delicate. The faster the scanner, the larger the scan pattern and the higher the accuracy required. Therefore, the cost and price of large format scanners and fast document graphics scanners (currently available in Α4 and A3 formats) are very high (up to 600,000 yuan for fast document graphics scanners; 600,000 yuan for A1 format flatbed scanners; 62 "The format roller scanner is up to 1.3 million yuan; large-format roller scanners over 2m wide will be more expensive because of higher precision requirements." 2 small-format low-range scanners are cheap, but slow and can't scan large pictures; The document graphics scanner can only scan the A4 and A3 format graphics, nor can it scan large images: Large format scanners are expensive and have limited lifespan. It is not possible to use them for long-term file scanning, even if it is occasionally used (relative to small size). Face scanners. There are limitations in the application of various scanners: both scanners and two scanners must be purchased to scan large images and quickly scan a large number of files. 3Because ordinary scanners have to adjust the relative position of the sensor and the drawing. Mobile devices, long-term use will cause wear, and mobile devices have a great impact on scanning accuracy, so the wear of mobile devices will scan the image accuracy - ordinary sweep Instrument accuracy will decrease with long-term use for large format .④ drum scanner, with great difficulty generating jammed sheet is not inserted, especially for old drawing - large format scanner drum vulnerable Drawings. 5 For large-format roller scanners, because the thickness of the paper is different, the finish is different (the friction coefficient is different), so the paper feed speed is not the same when the speed of the platen roller is the same - the roller scanner has a certain error. 6 Because of the high price and the attributes of precision instruments, large-format roller scanners need to be kept in a safe place, which limits more people to master graphic scanning technology. 7 Because of the high price and single function, many users who need large format scanners and fast file graphics scanners are unable to purchase, which reduces the popularity of such devices and slows down the digitization process. DISCLOSURE OF THE INVENTION The object of the present invention is to make up for the deficiencies in the prior art, and to provide a photographic imaging scanning method which is convenient to use, has high scanning precision, is versatile, and has low maintenance cost.

Research on photographic imaging scanning:

With the development of digital photography technology, high-definition, high-performance, high-resolution (12 million pixels) digital cameras have now dropped below 5,000 yuan. If such high-performance cameras are used for graphics and file scanning, the application level of digital cameras can be improved, and the cost of fast file scanning and large-format graphics scanning devices can be greatly reduced, thereby improving large-format graphics scanning devices and fast file scanning devices. The purchasing power, which will promote the development of the digital process to a certain extent.

However, the photos taken by any photographic equipment are subject to certain deformation and cannot be used directly for graphic scanning. For example, Figure 1 is a standard grid diagram (the standard grid diagram can also be composed of points or certain markers arranged in a regular pattern, but the line effect is better. The grid lines can be arranged horizontally and vertically; or they can be arranged in other ways. Like the polar coordinate system formed by the heart circle and the concentric ray; the grid spacing can be the same or different), Figure 2 is a photograph of the standard grid diagram, and the deformation of the photo can be clearly seen. Due to mechanical errors and assembly errors, the distortion of the photos taken by each camera is different. When the relative position and optical parameters of the camera and the drawing are different, the deformation law of the photos taken by the same camera is different.

Photographic imaging principle:

It has been found through practice that in the case of the immovable drawings, the position of the camera and the optical parameters, the images obtained by repeated photographing have the same amount of deformation, and their deformation laws are the same, which is determined by the imaging law of the optical lens. Due to the law of photographic imaging, in the case where the position and optical parameters of the digital camera are the same and the position of the drawing is the same, the deformation law of the other image can be known from the deformation law of one of the images. Therefore, each image can be subjected to image restoration transformation according to its deformation law. The standard mesh pattern shown in FIG. 1 is photographed to obtain the deformation law under the condition, and the image 2 is restored according to the deformation law to obtain the image shown in FIG. 3, and the image 4 is restored to obtain the image 5.

Due to the limited resolution and resolution of digital cameras, the resolution of a 12-megapixel digital camera with a 400mm X 300mm image is 10 dots per millimeter. For large format drawings, if only one photo is taken, the resolution is too low and it is not easy to distinguish the small pixels of the drawing. Therefore, according to the accuracy requirements and the size of the drawing, the drawing should be divided into M rows and N columns to take pictures, and an image array of M rows and N columns is formed; each picture is converted and restored according to its deformation law; , remove the overlap between each picture and other pictures, so that the image arrays of M rows and N columns are put together, there is no overlap, and no image is lost.

The above is the principle of photographic imaging scanning.

The technical solution of the present invention is as follows.

A photographic imaging scanning method for scanning text and images, which uses a digital camera to take pictures, and then corrects image distortion according to the deformation law of the pictures, and finally assembles the pictures according to the relative positions of the pictures.

The digital camera includes a general high-end camera and a bare-metal camera designed specifically for graphic scanning, that is, a camera having only one lens, a charge coupled device CCD, a set of circuits, a simple housing, and a necessary component.

The invention includes a standard drawing plate designed to obtain a photo deformation, and the standard drawing plate may be composed of grid lines, or may be composed of points arranged according to a certain regularity or some kind of mark, the former effect is better; the grid line or a certain The markings may be arranged horizontally or vertically, or may be arranged in other ways, such as a polar coordinate system composed of a heart circle and a concentric ray; the grid spacing may be the same or different.

The invention includes a method of restoring the picture taken to correct the picture distortion: Designing the standard board, taking the same conditions - camera optical parameters, position and lighting conditions, shooting the standard board and the graphics to be scanned, according to the standard board The deformation law of the photo determines the deformation law of other pictures, and corrects the deformation of the picture according to the deformation law of each picture.

The invention includes a method of splicing the corrected picture: dividing a large picture into a matrix of M rows and N columns by grid, forming M rows and N columns, each picture should completely contain the corresponding area graphic, After each picture is corrected, only the image in the corresponding area of the picture is retained, and then the corresponding relationship is pressed. Assembled into the entire drawing.

The scanning device of the photographic imaging scanning method of the present invention, as shown in FIG. 6, is provided with a digital camera of the same model number 1 to 12 on the frame, and the shutter lines 13 to 18 are disposed between the cameras of the same row, and the shutter line is disposed on the frame. The concentrating button 19 is provided with a drawing board 21 on which the drawing is placed in the fixing frame 20 of the apparatus, and a standard pattern of a standard grid is printed on the drawing board 21.

The drawing board 21 is printed with a standard pattern of a standard grid, the horizontal and vertical thin lines are used to determine the deformation parameters of the photos taken by each digital camera, and the horizontal and vertical thick lines are used to divide the effective image of each picture; The image is the No. 1 picture. The image in the thick line range of the 1 ' area of the picture is the effective area of the picture No. 1. The picture No. 1 will contain all the images in the thick line range of the Γ area, and the picture will be restored. After the image is stitched, only the effective area of the picture No. 1 is retained; and so on, the other pictures are processed in the same way, that is, the assembly of one drawing can be completed.

Secure multiple digital cameras to the frame in a defined arrangement and ensure that there are no small changes in camera position.

Each camera is controlled by a shutter release that can be used to focus on a centralized button for scanning large drawings, or it can be used separately to scan small-format graphic text.

The comparison between the photographic imaging scanning method and the ordinary scanner of the present invention is as follows - 1. Easy to use:

1 can design a mobile scanning device using the present invention, by which a graphic file scan can be performed during a business trip;

2 flat-panel equipment, drawings can be laid flat, drawings are easy to install;

8 can scan the image of the thick plate and the uneven image;

4 When scanning graphics, you don't need to connect to the computer, no special interface is needed, the computer graphics resources are not used when scanning graphics, and there is no computer scanning interface when scanning the device. The universal string that is commonly configured by digital cameras when inputting graphics is input. The line bus (USB) interface can be input to the computer;

5 When the lighting conditions are good, the power can also scan the map;

6 Large-format scanning equipment can be provided according to user's requirements (due to the characteristics of this scanning method, large-format scanning equipment only needs to increase a certain number of digital cameras and increase the area of the scanning table, but does not affect the scanning accuracy, so the cost increase is small. ). 2. High scanning accuracy - The drawing does not move during scanning, and the camera is fixed when taking pictures, so theoretically the scanning result can reach the accuracy of the standard drawing board.

3. Multi-purpose

The system's main hardware is a digital camera that can be used for scanning, camera and photography. It can be used for both office and entertainment. A photographic imaging scanner can be used for large-format drawing scanning or for fast document scanning (equivalent to multiple sets of small images). Face fast scanner).

4. Low maintenance costs and convenience:

Since the main hardware of a photographic imaging scanner is low in cost and consists of multiple cameras, it is still possible to use a camera (with an alternative camera) after a camera failure. The cost of repairing the camera is much lower than that of a normal scanner. cost.

5. Can be upgraded:

Due to the low cost of the main hardware digital cameras of photographic imaging scanners, with the upgrade of digital cameras, photographic imaging scanners can also be upgraded.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a standard grid diagram;

Figure 2 is a photograph of a standard grid diagram;

Figure 3 is a photo of the restored standard grid map;

Figure 4 is the same photo as the photo shooting environment of Figure 2;

Figure 5 is a photograph after restoration using the deformation parameters of the photograph of Figure 2;

Figure 6 is a schematic diagram of a fixed platform type photographic imaging scanning device;

Figure 7 is a schematic diagram of the scanner device after the installation drawings;

Figure 8 is a mosaic view of the effective image after the photo taken by each camera is restored;

Figure 9 shows the target area taken for picture number 1;

Figure 10 is the picture number 1. BEST MODE FOR CARRYING OUT THE INVENTION The photographic imaging scanning method of the present invention: 1 uses a digital camera (including bare metal: specifically designed for graphic scanning) The digital camera has a set of lenses, a charge coupled device CCD, a set of circuits, and a simple housing. The cost will be lower, the scanning effect is better) to take pictures; 2 correct the picture deformation according to the deformation law of each picture - image restoration; 3 assemble the picture according to the relative position of each picture (assemble the restored picture).

As shown in Fig. 6, the device is a structure for realizing a photographic imaging scanning method. In Figure 6, 1~

The 12th number is the same type of digital camera, the 13-18 is the digital camera's shutter release, and the 19th is the shutter release button (if the camera's shutter release is connected to the button, press the button Each camera can be photographed at the same time. The shutter of each camera can also be used separately, so that it can be photographed separately, No. 20 is the fixed frame of the device, and No. 21 is the drawing board (also the standard graphic printed with the standard grid), camera and The board is fixed on frame 20. On the 21st board, the horizontal and vertical lines are used to determine the deformation parameters of the photos taken by each digital camera, and the horizontal and vertical lines are used to divide the effective image of each picture. The image taken by the No. 1 camera is the No. 1 picture. The image in the thick line range of the apostrophe area of the board is the effective area of the No. 1 picture, and the No. 1 picture will contain all the images in the range of the thick line in the number area. When the restored image is spliced, only the effective area of the No. 1 picture is retained. For other pictures, you can complete the assembly of one drawing by the same method.

Figure 6 shows the operation of the photographic imaging scanning device: 1Install the digital camera (the digital camera can be disassembled and installed at will, and can be used for daily photography); 2System initialization - Shoot the board with the standard grid diagram to obtain The deformation law of each digital camera; 3 placing the drawings (see Figure 7) and taking pictures of the original picture with a specific deformation law; 4 inputting the pictures into the computer through the universal serial bus (USB) interface commonly configured by digital cameras and according to their respective The deformation law restores the picture; 5 The effective area of the restored picture is assembled into the specified area of the final picture, see Figure 8. If you scan large-format drawings, the shutter release of each camera should be connected to the centralized button. Press the button once to scan the entire drawing, sweep one image for less than one second; to scan a small size (A4 or A3) The format or file, the shutter release of each camera is used separately, each camera can be scanned separately, one camera can scan 30~60 graphics or files per minute, corresponding to the device shown in Figure 6, if there are 12 people operating , can reach the scanning speed of 300~700 graphics or files per minute.

The method of restoring the deformed image (take the No. 1 picture as an example), as shown in Figure 1: For the No. 1 image, the captured image should completely contain all the images within the thick line of the area. This area consists of a straight line 01-02.

02-03, 03-04, 04-01 circled, only the Ny strip level and Nx stripe used to determine the deformation law Straight grid lines (the text and shadow boxes in the figure are filled in to illustrate the restoration method), see the figure

9. The grid line spacing is D, and the height and width of the area are

H=Y04-Y01=D X (Ny-1)

W=X02—XO 1=D X (Nx- l)

Let the coordinates of 01 point be Χ, Υ (for the 2' area, Χ2 ', Υ2', and so on),

The coordinates of the defect are

Xi,j=Xl, +D X i

Yi, j=Yl, +D Xj Figure 10 is a picture of the apostrophe area, which has only lines that are nearly horizontal and vertical (the text and shadow squares in the figure are backfilled), which correspond to those in Figure 9. The horizontal and vertical lines are now curved. In the thick line range of the picture 10 (including the thick line), the Ny line Nx column intersection points can be determined by the near-vertical curve of the beam and the Ny-bar near-horizontal curve, and these intersection points correspond to the horizontal line and the vertical line in FIG. The intersection. In FIG. 10, the four intersections Q1 to Q4 correspond to the four intersections of Q1 to Q4 in FIG. 9, and the four-point range image (shaded area) can be restored by bilinear transformation, and the bilinear transformation defines the boundary of the region as The straight line segment; in Fig. 10, the nine intersection points of P1~P9 correspond to the nine intersection points of P1~P9 in Fig. 9. The image of the nine intersection points (shaded area) can be restored by double parabola transformation, and the double parabola transformation defines the area. The boundary is a parabola. The transformation makes the nine intersections of P1~P9 in Fig. 10 correspond to the nine intersections of P1~P9 in Fig. 9. Fig. 3 and Fig. 5 are obtained by using the transformation; Fig. 10 a~p16 The intersection points correspond to the sixteen points of a~p in Fig. 9. The double-cubic transformation can be used to recover the image (shaded area) within the four points of m, n, o, p. The bicubic transformation defines the boundary of the area as three times. curve. Both the 9-point transform and the 16-point transform are more accurate than the 4-point transform, but the 16-point transform is more complicated, so a 9-point transform should be used. Industrial applicability

The invention can be applied industrially and has industrial applicability.

Claims

Rights request

1. A photographic imaging scanning method for scanning text and images, characterized in that: a digital camera is used to take a picture, and then the picture is deformed according to the deformation law of the picture, and finally the picture is assembled according to the relative position of each picture.

2. The photographic imaging scanning method according to claim 1, wherein: said digital camera comprises a general high-end camera and a bare metal camera specially designed for graphic scanning, 艮卩: only one set of lenses, one charge coupled Device CCD, a set of circuits, a simple housing and a camera with the necessary components.

3. The photographic imaging scanning method according to claim 1, wherein: the standard drawing plate designed to obtain a photo deformation, the standard drawing plate may be composed of grid lines, or may be arranged by a certain regularity or A certain mark constitutes the former, and the former has better effect; the grid lines or some marks can be arranged horizontally and vertically, or can be arranged in other ways, like a polar coordinate system composed of a heart circle and a concentric ray; the grid spacing can be the same or different .

4. The photographic imaging scanning method according to claim 1, wherein: the method for restoring the taken picture in order to correct the image deformation: designing the standard drawing board under the same conditions - camera optical parameters, position and lighting conditions , shooting standard board and graphics to be scanned, determining the deformation law of other pictures according to the deformation law of the standard picture photos, and correcting the deformation of the picture according to the deformation law of each picture.

5. The photographic imaging scanning method according to claim 1, wherein: splicing the corrected image: dividing a large image into a matrix of M rows and N columns by a grid to form M rows and N columns For each area, each picture should contain the graphic of the corresponding area. After correcting each picture, only the image in the corresponding area of the picture is retained, and then the corresponding figure is assembled into the whole picture.

6. The scanning device of the photographic imaging scanning method according to claim 1, wherein: the same type of digital camera of 1 to 12 is mounted on the frame, and the shutter lines 13 to 18 are disposed between the cameras of the same row. A concentrating button 19 for the shutter release is provided on the frame, and a drawing board 21 on which the drawing is placed is provided in the fixing frame 20 of the apparatus, and a standard pattern of a standard grid is printed on the drawing board 21.

7. The photographic imaging scanning method according to claim 1, wherein: the drawing board 21 is printed with a standard pattern of a standard grid, and the horizontal and vertical thin lines are used to determine the photo taken by each digital camera. Deformation parameters, horizontal and vertical thick lines are used to divide the effective image of each picture; the image taken by the No. 1 camera is the No. 1 picture, and the image in the thick line range of the apostrophe area of the picture is the effective area of the picture No. 1, the picture No. 1 All the images in the range of the thick line in the number area are included, and only the effective area of the picture No. 1 is reserved when the image restored by the picture 1 is spliced; and so on, the other pictures are processed in the same way, that is, one can be completed. The assembly of the drawings.

8. A photographic imaging scanning method and scanning apparatus according to claims 1 and 6, characterized in that: a plurality of digital cameras are fixed to the frame in a certain arrangement and that there is no slight movement change in the position of the camera.

9. The photographic imaging scanning method and scanning apparatus according to claim 1, wherein each camera is controlled by a shutter release, and the shutter release can be concentrated on a centralized button for scanning a large drawing, or can be separately used. Use to scan small format graphic texts.