RU2298292C2 - System and method for scanning - Google Patents

Abstract

FIELD: systems and methods for scanning, primarily, objects in binding.

SUBSTANCE: system for scanning primarily bound objects includes, interconnected by data bus: scanner, digital camera, processor, program memory block, block for memorizing given parameters of documents and external displaying device, while objective of camera is facing end side of object, program memory block contains correction means, made with possible determining of parameters and method for image correction, processor has distortion correction module, made with possible analysis of object image, received from camera, and also correction of distortions of scanned image of object close to binding by determining and processing parameters, characterizing shape of bending of object pages close to binding.

EFFECT: high speed correction of geometric distortions of object in binding area with maximal usage of standard office equipment like scanners and digital cameras.

2 cl, 9 dwg

Description

The invention relates to systems and methods for scanning and copying mainly for objects in binding.

When scanning an object in binding, it is practically impossible to obtain a scanned image without distortion of geometric parameters (distortion) and darkening of the text in the binding area. In this regard, the problem of correcting distortions in scanned images, especially in text ones, is currently relevant, in particular for the purpose of subsequent text recognition, since text recognition systems can only work with an undistorted image of text.

When scanning objects in binding, such as books, or objects with an uneven surface, distortions in the scanned image of the object, namely distortion and darkening of the text, occur mainly in the binding area and are caused by a loose fit of the surface of the scanned object to the scanner substrate along which the object is scanned.

There are various ways to eliminate the above distortions that occur during scanning. The methods for eliminating distortions are divided into hardware, in which they affect the scanned object itself, and algorithmic, which are based on the correction of the scanned image of the object.

In hardware methods, various mechanical devices are mainly used, which allow the object to be pressed tightly against the substrate along which it is scanned. There is a risk of damage to the object. In the algorithmic systems and methods to which the claimed invention relates, areas with distortion are determined either by analyzing the scanned image of the object, or based on data from additional sensors that determine the height of the gap between the scanned object and the scanner substrate. Moreover, in many cases, the use of additional sensors requires a double pass of the scanning system along the document, which increases the scanning time.

US Pat. No. 6,330,050 describes a device for correcting distortions arising in the binding area of an open book during scanning. The device comprises a scanning tool, a memory, a distortion correction tool, configured to determine a distortion region in a scanned image of an object in the binding area of a book without using additional distance detectors and image correction. The disadvantage of this device is the low accuracy of determining the area of distortion and corrective ability, in addition, the device involves the use of significant memory resources.

US Published Application No. 2002/0085248 proposes a method for correcting a scanned image in which an uneven object is scanned. The method determines the data of the scanned image of the object, analyzes the data of the scanned image of the object, while determining the areas of distortion that occur in uneven places where the object does not fit snugly on the scanner substrate, and data for adjusting the area of distortion of the image of the object. Then restore the brightness of the image in the distortion region, using the data for correction. This method is selected as a prototype of the claimed invention.

The disadvantage of the prototype is the low accuracy of determining the areas of distortion of the image and the inability to adjust geometric distortions and sharpness of the image of the object that occur in the areas of distortion. In addition, the prototype method does not allow to determine the distortion region of the image of the object and the data for adjusting the distortion region directly during scanning, which increases the execution time of the method.

The objective of the claimed invention is to provide a system and method for scanning and copying mainly for an object in binding, allowing for high-speed correction of geometric distortions of an object in the binding area with the maximum use of standard office equipment such as scanners and digital cameras.

The technical result is achieved through the use of such a system that contains a scanner connected to each other by a data bus, a digital camera, a processor, a program memory block, a data memory block and an external display device, while the scanner has a substrate on which the object is located and along which it is scanned, and the digital camera is located on one of the sides of the scanner in such a way that the camera lens is aimed at the end of the binding at the bend of the opened pages, the program memory contains correction means, you Completed with the ability to set parameters and an image correction method, the processor has a distortion correction module, configured to analyze the image of the object received from the camera, as well as correct distortions of the scanned image of the object near the binding by determining and processing data characterizing the shape of the bending of the pages of the object near the binding in compliance with the parameters and correction method.

For the functioning of the system, it is essential that the scanner is a tablet.

For the system to function, it is important that the scanner backing is made of glass or other transparent material.

For the system to function, it is necessary that the scan object is placed on the substrate in expanded form, and the pages of the object are pressed against the substrate.

For the system to function, it is essential that the scan object is placed on the scanner substrate so that the edges of the object pages are parallel to the borders of the scanner substrate.

For the functioning of the system, it makes sense that at least one lighting device is arranged next to the camera, configured to be turned on during shooting to illuminate the object from the side of the camera.

For the functioning of the system, it is important that the camera is built into the scanner body or connected to the scanner at the time of scanning the object.

For the system to function, it is essential that the optical axis of the camera is located at an angle from 0 to 60 degrees to the plane of the scanner substrate.

For the functioning of the system, it is necessary that, as an external device, the system contains at least one device selected from the group including a monitor, printer, computer, or other similar devices.

The task is also achieved by creating a method of scanning and copying, which includes the following operations:

- remove the end part of the object with a digital camera;

- record the received image of the end part of the object in the data memory block;

- analyze the presence of distortions in the image recording in the processor distortion correction module in accordance with the parameters obtained from the correction means, while determining whether the object has distortions near the binding, if the object has no distortions near the binding, scan the object with a scanner and display the image of the object on an external device, if the object has distortions near the binding, the parameters characterizing the shape of the bending of the pages of the object near the binding are determined and stored in the data memory block;

- scan the object with a scanner;

- write the scanned image of the object in the data memory block;

- correct in the distortion correction module the distortion of the scanned image of the object stored in the data memory unit, in accordance with the parameters and the correction method set in the correction tool, using data characterizing the shape of the bending pages of the object near the binding;

- record the corrected scanned image of the object in the data memory block and output it to an external device.

For the operation of the method, it is essential that the object was shot with a digital camera, while illuminating the object with at least one lighting device from the side of the camera.

It is important for the functioning of the method to a picture object stored in the data memory unit, analyzed in a module processor distortion correction, the injected frame coordinate system of the object captured by the camera X _c Z _c, X-axis _which is parallel to, and Z _c axis - perpendicular the array, points in the frame coordinate system of the object captured by the camera were assigned an array of points on the scanner substrate that were equally spaced from each other, and the coordinates {x _c (i)} of each array point on the scanner substrate were determined for all points in the array with coordinate we {x _c (i)} determined the array of distances {h (i)} from each point on the scanner substrate z _c = 0 to its projection on the object page, as a result, we obtained an array of pairs of points {x _c (i), h (i )}, we determined the coordinate x _с0 corresponding to the maximum of the array of distances (the abscissa of the maximum):

x _c0 : h (x _c0 ) = max {h (i)},

analyzed the nature of the array {x _c (i), h (i)} to the left and right of x _c0 and determined whether the object has distortions near the binding or not.

For the method to function, it is essential that if the object has distortions near the binding, the data characterizing the shape of the bending of the object pages near the binding is determined, while the array {x _c (i), h (i)} is converted into two coordinate arrays {x _c (j) , z _ca (j)} and {x _c (j), z _cb (j)}, introduced the scanner coordinate system X _s Z _s whose axes coincide with the axes of the coordinate system of the frame of the object captured by the camera X _c Z _c , coordinate arrays {x _c (j), z _ca (j)} and {x _c (j), z _cb (j)} converted (converted) into the scanner coordinate system X _s Z _s :

согласует разрешение камеры и сканера по оси X, d согласует разрешение камеры и сканера по оси Z, а также учитывали угол наклона камеры к подложке, где tx переносит начало системы координат, в результате получали массивы координат {x_sa(j), z_sa(j)} из n элементов и {x_sb(j), z_sb(j)} из m элементов, определяющие форму изгиба страниц объекта вблизи переплета.Where

matches the resolution of the camera and scanner along the X axis, d matches the resolution of the camera and scanner along the Z axis, and also took into account the angle of inclination of the camera to the substrate, where tx transfers the origin of the coordinate system, as a result we got coordinate arrays {x _sa (j), z _sa ( j)} of n elements and {x _sb (j), z _sb (j)} of m elements, which determine the form of bending of the pages of the object near the binding.

For the method to function, it is important that the distortion correction module of the processor corrects the distortion of the scanned image of the object, while correcting the distortion and shadow in the bending area of the pages of the image of the object.

For the method to function, it is necessary that the distortion correction module corrects the distortions of the scanned image of the object, using piecewise linear interpolation over the coordinate arrays {x _sa (j), z _sa (j)} and {x _sb (j), z _sb ( j)}, it was believed that the number of pixels of each row of the corrected scanned image of the object exceeds the number of pixels of the scanned image by the value of Delta, which was specified by the following dependence:

Delta = round (L _a + L _b - (x _sb (m) -x _sa (n))),

where round is the function of rounding to the nearest integer, L _a and L _b are the total lengths of the interpolating segments, which corresponds to the number of pixels of the curved parts of the pages, while L _a and L _{b are} given by the following dependence:

to correct distortion along the interpolated curve formed by the cross section of the curved surface of the page of the object, the samples were selected in increments of a unit length of the scanner coordinate system, as a result of which the corresponding coordinate x _s 'was determined, from which the color components of R'G'B' were calculated from the scanned image corrected image by linear interpolation according to the values of the color components of the RGB pixels of the scanned image with the coordinates floor (x _s ') and floor (x _s ') +1, where floor is the function of taking the integer part:

R '= R (floor (x _s '), y) + (x _s ' -floor (x _s ')) * R (floor (x _s ') + 1, y),

for the values of G 'and B' was calculated similarly, the current coordinates of the samples (x _s ', z _s ') along the curve interpolated by the segments are calculated according to the parametric equations of the segments of the straight line from the point with coordinates (x _sa (i), z _sa (i)) to (x _sa (i-1), z _sa (i-1)), i = 2 ... n:

x _s ' = x _sa (i) + t * (x _sa (i-1) -x _sa (i)),

z _s ' = z _sa (i) + t * (z _sa (i-1) -z _sa (i)),

where t = 0 at the beginning of the segment (point (x _sa (i), z _sa (i)), t = 1 at the end of the segment (point (x _sa (i-1), z _sa (i-1)), and the change step t was set by the expression:

the current coordinates of the samples (x _s ', z _s ') were calculated along the interpolated curve for the array {x _sb (j), z _sb (j)} in the same way, the experimentally averaged dependence of the brightness of the distorted pixel on the brightness of the original pixel and the distance to the projection of this pixels on the page from the scanner substrate, then this dependence was converted into a distortion correction function of the scanned image of the object, which also takes into account the scanning resolution:

Y '(x, y) = f (Y (x _s ', y), z _s '),

where Y '(x, y) is the brightness of the adjusted pixel, Y (x _s ', y) is the pixel brightness of the scanned image, z _s 'is the distance to the page from the scanner substrate in x _s '.

For the method to function, it is essential that the distortion module corrects the distortions of the scanned image of the object, while correcting the distortions in each line of the scanned image, regardless of the other lines.

For a deeper understanding of the claimed invention, the following is a detailed description and drawings.

Figure 1. Block diagram of a scanning and copying system according to the invention.

Figure 2. A possible appearance of the scanning and copying system according to the invention.

Figure 3. A block diagram of the main steps of a scanning and copying method according to the invention.

Figure 4. The binding scheme of the object in the coordinate system of the camera according to the invention.

Figure 5. A block diagram of the steps of analyzing an object frame according to the invention.

6. Image of a fragment of book pages near the binding in the scanner coordinate system.

7. Graph of background brightness for a book page near the binding according to the invention.

Fig. 8. The block diagram of the correction of the scanned image, namely the correction of distortion and shadow for each scanned line in the bending area of the page according to the invention.

Fig.9. An example of the correction of an object according to the invention:

a) a scanned fragment of a page of a book without correction;

b) a scanned fragment of a book page with correction of distortion only;

c) a scanned fragment of a book page with correction for distortion and shadow.

Figure 1 shows a block diagram of a scanning and copying system according to the invention. The scanning and copying system comprises a scanner 1, a camera 2, a processor 3, comprising an image correction module 4, a document parameter data memory unit 5, a program memory unit 6 containing the correction means 7, an external device 8 and a data bus 9. The operation of the system is controlled by the processor 3. Camera 2 is located on one of the sides of the scanner 1, and the lens of camera 2 is located on the side of the object, directly opposite the end of the binding. In this case, the camera 2 is configured to obtain a snapshot of the object in the bending area of the pages near the binding. The correction means 7 is configured to set parameters and an image correction method. The distortion correction module 4 is configured to analyze the image of the object and correct distortions of the scanned image of the object. Data exchange between the scanner 1, camera 2, processor 3, data memory unit 5, program memory unit 6 and external device 8 is carried out via data bus 9. When scanning, processing can be performed for a single image line or for several lines.

Figure 2 illustrates a possible variant of the appearance of the scanning and copying system and shows the relative position of the substrate 10, camera 2 and object 11. The scanning object 11 is placed on the substrate 10 of the flatbed scanner 1 in the expanded form with the pages facing down, and the pages of the object 11 are pressed against the substrate 10. Moreover, the object 11 is placed on the substrate 10 with the end face of the binding opposite the mark 12 on the scanner body 1, and the pages of the object 11 are parallel to the borders of the substrate 10. To facilitate positioning on the borders of the substrate and on the body, the scanner is nano Mark 12 or use delimiters. On one of the sides of the scanner 1, behind the mark 12, there is a camera 2. In this version of the system, the camera 2 is a black and white or low-resolution color digital camera. The field of view of the camera 2 is such that the entire bending area of the pages near the binding of the object 11 falls on the image of the camera 2. The horizontal image size of the camera 10-12 cm is sufficient for books and similar bound objects up to A3 format. The angle of inclination of the camera 2 to the plane of the substrate 10 of the scanner 1 is set in the range from 0 to 60 degrees and depends on the optical system of the camera 2 and the design of the appearance of the system as a whole. The camera 2 is located on one of the sides of the scanner 1, and the lens of the camera 2 is located on the side of the object, along the binding line. Camera 2 can be integrated into the scanner 1 housing or connected at the time of scanning.

It is important to note that the use of camera 2 is one of the simplest methods for the average estimate of the bending of the pages of the object 11 near the binding. It will be shown below that this technical solution allows the correction of geometric and luminance distortions that occur when scanning intertwined objects.

A illuminator 13 is located on the same side of the system as the camera 2, which is turned on during the shooting of the object 11 by the camera 2 and provides such lighting that the lateral (i.e., end) part of the object 11 is illuminated and the rest remains dark. It is possible, however, and another solution to the problem of lighting, providing a pronounced contrast between the end part of the object and the rest of its parts. The scanning head 14 of the scanner 1 moves along the Y _s axis parallel to the substrate 10 of the scanner 1. From above, the object 11 on the substrate 10 is covered with a lid 15 or otherwise presses the object 11 against the substrate 10 of the scanner 1. The corrected scanned image of the object 11 is output to an external device 8, which in this embodiment of the system, a printer is integrated in the scanner body 1. FIG. 1 shows an output tray 14 of a printing apparatus.

Figure 1 also indicates the coordinate system used in the description below:

X _s Y _s Z _s - coordinate system of the scanned image of the object;

X _c Z _c - coordinate system of the frame of the object captured by the camera;

X _p Y _p is the coordinate system of the printed corrected image of the object.

Figure 3 shows a block diagram of the main steps of the method of scanning and copying objects in binding according to the invention. The user orientates the object 11 on the substrate 10 of the scanner 1, closes the lid 15 of the scanner 1 and gives the command to scan the object 11. Then turn on the side illuminator 13, shoot the object 11 with the camera 2, and a picture of the object 11 is obtained (step 1). Record the received snapshot of the object in block 5 of memory 0 data. Analyze in the distortion correction module 4 of the processor 3 the presence of distortions in the image of the object 11 from the data memory unit 5 in accordance with the parameters obtained from the correction means 7. In this case, it is determined whether the object 11 has distortions near the binding (step 2). If the object has distortions near the binding, the parameters characterizing the bending shape of the pages of the object 11 near the binding are determined and stored in the data storage unit 5 (step 3). Then scan the object 11 with the scanner 1 (step 4). The scanned image of the object 11 is recorded in the data memory unit 5. The distortions of the scanned image of the object from the data memory unit 5 are determined and corrected in the module 4 in accordance with the parameters and the correction method set in the correction means 7 (step 5). This uses data characterizing the shape of the bending pages of the object 11 near the binding. As a result, the corrected scanned image of the object 11 is recorded in the data memory unit 5 and output to an external device 8 (step 6), for example, a printer, and an adjusted copy of the object is obtained. If the object 11 does not have distortions near the binding, scan the object 11 with the scanner 1 (step 4) and output the scanned image of the object 11 to the external device 8 (step 6).

5 is a flowchart of steps for analyzing an object frame according to the invention. In the block diagram, the notation adopted earlier in FIG. 4 is used. At the first stage, the coordinate system of the frame of the object captured by the camera X _c Z _c is introduced, the X axis _with which is directed parallel to and the Z _c axis is perpendicular to the substrate, in the coordinate system of the frame of the object captured by the camera, an array of points on the scanner substrate are set apart each other at an equal distance, and determine the coordinates {x _c (i)} of each point in the array on the scanner substrate. For all points of the array with coordinates (x _c (i)}, an array of distances {h (i)} from each point on the scanner substrate z _c = 0 to its projection on the object page is determined. As a result, an array of pairs of points {x _c (i ), h (i)}. The coordinate x _с0 corresponding to the maximum of the array of distances (the abscissa of the maximum) is determined:

They analyze the nature of the array {x _c (i), h (i)} to the left and to the right of x _с0 and make a conclusion about whether the object has distortions near the binding or not. If the object does not have distortions near the binding, then the scanned image of the object is not corrected. If the object has distortions near the binding, the array {x _c (i), h (i)} is converted into two coordinate arrays (x _c (j), z _ca (j)} and {x _c (j), z _cb (j)}, as shown in Figure 4. Enter the coordinate system of the scanner X _s Z _s , the axes of which coincide with the axes of the coordinate system of the frame of the object captured by the camera X _c Z _c . Arrays of coordinates {x _c (j), z _ca (j)} and {x _c (j), z _cb (j)} convert (convert) to the scanner coordinate system X _s Z _s :

согласует разрешение камеры и сканера по оси X; d согласует разрешение камеры и сканера по оси Z, а также учитывают угол наклона камеры к подложке; tx переносит начало системы координат. В результате получают массивы координат

из n элементов и {x_sb(j), z_sb(j)} из m элементов, определяющие форму изгиба страницы объекта. Эти массивы используются для коррекции дисторсии и тени.Where

Coordinates the resolution of the camera and scanner along the X axis; d matches the resolution of the camera and scanner along the Z axis, and also takes into account the angle of inclination of the camera to the substrate; tx transfers the origin of the coordinate system. The result is an array of coordinates

of n elements and {x _sb (j), z _sb (j)} of m elements that determine the shape of the page bend of the object. These arrays are used to correct distortion and shadow.

имеют геометрические (дисторсия) и яркостные/цветовые (тень) искажения. В то же время пикселы вне этого диапазона не искажены и переносятся в результирующее изображение без преобразований.Consider a schematic image of the page of the object without correction in the coordinate system of the scanner (Fig.6). The pixel values of each line of the scanned image in the range

have geometric (distortion) and brightness / color (shadow) distortions. At the same time, pixels outside this range are not distorted and are transferred to the resulting image without transformations.

и

. Количество пикселов каждой строки скорректированного отсканированного изображения объекта превосходит количество пикселов отсканированного изображения на Delta:The reason for the distortion is that the image of the curved surface of the object page is projected onto the plane of the scanner substrate. With this orientation of the object and the scanning head, distortion occurs only along the X _s axis; therefore, each line of the scanned image can be adjusted independently of other lines. To restore the function of bending the pages of an object during correction of a scanned image, piecewise linear interpolation using coordinate arrays is used

and

. The number of pixels of each line of the adjusted scanned image of the object exceeds the number of pixels of the scanned image in Delta:

и

- суммарные длины интерполирующих отрезков, что соответствует количеству пикселов изогнутых частей страниц:where round is the function of rounding to the nearest integer,

and

- the total lengths of the interpolating segments, which corresponds to the number of pixels of the curved parts of the pages:

,

,

To correct distortion along the interpolated curve formed by the cross section of the curved surface of the page of the object, the samples are selected in increments of a unit length of the scanner coordinate system, as a result of which the corresponding coordinate x _s 'is determined, from which the values of the color components R'G'B' are calculated from the scanned image corrected image by linear interpolation according to the values of the color components of the RGB pixels of the scanned image with the coordinates floor (x _s ') and floor (x _s ') +1, where floor is the function of taking the integer part:

for G 'and B' is similar.

до

i=2...n:The current coordinates of the samples (x _s ', z _s ') along the curve interpolated by the segments are calculated by the parametric equations of the segments of the straight line from the point with the coordinates

before

i = 2 ... n:

, t=1 в конце отрезка (точка

шаг изменения t:where t = 0 at the beginning of the segment (point

, t = 1 at the end of the segment (point

change step t:

The calculation of the current coordinates of the samples (x _s ', z _s ') along the interpolated curve for the array {x _sb (j), z _sb (j)} is carried out similarly.

In the general case, the nature of the brightness and color distortions in the bending area of the page has a complex character, determined by many factors. A typical effect is the darkening of the image (shadowing) near the binding of the book. The graph of the brightness of the background page of the object for this area (along the segment AO of Fig.6) is shown in Fig.7. It was experimentally found that there is an average dependence of the brightness of the distorted pixel on the brightness of the original pixel and the distance to the projection of this pixel on the page from the scanner substrate. Next, this dependence is converted into a function that also takes into account the resolution of the scan:

where Y '(x, y) is the brightness of the corrected pixel, Y (x _s ', y) is the brightness of the pixel of the scanned image, z _s 'is the distance to the page from the scanner substrate in x _s '.

One of the simplest versions of f (Y (x _s ', y), z _s ') can be the following function:

where the brightness values are normalized in the range from 0 to 1, the function is measured for a resolution of 200 dpi and a specific scanner model. This function was used to correct the image from Fig. 9a; the correction result is shown in Fig. 9c. For each specific scanner model, this dependence can be measured previously.

To obtain the brightness value Y R'G'B '- the components of the scanned image after distortion correction are converted into a color system, one of the channels of which is brightness, and the other two encode color. It is effective to use linear conversion to a YIQ or YCbCr system, but more complex transformations, for example, to HSB or L * a * b * color systems, can be used within the framework of the present invention. After the correction of the brightness value Y-> Y ', the inverse is converted to the RGB color system, the other two channels encoding the color are not changed.

. На шаге 1' инициализируется счетчик цикла i для перебора интерполирующих отрезков прямой, построенных по массивам координат

и устанавливается начальное значение для переменной npix - счетчика количества скорректированных пикселов. На шаге 2' в соответствии с формулой (7) определяется величина dt, равная единице длины системы координат сканера, и переменная nsteps, равная количеству отсчетов, дискретизирующих текущий отрезок. На шаге 3' инициализируется счетчик цикла k для перебора отсчетов на отрезке, и для параметра прямой t устанавливается нулевое значение, которое соответствует началу отрезка. На шаге 4' отсчеты выбираются в соответствии с параметрическим уравнением прямой (6) и определяются значения (x_s', z_s'). На шаге 5' вычисляется значение индекса в массиве скорректированных пикселов. На шаге 6' из сканированного изображения вычисляются значения цветовых составляющих R'G'B' скорректированного изображения путем линейной интерполяции (5) по значениям цветовых составляющих пикселов RGB сканированного изображения с координатами floor(x_s') и floor(x_s')+1. На шаге 7' выполняется конвертирование значений цветовых координат текущего пиксела R'G'B' в цветовое пространство YCbCr. На шаге 8' осуществляется коррекция яркостных (теневых) искажений. Обратное преобразование из цветового пространства YCbCr, где значение яркости Y уже скорректировано на предыдущем шаге, выполняется на шаге 9'. На шаге 10' увеличиваются на единицу количество скорректированных пикселов и число взятых на отрезке отсчетов. При этом на величину dt увеличивается параметр прямой t. Шаги с 4' по 10' выполняются в цикле пока условие 11' истинно.On Fig shows a block diagram of the correction of distortion and shadow for each scanned line with coordinate y in the range of page bending

. At step 1 ', the loop counter i is initialized to iterate over the interpolating line segments constructed from the coordinate arrays

and sets the initial value for the variable npix - the counter of the number of adjusted pixels. In step 2 ', in accordance with formula (7), the quantity dt is determined equal to the unit length of the scanner coordinate system and the variable nsteps is equal to the number of samples discretizing the current segment. At step 3 ', the loop counter k is initialized to iterate over the samples on the segment, and for the parameter of the line t, a zero value is set that corresponds to the beginning of the segment. At step 4 ', the samples are selected in accordance with the parametric equation of the straight line (6) and the values (x _s ', z _s ') are determined. At step 5 ', the index value in the corrected pixel array is calculated. In step 6 ', the values of the color components R'G'B' of the corrected image are calculated from the scanned image by linear interpolation (5) from the values of the color components of the RGB pixels of the scanned image with the coordinates floor (x _s ') and floor (x _s ') +1 . In step 7 ', the color coordinates of the current pixel R'G'B' are converted into the YCbCr color space. At step 8 ', the correction of luminance (shadow) distortions is carried out. The inverse transformation from the YCbCr color space, where the brightness value Y has already been adjusted in the previous step, is performed in step 9 '. At step 10 ', the number of corrected pixels and the number of samples taken in the interval of samples are increased by one. In this case, the parameter t of the line increases by dt. Steps 4 'to 10' are executed in a loop until condition 11 'is true.

At step 12 ', the counter of cycle i is reduced by one to iterate over the interpolating line segments.

Steps 2 'to 12' are performed in a cycle for each segment while condition 13 'is true.

For correction in the range [x _sb (1), x _sb (m)], the block diagram is generally similar.

The fact that for each line of the scanned correction is carried out independently allows you to significantly optimize the correction time by pre-computing the look-up table (LUT - LookUp Table) with values (x _s ', z _s '), the index in the LUT is the x coordinate of the corrected image.

After performing the correction, the image is displayed on a printing device, and a copy of the object is obtained. Before printing, the coordinates of the pixels (x, y) of the corrected image are converted into the coordinates of the printing device:

where ρ matches the resolution of the scanner and printer; tx1 and ty1 carry the origin of the coordinate system.

The present invention contemplates framing along the borders of an object or pages.

Various logic of the system’s operation during copying is also possible, for example, when only one page of a book is copied or two pages of a book are copied onto 1 sheet with scaling and the like.

Within the framework of the present invention, it is possible to move the scanning head of the flatbed scanner along the X _s axis, which will require correction of several lines and a slight change in the correction tool (in particular, shown in Fig. 8), but will not affect the basic principles of the claimed system and scanning method.

Figure 9 shows an example of correction of a fragment of a page of a book. Fig. 9a shows a scanned fragment of a book page without correction, Fig. 9b - correction of distortion only, Fig. 9c - correction of distortion and shadow.

It should be noted that the present invention can be used not only for books, but also for any other objects (for example, documents and photographs) as a regular copier or flatbed scanner. Accordingly, the best embodiment of the invention is the modification of a copier, flatbed scanner or multifunction peripheral device (MFP).

The most significant improvements that are provided on the basis of the claimed invention are:

a) the correction of geometric distortions and the alignment of relative illumination during scanning of thick-bound objects, in particular books, is ensured in almost real time by using the image processing procedure generated by the camera located to the side of the scanned object;

b) analysis of the shape of the bending of the object’s pages in preliminary mode, namely at the time of the start of the scanning process, makes it possible to provide acceptable requirements for the system’s hardware resources (storing in the preferred embodiment of the invention only one or more scanning lines);

c) the possibility of using the proposed method for correcting this kind of interference in other devices of this kind, for example, scanners that scan from above, or in cases where an image of the object is formed using the main digital camera of high resolution.

The above embodiment of the invention is set forth to illustrate the claimed invention, while it is clear to those skilled in the art that various modifications, additions and substitutions are possible without departing from the scope and meaning of the present invention disclosed in the claims.

Claims (17)

1. A system for scanning documents, predominantly bound, comprising a scanner, a digital camera, a processor, a program memory unit, a document parameter data memory unit and an external display device, the scanner having a substrate on which the document is located and along which and the digital camera is located on one of the sides of the scanner opposite the binding line of the document in binding, and the lens of the digital camera is directed to the front side of the document, while the digital camera made with the possibility of obtaining a snapshot of the document in the area of bending pages near the binding, the program memory unit contains correction means configured to set parameters and an image correction method, the processor has a distortion correction module configured to analyze a snapshot of a document received from a digital camera, and correction of distortions of the scanned document near the binding by determining and processing parameters characterizing the form of bending of the pages of the document near the binding in accordance with the parameters and correction method. 2. The system according to claim 1, characterized in that the scanner is made flatbed. 3. The system according to claim 1, characterized in that the scanner substrate is made of glass or other transparent material. 4. The system according to claim 1, characterized in that the scanned document is placed on the substrate in expanded form, and the pages of the document are pressed against the substrate. 5. The system according to claim 1, characterized in that the scanned document is placed on the scanner substrate in such a way that the edges of the pages of the document are parallel to the borders of the scanner substrate. 6. The system according to claim 1, characterized in that at least one lighting device is arranged next to the digital camera, configured to turn on the digital camera while the document is being shot and the document is illuminated. 7. The system according to claim 1, characterized in that the digital camera is built into the scanner body. 8. The system according to claim 1, characterized in that the digital camera is connected to the scanner at the time of scanning the document. 9. The system according to claim 1, characterized in that the digital camera is located at an angle from 0 to 60 ° to the scanner substrate. 10. The system according to claim 1, characterized in that, as an external display device, the system comprises at least one device selected from the group including a monitor, printer, or computer. 11. A method of scanning documents, including the following operations: remove a document with a digital camera; write the received snapshot of the document in the memory unit data parameters of the documents; analyze in the processor distortion correction module the presence of distortions in the document image stored in the document parameter data memory in accordance with the parameters obtained from the correction means of the document parameter data memory block, and it is determined whether the document has distortions near the binding, if the document has no distortions near the binding, the document is scanned by the scanner and the image of the document is displayed on an external display device, in case the document has distortions near the binding, fissionable and stored in the data memory unit parameters documents parameters characterizing the shape bend near the binding of document pages; scan a document with a scanner; write the scanned image of the document in the memory unit data parameters of the documents; correct in the distortion correction module of the processor of the distortion of the scanned image of the document from the data unit of the document parameter data in accordance with the parameters and the correction method set in the correction tool of the memory unit of the document parameter data, using parameters characterizing the shape of the bending of the document pages near the binding; write the corrected scanned image of the document in the memory unit of the data parameters of the documents and output it to an external display device. 12. The method according to claim 11, characterized in that they remove the document with a digital camera, while illuminating the document with at least one lighting device from the side of the digital camera. 13. The method of claim 11, wherein the analyzing module in the distortion correction processing of a picture object document data memory unit parameters, the injected frame coordinate system of the object captured by the camera X _c Z _c, X-axis _which is parallel to and the axis Z _c perpendicular to the substrate, in the frame coordinate system of the object captured by the camera, specify an array of points on the scanner substrate that are equally spaced from one another, and determine the coordinates {x _c (i)} of each array point on the scanner substrate for all points of the array with coordinates and {x _c (i)} determine the array of distances {h (i)} from each point on the scanner substrate z _c = 0 to its projection on the object page, as a result, an array of pairs of points {x _c (i), h (i )}, determine the coordinate x _с0 corresponding to the maximum of the distance array: x _c0 : h (x _c0 ) = max {h (i)}, analyze the nature of the array {x _c (i), h (i)} to the left and right of x _c0 and determine whether the object has distortions near the binding or not. 14. The method according to item 13, characterized in that if the object has distortions near the binding, determine the data characterizing the shape of the bending pages of the object near the binding, while the array {x _c (i), h (i)} is converted to two coordinate arrays {x _c (j), Z _ca G)} and {x _c (j), z _c bG)} introduce the coordinate system of the scanner X _s Z _s whose axes coincide with the axes of the coordinate system of the frame of the object captured by camera X _c Z _c , the coordinate arrays {x _c (j), z _ca G)} and {x _c (j), z _cb (j)} are converted to the scanner coordinate system X _s Z _s :

where a agrees the resolution of the camera and scanner along the X axis, d agrees the resolution of the camera and scanner along the Z axis, and also takes into account the angle of inclination of the camera to the substrate, where t _x transfers the origin of the coordinate system, resulting in coordinate arrays {x _sa (j), z _sa (j))} of n elements and {x _sb (j), z _sb (j)} of m elements, which determine the form of bending of the pages of the object near the binding. 15. The method according to claim 11, characterized in that the distortion correction module of the processor for distorting the scanned image of the object is corrected, while the distortion and shadow in the bending area of the pages of the image of the object are corrected. 16. The method according to 14, characterized in that the distortion correction module corrects for distortion of the scanned image of the object, using piecewise linear interpolation over the coordinate arrays {x _sa (j), z _sa (j} and {x _sb (j ), z _sb (j)}, it is believed that the number of pixels of each row of the corrected scanned image of the object exceeds the number of pixels of the scanned image by the value Delta, which is specified by the following dependence: Delta = round (L _a + L _b - (x _sb (m) -X _sa (n))), where round is the function of rounding to the nearest integer, L _a and L _b are the total lengths of the interpolating segments, which corresponds to the number of pixels of the curved parts of the pages, while L _a and L _{b are} given by the following dependence:

to correct distortion along the interpolated curve formed by the cross section of the curved surface of the object page, the samples are selected in increments of a unit length of the scanner coordinate system, as a result of which the corresponding coordinate x _s 'is determined, from which the values of the color components R'G'B' are calculated from the scanned image corrected image by linear interpolation according to the values of the color components of the RGB pixels of the scanned image with the coordinates floor (x _s ') and floor (x _s ') +1, where floor is the function of taking the integer part: R '= R (floor (x _s '), y) + (x _s ' -floor (x _s ')) ^* R (floor (x _s ') +1, y), for the values of G 'and B', they are calculated similarly, the current coordinates of the samples (x _s ', z _s ') along the curve interpolated by the segments are calculated by the parametric equations of the segments of the straight line from the point with coordinates (x _sa (i), z _sa (i)) to (x _sa (i-1), z _sa (i-1)), i = 2 ... n: x _s ' = x _sa (i) + t (x _sa (i-1) -x _sa (i), z _{× s} ' = z _sa (i) + t (z _sa (i-1) -z _sa (i)), where t = 0 at the beginning of the segment (point (x _sa (i), z _sa (i)), t = 1 at the end of the segment (point (x _sa (i-1), z _sa (i-1)), and the change step t is given by the expression:

the calculation of the current coordinates of the samples (x _s ', z _s ') along the interpolated curve for the array {x _sb (j), z _sb (j)} is carried out similarly, the experimentally averaged dependence of the brightness of the distorted pixel on the brightness of the original pixel and the distance to the projection of this pixels on the page from the scanner substrate, then this dependence is converted into a distortion correction function of the scanned image of the object, which also takes into account the scanning resolution: Y '(x, y) = f (Y (x _s ', y), z _s '), where Y '(x, y) is the brightness of the adjusted pixel, Y (x _s ', y) is the pixel brightness of the scanned image, x _s 'is the distance to the page from the scanner substrate in x _s '. 17. The method according to claim 11, characterized in that using the distortion correction module, distortions are corrected for the scanned image of the object, and the distortions in each line of the scanned image are corrected independently of other lines.

Images