MC200052A1 - Method and system for digitizing images - Google Patents

Description

002^59

2 0 0 0 Hl1 5

Mr. 2001

OUPUCTA

5

PATENT FOR INVENTION

10 IMAGE DIGITIZATION METHOD AND SYSTEM.

Gérard POPINEAU

TECHNICAL FIELD OF THE INVENTION.

15 The present invention relates to a method and system for digitizing images.

TECHNOLOGICAL BACKGROUND OF THE INVENTION.

The problem of digitizing large and/or high-resolution images arises for very diverse applications.

This can involve reproducing assembled or bound "paper" documents (books, registers, journals, etc.) as facsimile images, an operation that has become increasingly common since advances in electronic document management (EDM) have been sufficient to make its implementation routine in businesses and institutions (libraries, research institutes, local authorities, etc.). In this case, the most common solution is the use of static digital cameras of the type developed and marketed by companies such as MINOLTA (PS3000®), ZEUTSCHEL (OMNISCAN3000©), or IMAGEWARE (BOOKEYE®).

The company MINOLTA is the holder of several patents in this field, for example US patent US5995245, published on November 30, 1999 in the name of F. MORO, and patent US5969795, published on October 19, 1999 in the name of T. HONDA.

(/

The document is placed open and flat under the camera lens, and the pages are turned by an operator. This work is slow and tedious (approximately 200 pages per hour).

The operator's speed is limited in any case by the performance of existing cameras: several seconds, or even tens of seconds, are needed to acquire a 5 double A4 page with a resolution of 300dpi, even in monochrome mode.

The CCD sensors used in these devices are not generally of the "matrix" type, as they are far too expensive considering the number of pixels required.

These are arrays of the type used in line-of-sight cameras, which therefore necessitates the provision of means to move either the image or the object. The movements must be very precisely controlled to obtain good image quality. In known systems, the scanning is linear, the movement is therefore alternating, and necessarily slow to avoid vibrations.

The performance of electronic sensors, such as those marketed by the company DALSA, reaches several hundred million pixels per second: existing mechanical linear scanning devices do not make the best use of this resource.

15 Methods for analyzing or generating an image other than linear scans are known from the prior art.

The Nipkow disk used a spiral scan to create the first television images.

US patent 570186, published on February 9, 1999, in the name of P. MOGAN et al., describes a radial scan for analyzing a dust collector disk using a CCD camera. The method implemented aims to provide a large collecting area, with image quality and acquisition speed being secondary.

US patent 4620235, issued on October 28, 1986, in the name of P. Watt, discloses another implementation of radial scanning for digitizing a film disc using a linear 25 CCD sensor. The disc rotates in front of the camera, and a Dovc prism is used to compensate for this movement and reduce linearity errors.

In US patent 3647961, published on March 7, 1972 in the name of F. BLITCHINGTON, the Dove prism is used to rotate the image of a conductive printed circuit pad in front of a video camera and thus detect possible defects.

30 It therefore appears from the documents cited above that the radial scanning devices known in the prior art have not, until now, been designed to utilize the performance of CCD sensors.

3

commercially available linear components for manufacturing high-speed, high-resolution image scanning systems.

GENERAL DESCRIPTION OF THE INVENTION.

5

The present invention therefore aims to obtain a device for digitizing images with high performance.

Its specific purpose is a process for digitizing images at a given resolution, these images being contained within a rectangular frame of the image plane, and consisting of producing a sequence of 10 digital data representing, line by line, the intensity and color of each of the elements,

Regularly distributed, corresponding rectangular images. The method of the present invention is remarkable in that it comprises:

a) a first step in which the circular images of a disk encompassing the frame of the image plane are analyzed, ray by ray, and generate a sequence of digital data representing

15 the intensity and value of each of the regularly distributed elements of these circular images.

b) a second step during which this sequence is transformed by computational means into the sequence of numerical data representing the corresponding rectangular images according to the desired resolution.

Advantageously, the image digitization process according to the invention includes an intermediate step during which the sequence of digital data is stored by storage means.

In order to implement the method according to the invention, the image digitization system, according to a first embodiment, comprises at least one linear CCD sensor for image acquisition mounted in the image plane on a support rotating around an axis of rotation 25 perpendicular to this plane.

According to a second embodiment, the system includes an optical device for rotating images, preferably a Dovc prism rotating around an axis of rotation perpendicular to the image plane.

An additional feature of the system according to either embodiment is that the instantaneous center of rotation of the support or images is located between the ends of at least one of the sensors.

4

Preferably, the instantaneous center of rotation is located outside the ends of several radially arranged sensors.

Advantageously, the sensor(s) are located symmetrically with respect to this instantaneous center of rotation.

5 Benefits will be derived from the application of the image digitization system according to the invention as an imager for a digital static camera, a digital video camera or a digital cinematographic shooting camera.

These few essential specifications make evident to the person skilled in the art the advantages brought by the process and the image digitization system according to the invention compared to the state of the prior art.

BRIEF DESCRIPTION OF THE DRAWINGS.

15 Figures 1a and 1b schematically present the process according to the invention.

Figures 2a, 2b, 2c and 2d show the preferred arrangement of linear CCD sensors respectively in the case of a single sensor, several sensors for high resolution, a single sensor for double the frame rate, several sensors for high resolution and double the frame rate.

20 Figure 3a is a schematic view of the image digitization system according to a first embodiment comprising a rotating sensor support.

Figure 3b is a schematic view of the system according to a second embodiment comprising a rotating Dove prism.

25

DESCRIPTION OF THE PREFERRED FORMS OF THE INVENTION.

Figure 3 schematically illustrates how a rectangular image ABCD is analyzed by radial scanning. Each pixel of the image is located on a polar axis (U) with a polar angle of 9. The 30-digitization system of the signals from the linear CCD sensor(s) produces an array of N

lines each containing P data C(I, J) corresponding to the ray-by-ray analysis up to N rays at

5

average of an equivalent linear CCD of P points and length R, C(I,J) being the intensity and color of the pixel with polar coordinates (RI/P, 2hJ/N). for 1= 1... N and J= 1... P.

Figure 1b shows the algorithm for simulating the data T(G,H) corresponding to a linear scan of the image ABCD from the data C(I,J) generated by a radial scan. 5 At the beginning of the algorithm (1), the integer variables I and J are equal to 1 (2). For each pair of values (I,J), the data C(I,J) is read (3) and the Cartesian coordinates (X,Y) of the pixel are calculated (4):

X = (R1/P)cos(2tiJ/N)

Y = (RI/P)sin(27tJ/N)

10

If the pixel is a pixel of the image ABCD, of length L and width 1, that is to say if (5):

-L/2 <= X <= L/2 -1/2 <= Y <= 1/2

15

the data T(G,H) is calculated (6):

T(G,H) = C(I,J) with

G= Ent((X+L/2)/a) + 1 H= Ent((Y+l/2)/a) + 1 where a is the final resolution of the image ABCD.

Depending on the case, variables 1 (7) and/or J (8) are incremented by 1 (9,10) in order to read the following pair (I,J) (3).

The algorithm is finished when I=N and J=P (1 l). The data simulating a linear scan of 25 the image are contained in the table T(G,H), with G= 1 ... Ent(L/a)+1 and H= 1 ... Ent(l/a)+1.

Figures 2a, 2b, 2c, and 2d show various arrangements of linear CCD sensors for radial image scanning. The simplest case is shown in Figure 1a. A single sensor (12)

A radially arranged array analyzes a disk A. Depending on the resolution to be achieved for the final image, 30 of the arrays having a sufficient number P of elements and a sufficiently high line frequency may not be used.

6

not be available on the market. In this case the provisions indicated in Figures 2b, 2c and 2d compensate as needed for the insufficient performance of the sensors (13,14,15,16,17).

In Figures 2b and 2d, several low-resolution but high-line frequency sensors (13, 16, 17) are equivalent to a single very high-resolution sensor. The juxtaposition of the sensors (13, 16, 17), whose ends overlap, creates radial shifts that are compensated for by appropriate processing of the data from the circular areas A, B, and C.

The use of sensors (14,15; 16.17) distributed over a diameter (XY,X'Y',X"Y"), as shown in Figures 2c and 2d, allows us to divide by two the image acquisition time for a given line frequency of the sensors (14,15.16).

10 These sensor arrangements (12,13,14,15,16,17) are used in radial scanning camera optical heads as shown in Figures 3a and 3b.

Figure 3a shows an embodiment in which the support (18) for the CCD(s) (19), placed in the image plane (20) of a lens (21), is rotating. This implies the implementation of a rotating commutator (22) on the shaft of the drive motor (23) to provide electrical connections with the image sensor (19).

In a second embodiment shown in Figure 3b, the support (24) of the sensor(s) (25) is fixed in the secondary image plane (26) of an optical translator (27). This translator comprises two lenses (28, 29) between which a Dove prism (30) is placed. The rotation of the prism (30), driven by a motor (31), produces the rotation in the secondary image plane (26) of the image 20 produced in the primary image plane (32) by the camera lens (33).

As is self-evident, the invention is not limited to the preferred embodiments described above. On the contrary, it encompasses all possible variants of embodiment.

25

30

7

Claims (1)

DEMANDS 1) A method for digitizing images at a given resolution, said images being contained within a rectangular frame (ABCD) of the image plane (20,26), and consisting of producing a sequence of data 5 numerical representations, line by line, of the intensity and color of each of the regularly distributed elements of the corresponding rectangular images (ABCD), characterized in that it comprises: a) a first step in which the circular images of a disk encompassing said frame (ABCD) are analyzed, ray by ray, and generate a sequence of digital data representing the intensity and value of each of the regularly distributed elements of said circular images. b) a second step in which said sequence is transformed by computational means into said sequence according to said resolution. 2) Image digitization method according to claim 1, characterized in that it comprises an intermediate step during which said sequence is stored by storage means. 15 3) Image digitization system adapted to the implementation of the method according to any one of claims 1 or 2, comprising at least one linear CCD sensor (12,13,14,15,16,17,19) intended for the acquisition of said images, characterized in that said sensors (12,13,14,15,16,17,19) are mounted in said plane (20) on a support rotating about an axis of rotation perpendicular to said plane (20). 20 4) An image digitization system adapted to implementing the method according to any one of claims 1 or 2, comprising at least one linear CCD sensor (12, 13, 14, 15, 16, 17, 25) for acquiring said images, characterized in that said system comprises an optical device for rotating said images, specifically a Dove prism (30) rotating about an axis of rotation 25 perpendicular to said plane (26). 5) Image digitization system according to any one of claims 3 or 4, characterized in that the instantaneous center of rotation (O) of said support (18) or of said images is located between the ends of at least one of said sensors (12,13,14,15,16,17). 30 8 6) Image digitization system according to claim 5, characterized in that said instantaneous center of rotation (O) is located outside the ends of several of said radially arranged sensors (13,16,17). 5 7) Image digitization system according to any one of claims 5 or 6, characterized in that said sensor(s) (14,15;16,17) are located symmetrically with respect to said instantaneous center of rotation (O). 8) Application of the image digitization system according to any one of claims 3 to 7, 10 characterized in said system constitutes the imager of a static digital camera. 9) Application of the image digitization system according to any one of claims 3 to 7, characterized in said system constitutes the imager of a digital video camera. 15 10) Application of the image digitization system according to any one of claims 3 to 7, characterized in said system constitutes the imager of a digital cinematographic shooting camera.