Method for producing a geometrical transformation on a video image and devices for carrying out said method

• US 4,752,828 A
• Filed: 07/02/1987
• Issued: 06/21/1988
• Est. Priority Date: 05/11/1983
• Status: Expired due to Term

• Current Assignee: Thomson-CSF (Thales SA)
• Original Assignee: Thomson-CSF (Thales SA)
• Inventors: Chapuis, Franck, Xuan, Kiet N., Oyaux, Jacques
• Primary Examiner(s): Groody, James J.
• Assistant Examiner(s): Bauer, Robert M.
• Application Number:
  US07/070,128
• Time in Patent Office: 355 Days
• Field of Search: 358/22, 358/183, 358/21 R, 358/100, 382/41, 382/44, 382/45, 382/46, 382/47, 340/798, 364/521, 364/522
• US Class Current: 348/580
• CPC Class Codes: G06T 3/00 Geometric image transformations in the plane of the image
• CPC Class Codes: H04N 5/2628 Alteration of picture size, shape, position or orientation, e.g. zooming, rotation, rolling, perspective, translation

Abstract

Special effects are produced on video images by means of a method of geometrical transformation which can involve a translation, a rotation, an affine transformation or an effect of perspective. The method consists:

in locating each point of the initial image by means of coordinates (X, Y, O) in a cartesian reference frame designated as a movable reference frame and related to the initial image;

in representing each point of the initial image by a binary word;

in storing the initial image in a storage device having two read address inputs and one data output, the function of the device being to deliver a binary word corresponding to the point with coordinates (X, Y, O) in the initial image when read address values X and Y are applied respectively to the two inputs:

in causing each point of the transformed image resulting from an initial-image transformation to be located by means of coordinates (X₃, Y₃, O) in a cartesian reference frame which is related to the transformed image and designated as a fixed reference frame;

in representing each point of the transformed image by a binary word M supplied by the storage device by applying values of X and Y to the read address inputs of the device, these values being computed as a function of the coordinates (X₃, Y₃, O) of said point and as a function of the geometrical transformation parameters.

Claims

1. A method for producing a geometrical transformation on a video image designated as an initial image, comprising the steps of:

identifying each point of the initial image by a plurality of first coordinates (X, Y, O) in a first cartesian reference frame which is designated as a movable reference frame and is related to the initial image;

representing each point of the initial image by a digital word;

storing the digital words representing each point of the initial image at a plurality of addressable locations in a storage device having address inputs and a data output, so that a digital word corresponding to the each point and having coordinates (X, Y, O) in the initial image is delivered when address values corresponding to the each point are applied to the storage device address inputs;

performing a transformation of the initial image using predetermined tranformation parameters to yield a transformed image;

identifying each point of the transformed image by a plurality of second coordinates (X₃, Y₃, O) in a second cartesian reference frame which is related to the transformed image and is designated as a fixed reference frame;

determining a relationship between the second coordinates and the first coordinates, based on the transformation parameters and calculating one computed address value; and

representing each point of the transformed image by a digital word supplied from the storage device by applying computed address values to the address inputs of said storage device, said computed values being computed as a sum of the computed address value of second coordinates (X₃, Y₃, O) of said point and, an increment value.

2. A method of transformation according to claim 1, wherein the geometrical transformation is an affine transformation having coefficients (A_x, A_y) and includes a rotation through an angle α

, both transformations being performed in the plane of the initial image, said transformation also including a projection in perspective on the plane of the transformed image, this projection being defined by a direction of view and a vanishing point, said direction of view being defined by two predetermined angular values β and

γ and

said vanishing point of said projection being located at a distance ρ

_PF from the origin of the fixed reference frame, said origin being the first point of the first line of the transformed image, and said transformation further including a translation (X_O3, Y_O3) in the plane of the transformed image, comprising the further steps of;

computing the read addresses as a function of those computed in the immediately preceding point in the transformed image when the each point in the transformed image has coordinates (X₃ >

O, Y₃ O,O) said computing being in accordance with the formulae;

space="preserve" listing-type="equation">X(X.sub.3, Y.sub.3)=X(X.sub.3 -1, Y.sub.3)+k.D.sub.xx

space="preserve" listing-type="equation">Y(X.sub.3, Y.sub.3)=Y(X.sub.3 -1, Y.sub.3)+k.D.sub.yx (

     15)

initializing values of the each point having coordinates (X₃ =O, Y₃ >

O,O) and which are located at the beginning of a line of the transformed image by the values;

space="preserve" listing-type="equation">X(O, Y.sub.3)=X(O, Y.sub.3 -1)+k.D.sub.xy

space="preserve" listing-type="equation">Y(O, Y.sub.3)=Y(O, Y.sub.3 -1)+k.D.sub.yy (

     16)

where X(O, Y₃ -1), Y(O, Y₃ -1) being the read address values computed in respect of the first point of the immediately preceding line in the transformed image; and

calculating the first point of the transformed image (X₃ =O, Y₃ =O) according to the relation;

##EQU11##

3. A method according to claim 2, wherein the vanishing point is assumed to be located at infinity, and wherein k=1.

4. A method according to claim 3 for producing an effect of multiplication of the initial image in addition to the geometrical transformation in the transformed image, wherein said method comprises the further steps of:

periodically reinitializing the value of (X, Y) in accordance with the formulae;

space="preserve" listing-type="equation">X(O, Y.sub.3)=X(O, Y.sub.3 -1)+k.D.sub.xy

space="preserve" listing-type="equation">Y(O, Y.sub.3)=Y(O, Y.sub.3 -1)+k.D.sub.yy (

     16)

during each transformed-image line scan, with a period being a submultiple of the line scanning interval;

periodically reinitializing the value of (X, Y) in accordance with the formulae;

space="preserve" listing-type="equation">X(O,O)=-k.D.sub.xx.X.sub.O3 -k.D.sub.xy.Y.sub.O3

space="preserve" listing-type="equation">X(O,O)=-k.D.sub.yx.X.sub.O3 -k.D.sub.yy.Y.sub.O3 (

     17)

during each transformed-image field scan, with a period being longer than the duration of one line scan and being a submultiple of the field scanning interval.

5. A method according to claim 4 for producing a mirror effect in addition to the geometrical transformation and the multiplication effect, and comprising the further steps of:

reversing the sign of D_xx and D_yx at the time of each reinitialization of (X, Y) with the values given by the formulae;

space="preserve" listing-type="equation">X(O, Y.sub.3)=X(O, Y.sub.3 -1)+k.D.sub.xy

space="preserve" listing-type="equation">Y(O, Y.sub.3)=Y(O, Y.sub.3 -1)+k.D.sub.yy (

     16)

in order to obtain symmetries with respect to vertical axes;

reversing the sign of D_yy and D_xy at the time of each reinitialization of (X, Y) with the values given by the formula

space="preserve" listing-type="equation">X(O,O)=-k.D.sub.xx.X.sub.O3 -k.D.sub.xy.Y.sub.O3

space="preserve" listing-type="equation">X(O,O)=-k.D.sub.yx.X.sub.O3 -k.D.sub.yy.Y.sub.O3 (

     17)

in order to obtain symmetries with respect to horizontal axes.

6. A method for producing a geometrical transformation of an initial video image, comprising the steps of:

identifying each point of the video image by a first address in a first coordinate system;

storing a digital word representing characteristics with each point in the video image at the first address;

performing a transformation of the original image using predetermined transformation parameters to yield a transformed image;

identifying each point in the transformed image by a second address in a second coordinate system;

computing a value of an initial point in the transformed image;

determining a first recurrence, which when added to a value of an adjacent point yields a value of a current point;

determining each value of each point from the value of an adjacent point; and

using such value as an address to access the stored digital word representing the characteristics of each point.

7. A method as in claim 6 wherein said determining a recurrence step includes the steps of:

determining a first recurrence between any point and a point adjacent to such any point on a same line; and

determining a second recurrence between a point at the end of one line and a point on another line.

8. A method as in claim 7 wherein said computing a value of an initial point step is performed during a time interval which elapses between two successive images.

9. A method as in claim 8 wherein said geometrical transformation includes a rotation, an affine transformation and a projection to produce a perspective effect.

10. An apparatus for performing a transformation on a video image, comprising:

image memory means for storing a digital word representing characteristics of each point of an initial image;

processing means for;

(1) determining initial values for a point of a transformed image, and (2) determining a recurrence value between each point and an adjacent point of the transformed image; and

adding means for producing a current address from an adjacent address by adding said recurrence value and one of (a) said initial value when said initial value is said adjacent address, and (b) the previous current address output by said adding means to thereby perform an adding operation between a recurrence value and the previous address used to produce a current address in real time.

11. Apparatus as in claim 10 wherein there are two recurrence values, a first recurrence value for calculating adjacent points which are on the same line, and a second recurrence value for calculating the value of another point based on a point which is on a different line.

12. Apparatus as in claim 11 further comprising means for multiplexing said initial values and said previous current address so that one of said initial values and said previous current address is applied to a first input of said adding means.

13. Apparatus as in claim 12 further comprising first register means coupled between an output of said processing means and a second input of said adding means for temporarily storing information output by said processing means;

and

second register means coupled between an output of said processing means and an input of said multiplex means for temporarily storing said initial values output by said processiog means and for providing said initial values at an output thereof to said multiplex means.

14. Apparatus as in claim 10 wherein said processing means is also for (3) controlling said multiplex means so that a proper input is coupled to said adding means.

15. Apparatus as in claim 14 wherein said adding means and said multiplex means are for one component of each point of the video image and wherein there are second adding means and second multiplex means for another component thereof.

16. A device according to claim 10 in which the initial video image is any desired image, further comprising:

at least one counter for generating write address values at the point scanning frequency;

wherein the image memory means is a memory having a capacity equal to one image, alternately written at an address supplied by the counters and read at an address supplied by the processing means at the point scanning frequency, the written data being the value of a digitized video signal corresponding to one field of the initial image and the read values being such as to constitute the values of a video signal corresponding to one field of the transformed initial image as a result of the geometrical transformation thereby resulting in an efficiency of operational transformation.

17. A device according to claim 10 further comprising means for generating control signals for a variable-gain video switcher, wherein the image memory means is for storing data defining a geometric figure.