Search

EP-4740168-A1 - METHOD FOR OBTAINING A STACK OF IMAGES OF A SCENE, COMPUTER PROGRAM, DEVICE, APPARATUS AND VEHICLE IMPLEMENTING SUCH A METHOD

EP4740168A1EP 4740168 A1EP4740168 A1EP 4740168A1EP-4740168-A1

Abstract

The invention relates to a method (100) for obtaining a stack of images of a scene from a stack of source images (PILs), referred to as source stack, which is acquired with a camera module comprising at least one optical lens and at least one image sensor, and includes at least two source images acquired with different focuses, said method (100) comprising the following steps for at least one source image from said source stack: - correcting (102) geometric aberration(s) in the source image in relation to a reference format by means of a coordinate transformation function, and - storing (108) said corrected image in a stack of images (PILc), referred to as corrected stack, which includes the at least one corrected image instead of the source image. The invention also relates to a computer program and a device implementing such a method.

Inventors

  • LEGROS, ERIC
  • PETITGRAND, Sylvain
  • LUONG, BRUNO

Assignees

  • Fogale Optique

Dates

Publication Date
20260513
Application Date
20230708

Claims (15)

  1. 1. Method (100;200;300) for obtaining a stack of images of a scene, from a stack of source images (PIL S ), called source stack, of said scene, said source stack (PIL S ): - comprising at least two source images acquired at different focuses, such that each of said at least two source images have different areas of sharpness for said scene, and - having been acquired with a camera module comprising at least one optical lens and at least one image sensor; said method (100;200;300) comprising the following steps for at least one source image of said source stack: - correction (102) of geometric aberration(s) in said source image, predetermined, by application, to at least one pixel of said source image, of a predetermined coordinate transformation function relative to a reference format, said correction providing a corrected image, and - storing (108) said corrected image in an image stack (PILc), called corrected stack, comprising said at least one corrected image in place of said source image.
  2. 2. Method (100;200;300) according to the preceding claim, characterized in that the reference format is: - the format of a source image from the source stack, or - the format of the image sensor of the camera module.
  3. 3. Method (100; 200; 300) according to the preceding claim, characterized in that, for at least one source image, the step (102) of correcting geometric aberration(s) in said source image comprises the following steps carried out for at least one pixel of said source image: - calculation (104), by the predetermined coordinate transformation function taking as input the position of said pixel in said source image, of a position, called corrected, of said pixel in the reference format, and - resetting (106) said pixel to said corrected position.
  4. 4. Method (100; 200; 300) according to any one of the preceding claims, characterized in that, for at least one source image, the transformation function is a function of the focus of said source image, so that two source images acquired at two different focuses are corrected with different transformation functions.
  5. 5. Method (100;200;300) according to any one of the preceding claims, characterized in that at least one transformation function is a position correspondence matrix.
  6. 6. Method (200; 300) according to any one of the preceding claims, characterized in that it further comprises, prior to the correction step (102), a phase (202), called the calibration phase, for determining at least one coordinate transformation function.
  7. 7. Method (200; 300) according to any one of the preceding claims, characterized in that the calibration phase (202) is carried out with the same camera module as that used for the acquisition of the source stack (PIL S ) so that at least one transformation function is specific to said camera module.
  8. 8. Method (200; 300) according to any one of claims 6 or 7, characterized in that the calibration phase (202) comprises the following steps: - acquisition (204) of an image stack (PILcai), called a calibration stack, with at least the same focuses as those of the source images of the source stack (PIL S ), and - deduction (206) from said calibration stack (PILcai) of at least one coordinate transformation function.
  9. 9. Method (100; 200; 300) according to any one of claims 1 to 5, characterized in that at least one coordinate transformation function is received from a device other than that implementing the correction step. (102), and in particular when the source stack (PILs) has been acquired by said other device.
  10. 10. Method (300) according to any one of the preceding claims, characterized in that it further comprises a step (302) of acquiring the source stack (PILs).
  11. 11. Computer program comprising executable instructions which, when executed by a computing device, implement all the steps of the method (100; 200; 300) according to any one of the preceding claims.
  12. 12. Processing device (400) comprising means configured to implement all the steps of the method (100; 200; 300) according to any one of claims 1 to 10.
  13. 13. Apparatus (510;520;530) comprising: - at least one camera module (430) comprising an optical lens and an image sensor, and - at least one calculation unit (400); configured to implement all the steps of the method (100;200;300) according to any one of claims 1 to 10.
  14. 14. Apparatus (510;520;530) according to the preceding claim, characterized in that it is: - a smartphone (510), - a tablet - a computer, - a television, - a virtual reality or augmented reality headset (520), or - a medical imaging device (530).
  15. 15. Vehicle (600) comprising: - at least one camera module (430) comprising an optical lens and an image sensor, and - at least one calculation unit (400); configured to implement all the steps of the method (100; 200; 300) according to any one of claims 1 to 10.

Description

DESCRIPTION Title: Method for obtaining a stack of images of a scene, computer program, device, apparatus and vehicle implementing such a method. [0001] The present invention relates to a method for obtaining a stack of images of a scene, and in particular to a stack of images in which at least two images, and in particular all the images, represent the scene with a different focus. It also relates to a computer program and a device implementing such a method. It further relates to an apparatus and a vehicle implementing such a method. [0002] The field of the invention is the field of obtaining a stack of images of a scene at different focuses. State of the art [0003] Techniques are known for representing a scene from a stack of images taken at different focusing distances, also called focus, depths of field, different in the present application. For example, the technique called focus bracketing, or focus stacking, "stacking of focus points" in French, is known, which makes it possible to capture a stack of images of a scene, each image being captured at a different depth of field. Then, the scene is represented on a display medium, with a greater depth of field, by exploiting the stack of images. [0004] However, current techniques, although improving the depth of field of the image of a scene, have drawbacks. When acquiring the stack of images by a camera module comprising an image sensor and an optical lens, the distance between the image sensor and the optical lens is modified to change the focus, in particular for the acquisition of each image. The inventors have noticed that this change in focus causes distortions, or shifts, in the images and therefore degrades the representation of the scene on a display medium from the stack of images thus obtained. [0005] An aim of the present invention is to remedy at least one of the drawbacks of the state of the art. [0006] Another aim of the invention is to propose a solution making it possible to obtain a stack of images representing the scene at different focuses comprising fewer, or no, geometric aberrations, thus allowing a representation of said scene with fewer deformations or shifts. Disclosure of the invention [0007] The invention proposes to achieve at least one of the aforementioned aims by a method of obtaining a stack of images of a scene, from a stack of source images, called source stack, of said scene, said source stack: - comprising at least two source images acquired at different focuses, such that said at least two source images have different areas of sharpness for said scene, and - having been acquired with a camera module comprising at least one optical lens and at least one image sensor; said method comprising the following steps for at least one source image of said source stack: - correction of geometric aberration(s) in said source image, predetermined, by application, to at least one pixel of said source image, of a predetermined coordinate transformation function relative to a reference format, said correction providing a corrected image, and - storing a stack of images, called corrected, comprising said at least one corrected image in place of said source image. [0008] Thus, the invention proposes to obtain a stack of images, called a corrected stack, in which the geometric aberrations due to a change of focus during the acquisition of the stack of source images are corrected, in at least one image, and in particular all the stack images (except possibly one). As a result, the corrected stack does not have, or has fewer, geometric aberrations. By using the corrected image stack, it is then possible to produce a representation of the scene on a display medium with little or no deformation or offset during said representation, in particular when switching from one of the images in the corrected stack to another, for example to modify the sharp area of the scene displayed on the display medium. In addition, the use of the corrected stack does not require any registration of the images forming said corrected stack between them. [0009] It should be noted that the correction of geometric aberrations does not consist of, or is different from, the registration of the images between them. Indeed, the registration of a first image relative to a second image applies a coordinate transformation that is uniform or identical to all the pixels: for example a translation or a rotation, identical to all the pixels of the image. In a different way, the correction of geometric aberrations is performed with a coordinate transformation that is different for at least two pixels of the image, and in particular all pixels of the image. [0010] By image is meant a digital image, and in particular a matrix image. [0011] According to the invention, the display medium can be any type of display medium. [0012] For example, the display medium may be a display screen, such as for example a touch screen, etc. [0013] For example, the display medium may be a surfac