Method and device for learning a depth prediction model to reduce the loss of consistency of a stereoscopic vision system.

Abstract

Method or device for learning a depth prediction model implemented by a neural network associated with a vision system. Indeed, the method comprises the generation (32, 34) of depth maps associated with two images acquired by two cameras, the generation (35) of images from the acquired images, depth maps and extrinsic parameters of the vision system as well as the generation of depth maps associated with the generated images, the determination (36) of a first consistency error by comparing first and fourth depth maps and a second consistency error by comparing second and third depth maps and the learning (37) of the depth prediction model by minimizing a loss error determined from a result of a comparison of the first and second consistency errors. Figure for the abstract: Figure 3

Description

Title of the invention: Method and device for learning a depth prediction model to reduce the loss of consistency of a stereoscopic vision system. Technical field

[0001] The present invention relates to methods and devices for determining a depth by a vision system on board a vehicle, for example in a motor vehicle. The present invention particularly relates to a method and a device for learning a depth prediction model associated with a stereoscopic vision system on board a vehicle. Technological background

[0002] Many modern vehicles are equipped with so-called AD AS (Advanced Driver-Assistance System). Such AD AS systems are passive and active safety systems designed to eliminate the element of human error in the driving of vehicles of all types. AD AS use advanced technologies to assist the driver while driving and thus improve their performance. AD AS use a combination of sensor technologies to perceive the environment around a vehicle, then provide information to the driver or act on certain vehicle systems.

[0003] There are several levels of ADAS, such as rearview cameras and blind spot sensors, lane departure warning systems, adaptive cruise control and automatic parking systems.

[0004] The AD AS embedded in a vehicle are supplied with data obtained from one or more embedded sensors such as, for example, cameras. These cameras make it possible in particular to detect and locate other road users or possible obstacles present around a vehicle in order, for example: • to adapt the vehicle's lighting according to the presence of other users; • to automatically regulate the vehicle speed; • to act on the braking system in the event of a risk of impact with an object.

[0005] The prediction of a distance separating an object from a vehicle, that is to say the depth associated with a pixel of an image acquired by a camera, the pixel representing the object, requires high precision in order to feed the AD AS. The prediction of a depth associated with a pixel notably calls upon a depth prediction model, the depth prediction model being adapted to the vision system on board the vehicle. Such a prediction model is commonly implemented by a neural network, which is notably learned in a training phase in order to predict accurate depths in relation to the on-board vision system and the type of environment in which the vehicle operates. A training phase then requires data, for example annotated data from libraries or annotated via other more precise on-board systems such as a LIDAR®. However, this type of learning requires a massive accumulation of data or the presence of this other on-board system which the vehicle does not always have. It is then possible to carry out a training phase of the depth prediction model from data acquired by the on-board vision system but the existing solutions are not adapted to the configuration of any on-board vision system in a vehicle. Summary of the present invention

[0006] An object of the present invention is to solve at least one of the problems of the technological background described above.

[0007] Another object of the present invention is to improve the learning phase of a depth prediction model from data resulting from the processing of images acquired by a vision system, in particular by a depth prediction model implemented by a neural network associated with a stereoscopic vision system.

[0008] Another object of the present invention is to improve road safety, in particular by improving the operational safety of AD AS systems supplied by data obtained from a camera of a vision system.

[0009] According to a first aspect, the present invention relates to a method for learning a depth prediction model implemented by a convolutional neural network associated with a stereoscopic vision system on board a vehicle, the vision system comprising a first camera and a second camera arranged so as to each acquire an image of a three-dimensional scene from a different point of view, the method being implemented by at least one processor, and being characterized in that it comprises the following steps: - reception of data representative of a first image and a second image acquired by the first camera and the second camera respectively at the same acquisition time instant; - generating a first depth map comprising depths associated with a set of pixels of the first image, the depths being predicted with the depth prediction model from the first and second images; - generation of a first image returned by symmetry of the first image with respect to a central vertical axis of the first image and generation of a second image flipped by symmetry of the second image with respect to a central vertical axis of the second image; - generating a second depth map comprising depths associated with a set of pixels of the second image, the depths being predicted with the depth prediction model from the returned second and first images; - generating a third image and a third depth map from the first image, the first depth map and extrinsic parameters of the stereoscopic vision system, the third depth map comprising depths associated with a set of pixels of the third image, and generating a fourth image and a fourth depth map from the second image, the second depth map and extrinsic parameters of the stereoscopic vision system, the fourth depth map comprising depths associated with a set of pixels of the fourth image; - determining a first consistency error by comparing the first and fourth depth maps and a second consistency error by comparing the second and third depth maps; - learning the depth prediction model by minimizing a loss error determined from a result of a comparison of the first and second consistency errors.

[0010] According to a variant of the method, the first consistency error and the second consistency error are determined respectively by the following function: (Dt(p)-Dt(p)Y With : • corresponding to the first consistency error L^p), respectively the second consistency error p Y for a pixel P, • D(p) a depth of a pixel Pt obtained from the first depth map, respectively from the second depth map, for a pixel P, and • D'(p) a depth of a pixel Pt obtained from the fourth depth map, respectively from the third depth map, for the pixel P.

[0011] According to another variant, the method further comprises determining a first photometric error determined by comparing pixel values ​​of the first and fourth images and a second photometric error determined by comparing pixel values ​​of the second and third images, the loss error being determined from the first and second photometric errors.

[0012] According to a further variant of the method, the loss error is determined by the following function: L^^min^L^p), Lc2(p)) + min(Lpi(p), Lp2(p))) With : • The loss error, • Lp^p) the first photometric error for a pixel P, P being a pixel defined by its coordinates in an image, • Lp2 the second photometric error for pixel P, • Lcl the first consistency error for pixel P, and • Le2 the second consistency error for pixel P.

[0013] According to yet another variant of the method, a photometric error is determined by the following function: L p *(p) - 0-«)■ R(p)-Hp) 1 +a ' With : • Lp*(p) the first photometric error noted Lpi(p), respectively the second photometric error noted Lp2(p), P being a pixel defined by its coordinates in an image, • I(p) a value of the pixel P in the first image, respectively second image, • a value of the pixel P in the fourth image, respectively third picture, • SSIM a function that takes into account a local structure, and • has a weighting factor depending in particular on the type of environment.

[0014] According to yet another variant of the method, the third and fourth depth maps are generated by the following function: D(pt) ) ] ) With : • Ps the coordinates of a pixel of the third depth map, respectively of the fourth depth map, • Æ a function to go from homogeneous coordinates to pixel coordinates by removing a dimension from a vector, • K' an intrinsic matrix of the second camera, respectively to the first camera, • K an intrinsic matrix of the first camera, respectively to the second camera, • T an extrinsic matrix of the stereoscopic vision system, and • 0 a projection function in the three-dimensional scene of a pixel Pt as a function of its coordinates in the second image, respectively first image, and of the depth associated with it in the second depth map, respectively first depth map.

[0015] According to an additional variant of the method, the generation of the second depth map comprises a generation of another depth map associated with the second returned image, a depth of a pixel of the second depth map having as its value the depth of another pixel of the other depth map, a correspondence of a pixel of the other depth map with a pixel of the second depth map being determined during the generation of the second returned image.

[0016] According to a second aspect, the present invention relates to a device configured to learn a depth prediction model by a vision system on board a vehicle, the device comprising a memory associated with at least one processor configured to implement the steps of the method according to the first aspect of the present invention.

[0017] According to a third aspect, the present invention relates to a vehicle, for example of the automobile type, comprising a device as described above according to the second aspect of the present invention.

[0018] According to a fourth aspect, the present invention relates to a computer program which comprises instructions adapted for executing the steps of the method according to the first aspect of the present invention, in particular when the computer program is executed by at least one processor.

[0019] Such a computer program may use any programming language and be in the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form, or in any other desirable form.

[0020] According to a fifth aspect, the present invention relates to a computer-readable recording medium on which is recorded a computer program comprising instructions for carrying out the steps of the method according to the first aspect of the present invention.

[0021] On the one hand, the recording medium may be any entity or device capable of storing the program. For example, the medium may comprise a storage means, such as a ROM memory, a CD-ROM or a microelectronic circuit type ROM memory, or a magnetic recording means or a hard disk.

[0022] Furthermore, this recording medium may also be a transmissible medium such as an electrical or optical signal, such a signal being able to be conveyed via an electrical or optical cable, by conventional or hertzian radio or by self-directed laser beam or by other means. The computer program according to the present invention may in particular be downloaded from an Internet-type network.

[0023] Alternatively, the recording medium may be an integrated circuit in which the computer program is incorporated, the integrated circuit being adapted to perform or to be used in performing the method in question. Brief description of the figures

[0024] Other characteristics and advantages of the present invention will emerge from the description of the particular and non-limiting exemplary embodiments of the present invention below, with reference to the appended figures 1 to 4, in which:

[0025] [Fig-1] schematically illustrates a vision system equipping a vehicle, according to a particular and non-limiting example of embodiment of the present invention;

[0026] [Fig.2] illustrates a flowchart of the different steps of a method for determining a depth of a pixel of an image by a depth prediction model associated with a vision system on board the vehicle of [Fig.l], according to a particular and non-limiting exemplary embodiment of the present invention;

[0027] [Fig.3] illustrates a flowchart of the different steps of a method for learning the depth prediction model used in the method of [Fig.2], according to a particular and non-limiting exemplary embodiment of the present invention; and

[0028] [Fig.4] schematically illustrates a device configured to learn a depth prediction model by a vision system on board the vehicle of [Fig.l], according to a particular and non-limiting exemplary embodiment of the present invention. Description of examples of implementation

[0029] A method and a device for learning a depth prediction model implemented by a convolutional neural network associated with a stereoscopic vision system on board a vehicle will now be described in what follows with joint reference to Figures 1 to 4. The same elements are identified with the same reference signs throughout the description which follows.

[0030] The terms "first(s)", "second(s)" (or "first(s)", "second(s)"), etc. are used in this document by arbitrary convention to enable different elements (such as operations, means, etc.) implemented in the embodiments described below to be identified and distinguished. Such elements may be distinct or correspond to a single element, depending on the embodiment.

[0031] According to a particular and non-limiting example of embodiment of the present invention, the depth prediction model is learned in a learning phase comprising the generation of depth maps associated with two images acquired by two cameras, the generation of images and other depth maps from the acquired images, previously generated depth maps and parameters extrinsic components of the stereoscopic vision system as well as the generation of depth maps associated with the generated images.

[0032] A first and a second consistency errors are determined by comparing the depth maps associated with the acquired images to depth maps associated with the generated images and the depth prediction model is learned by minimizing a loss error determined from a result of a comparison of the first and second consistency errors.

[0033] [Fig. 1] schematically illustrates a vision system equipping a vehicle, according to a particular and non-limiting exemplary embodiment of the present invention.

[0034] Such an environment 1 corresponds, for example, to a road environment formed of a network of roads accessible to the vehicle 10.

[0035] In this example, the vehicle 10 corresponds to a vehicle with a thermal engine, with electric motor(s) or even a hybrid vehicle with a thermal engine and one or more electric motors. The vehicle 10 thus corresponds, for example, to a land vehicle such as an automobile, a truck, a bus, a motorcycle. Finally, the vehicle 10 corresponds to an autonomous vehicle or not, that is to say a vehicle traveling according to a determined level of autonomy or under the total supervision of the driver.

[0036] The vehicle 10 advantageously comprises at least two on-board cameras, a first camera 11 and a second camera 12, configured to acquire images of a three-dimensional scene taking place in the environment of the vehicle 10 from separate observation positions. The first camera 11 and the second camera 12 form a stereoscopic vision system when used together as illustrated in [Fig.l]. The first camera 11 forms a monoscopic vision system when used alone, likewise the second camera 12 forms another monoscopic vision system when used alone. The present invention, however, extends to any vision system comprising at least two cameras, for example 2, 3 or 5 cameras.

[0037] The intrinsic parameters of the first camera 11 characterize the transformation which associates, for an image point, subsequently called “point”, its three-dimensional coordinates in the reference frame of the first camera 11 with the pixel coordinates in an image acquired by the first camera 11. These parameters do not change if the first camera 11 is moved. The intrinsic parameters of the first camera 11 include in particular a first focal distance fl associated with the first camera 11.

[0038] The intrinsic parameters of the second camera 12 characterize, for their part, the transformation which associates, for an image point, its three-dimensional coordinates in the reference frame of the second camera 12 with the pixel coordinates in an image acquired by the second camera 12. These parameters do not change if the second camera 12. The intrinsic parameters of the second camera 12 include in particular a second focal length f2 associated with the second camera 12.

[0039] The distortions, which are due to imperfections in the optical system such as defects in the shape and positioning of the camera lenses, will deflect the light beams and therefore induce a positioning deviation for the projected point compared to an ideal model. It is then possible to complete the camera model by introducing the three distortions which generate the most effects, namely radial, decentering and prismatic distortions, induced by defects in curvature, parallelism of the lenses and coaxiality of the optical axes. In this example, the cameras are assumed to be perfect, that is to say that the distortions are not taken into account, that their correction is processed at the time of image acquisition or at the time of calibration.

[0040] These two cameras 11, 12 are arranged so as to each acquire an image of a scene from a different point of view, the first point of view is for example located on or in the left rearview mirror of the vehicle 10 or at the top of the windshield of the vehicle 10, the second point of view is for example located on or in the right rearview mirror of the vehicle 10 or at the top of the windshield of the vehicle 10. In the case where the two cameras are located at the top of the windshield of the vehicle, they are then placed at a certain distance. In this example, the first camera 11 is located at the top of the windshield of the vehicle 10, the second camera 12 is located in the right rearview mirror of the vehicle 10.

[0041] A first reference point is associated with the first camera 11: - the direction of the x axis is defined horizontal and normal to the optical axis Cl of the first camera 11. The distance B separating the optical center of the first camera 11 from the projection of the optical center of the second camera 12 on the horizontal plane passing through the optical center of the first camera 11 is called the reference base (in English “baseline”); - the direction of the y axis is defined vertical and normal to the optical axis Cl of the first camera 11; - the direction of the z axis is defined orthogonal to the directions of the x and y axes. The three axes x, y and z thus form an orthonormal reference frame.

[0042] According to a particular exemplary embodiment, the optical axes C1, C2 of the first and second cameras are included in the same plane but are not parallel.

[0043] The extrinsic parameters linked to the position of the cameras 11, 12 are the following parameters: - three translations in the x, y and z directions: Tx, Ty and Tz constituting the translation vector T; and - three rotations in the x, y and z directions: 0x, 0y and 0z.

[0044] An extrinsic matrix of the vision system then includes the previously defined extrinsic parameters.

[0045] The extrinsic parameters are determined, for example, during a calibration phase of the stereoscopic vision system comprising the first camera 11 and the second camera 12.

[0046] A main constraint of the stereoscopic vision system used in automobiles is, for example, the large distance between the two cameras. Indeed, to be able to cover a measurement range of 200 meters, the reference base must reach 60cm for the cameras commonly used in this field.

[0047] The two cameras 11, 12 acquire images of a scene located in front of the vehicle 10, the first camera 11 covering only a first acquisition field 13, the second camera 12 covering only a second acquisition field 14 and the two cameras 11, 12 both covering a third acquisition field 15. The first and third acquisition fields 13, 15 thus allow a monoscopic view of the scene by the first camera 11, the second and third acquisition fields 14, 15 allow a monoscopic view of the scene by the second camera 12 and the third acquisition field 15 allows a stereoscopic view of the scene by the stereoscopic vision system composed of the two cameras 11, 12.

[0048] An obstacle 18 is placed in the acquisition field of the cameras, for example in the third acquisition field 15. The presence of the obstacle 18 defines an occlusion field for the stereoscopic vision system composed here of the three fields 16, 17 and 19.

[0049] Among these three fields, field 16 is visible from the second camera 12. The part of the scene present in this field 16 is therefore observable using the monoscopic vision system comprising the second camera 12.

[0050] The field 17 is visible from the first camera 11. The part of the scene present in this field 17 is therefore observable using the monoscopic vision system comprising the first camera 11.

[0051] Finally, field 19 is not visible to any of the cameras. The part of the scene present in this field 19 is therefore not observable.

[0052] According to a particular exemplary embodiment, the field of vision of the second camera 12 covers at least half of the field of vision of the first camera 11.

[0053] It is obvious that it is possible to use such a stereoscopic vision system to take images of scenes located on the sides or behind the vehicle 10 by equipping it with differently placed and oriented cameras.

[0054] The images acquired by the cameras 11, 12 at an acquisition time instant are presented in the form of data representing pixels characterized by: - coordinates in each image; and - data relating to the colors and brightness of objects in the observed scene in the form, for example, of RGB colorimetric coordinates (from the English “Red Green Blue”) or TSL (Tone, Saturation, Brightness).

[0055] Each pixel of the acquired image is representative of an object in the three-dimensional scene present in the camera's field of vision. Indeed, a pixel of the acquired image is the smallest visible unit and corresponds to a luminous point resulting from the emission or reflection of light by a physical object present in the three-dimensional scene. When the light strikes this object, photons are emitted or reflected, captured by a photosensitive sensor of the camera after passing through its lens. This sensor divides the three-dimensional scene into a grid of pixels. Each pixel records the light intensity at a specific location, thus capturing visual details. The combination of millions of pixels creates an image faithfully representing the physical object observed by the camera. An image point previously presented is thus a point on a surface of an object in the three-dimensional scene.

[0056] The images acquired by the cameras 11, 12 represent views of the same scene taken from different viewpoints, the positions of the cameras being distinct. On this scene are found for example: - buildings; - road infrastructure; - other stationary users, for example a parked vehicle; and / or - other mobile users, for example another vehicle, a cyclist or a moving pedestrian.

[0057] According to a particular exemplary embodiment, a field of view of the first camera 11 covers at least half of a field of view of the second camera 12 and a field of view of the second camera 12 covers at least half of a field of view of the first camera 11. In other words, more than half of the pixels of an image acquired by the first camera 11 correspond to an object of the three-dimensional scene seen by the second camera 12, pixels of an image acquired by the second camera 12 also corresponding to this object of the three-dimensional scene. Similarly, more than half of the pixels of an image acquired by the second camera 12 correspond to an object of the three-dimensional scene seen by the first camera 11, pixels of an image acquired by the first camera 11 also corresponding to this object of the three-dimensional scene.

[0058] The images acquired by the first camera 11 and by the second camera 12 are sent to a computer of a device equipping the vehicle 10 or stored in a memory of a device accessible to a computer of a device equipping the vehicle 10.

[0059] A method for determining a depth by a vision system on board the vehicle 10 is advantageously implemented by the vehicle 10, that is to say by a processor, a computer or a combination of computers of the on-board system of the vehicle 10, for example by the computer(s) in charge of the vision system of the vehicle 10.

[0060] [Fig. 2] illustrates a flowchart of the different steps of a method 2 for determining a depth of a pixel of an image by depth prediction model implemented by a convolutional neural network associated with a vision system embedded in a vehicle, for example in the vehicle 10 of [Fig. 1], according to a particular and non-limiting exemplary embodiment of the present invention. The method 2 is for example implemented by a device of the vision system embedded in the vehicle 10 or by the device 4 of [Fig. 4].

[0061] In a step 21, data representative of an image acquired by the first camera 11 and of an image acquired by the second camera 12 are received.

[0062] In a step 22, depths associated with a set of pixels of one of the received images are determined by the depth prediction model from the two received images.

[0063] Each determined depth then corresponds to a distance separating the vehicle 10 or a part of the vehicle 10 from an object of the three-dimensional scene with which a pixel is associated, the determination of a depth of a pixel then corresponding to a measurement of a distance separating an object from the vehicle carrying the vision system.

[0064] If the ADAS uses these depths or distances as input data to determine the distance between a part of the vehicle 10, for example the front bumper, and another user present on the road, the ADAS is then able to determine this distance precisely. For example, if the ADAS has the function of acting on a braking system of the vehicle 10 in the event of a risk of collision with another road user and the distance separating the vehicle 10 from this same road user decreases significantly, then the ADAS is able to detect this sudden approach and act on the braking system of the vehicle 10 to avoid a possible accident.

[0065] [Fig. 3] illustrates a flowchart of the different steps of a method for learning the depth prediction model used in a method for determining a depth of a pixel of an image, for example in method 2 of [Fig. 2], according to a particular and non-limiting exemplary embodiment of the present invention.

[0066] The learning method 3 is for example implemented by the device on board the vehicle 10 implementing the method for determining a depth by a vision system on board a vehicle or by the device 4 of [Fig.4].

[0067] In a step 31, data representative of a first image and a second image are received, the first image being acquired by the first camera 11 at an acquisition time instant and the second image being acquired by the second camera at the same acquisition time instant.

[0068] According to a particular exemplary embodiment, the first image and second image have the same definition, that is to say they comprise the same number of pixels, have the same number of pixels according to their height and the same number of pixels according to their width.

[0069] According to another particular embodiment, the first image and second image are not of the same definition. An additional step then consists of resizing or cropping them to obtain a first image and a second image of the same definition.

[0070] In a step 32, a first depth map comprising depths associated with a set of pixels of the first image is determined. The depths are predicted with the depth prediction model from the first and second images.

[0071] The depth prediction model, implemented by a convolutional neural network, is known to those skilled in the art and is for example presented in the document “UnOS: Unified Unsupervised Optical-flow and Stereo-depth Estimation by Watching Videos” by Yang Wang, Peng Wang, Zhenheng Yang, Chenxu Luo, Yi Yang and Wei Xu published in June 2019, adapted to a stereoscopic vision system comprising cameras whose optical axes are included in the same plane, or in the document “Unifying Flow, Stereo and Depth Estimation” written by Haofei Xu, Jing Zhang, Jianfei Cai, Hamid Rezatofighi, Fisher Yu, Dacheng Tao and Andréas Geiger, published in July 2023 adapted for any stereoscopic vision system including those comprising cameras whose optical axes are not included in the same plane.

[0072] Thus, a predicted depth is associated with each pixel of a set of pixels of the first image, the first depth map comprising the coordinates of each pixel of the set of pixels of the first image and an associated depth recorded for example in a correspondence table, in a memory accessible to the processor implementing this learning method 3.

[0073] In a step 33, a first flipped image is generated by symmetry of the first image with respect to a central vertical axis of the first image and a second flipped image is generated by symmetry of the second image with respect to a central vertical axis of the second image (such an operation being called “flip” in English). The central vertical axis of an image is the axis which divides this image into two equal parts, for example each comprising the same number of pixels.

[0074] The pair comprising the second returned image and the first returned image is equivalent to a pair of images acquired by the first 11 and second 12 cameras acquiring a virtual three-dimensional scene corresponding to the symmetry of a real scene with respect to a plane, the plane being normal to the axis separating the positions of the two cameras and equidistant from the positions of the two cameras.

[0075] The pair of returned images thus allows the depth prediction model to predict depths associated with the pixels of the second returned image, the latter having the appearance of a first image.

[0076] In a step 34, a second depth map is generated, the second depth map comprising depths associated with a set of pixels of the second image, the depths being predicted with the depth prediction model from the returned second and first images.

[0077] The generation of the second depth map comprises, for example, a generation of another depth map associated with the second returned image, the depths being predicted by the depth prediction model. A depth of a pixel of the second depth map then has as its value the depth of another pixel of the other depth map, a correspondence of a pixel of the other depth map with a pixel of the second depth map being determined during the generation of the second returned image. In other words, the depth prediction model predicts depths associated with the pixels of the second returned image. These depths are then associated with the pixels of the second image, each pixel of the second returned image corresponding to a pixel of the second image of which it is the symmetrical one. The second depth map is then symmetrical with respect to the other depth map.

[0078] Similar to the first depth map, a predicted depth is associated with each pixel of a set of pixels of the second image, the second depth map comprising the coordinates of each pixel of the set of pixels of the second image and an associated depth recorded for example in the previously described lookup table.

[0079] In a step 35, a third image and a third depth map are generated from the first image, the first depth map and extrinsic parameters of the stereoscopic vision system, the third depth map comprising depths associated with a set of pixels of the third image.

[0080] The generation of the third depth map from the first image, the first depth map and extrinsic parameters of the stereoscopic vision system consists of: • determining spatial coordinates in the three-dimensional scene, in a reference system associated with the first camera 11, of a point associated with a first pixel of the first depth map from the coordinates of the first pixel in the first image and the depth which is associated with this first pixel, its coordinates and depth being recorded in the first depth map. It should be noted that the coordinates of the first pixel in the first image comprise two components while the spatial coordinates of the point associated with the first pixel comprise three components; • the determination of spatial coordinates in the three-dimensional scene of the point associated with the first pixel in a reference frame associated with the second camera 12, in other words a change of reference frame from that associated with the first camera 11 to that associated with the second camera 12, using the extrinsic parameters of the stereoscopic vision system; and • the projection of the point associated with the pixel in the image plane of the second camera 12, this image plane corresponding to that of the second image, making it possible to determine the arrival coordinates of the first pixel of the first image in an image such as the second camera 12 would have acquired it. The image plane of a camera corresponds to a plane defined in the frame of reference of the camera, normal to the optical axis of the camera and located at the first focal length of the camera. The third depth map then includes the arrival coordinates of the first pixels and the depth associated with the first pixel.

[0081] The arrival coordinates of a pixel in the third depth map are thus determined and are the same as the arrival coordinates of a pixel in the third image. In the third depth map, the depth of a pixel corresponds to a depth determined from pixel depths of the first depth map, while in the third image, the value of a pixel corresponds to a value of a pixel determined from pixel values ​​of the first image.

[0082] Similarly, a fourth image and a fourth depth map are generated from the second image, the second depth map and the extrinsic parameters of the stereoscopic vision system, the fourth depth map comprising depths associated with a set of pixels of the fourth image.

[0083] According to a particular exemplary embodiment, the third depth map is generated using the following function: [Math.l] Px = ^(K[T^(piK,~^ D(PT) ) ] )

[0084] With : • Ps the coordinates of a pixel of the third depth map, • 77 a function to go from homogeneous coordinates to pixel coordinates by removing a dimension from a vector, • K' the intrinsic matrix of the first camera 11, • K the intrinsic matrix of the second camera 12, • T an extrinsic matrix of the stereoscopic vision system, and • 0 a projection function in the three-dimensional scene of a first pixel Pt as a function of its coordinates in the first image and the depth D^p) associated with it in the first depth map.

[0085] The determination of the fourth depth map is similar to the determination of the third depth map, the cameras being in particular inverted. The arrival coordinates of a pixel in the fourth depth map are thus determined and are the same as the arrival coordinates of a pixel in the fourth image. In the fourth depth map, the depth of a pixel corresponds to a depth determined from pixel depths of the second depth map, while in the fourth image, the value of a pixel corresponds to a value of a pixel determined from pixel values ​​of the second image.

[0086] Thus, the previous function for determining the arrival coordinates of a first pixel of the first depth map is applicable to determine the arrival coordinates of a second pixel of the second depth map, the function being adapted as follows: [Math.2] P, = T^PkK^ D(pr} ) ] )

[0087] With : • Ps the coordinates of a pixel of the fourth depth map, •77 a function to go from homogeneous coordinates to pixel coordinates by removing a dimension from a vector, • K' the intrinsic matrix of the second camera 12, • K the intrinsic matrix of the first camera 11, • T an extrinsic matrix of the stereoscopic vision system, and • 0 a projection function in the three-dimensional scene of a second pixel Pt as a function of its coordinates in the second image and of the depth [^pj which is associated with it in the second depth map.

[0088] According to an alternative embodiment, depth values ​​associated with pixels of the third and fourth images, i.e. depths of the third and fourth depth maps, are obtained by interpolating the depths using a function such as torch.nn.functional.grid_sample() in Python® language which requires as arguments the pixel coordinates of the third and fourth depth maps and the depths determined in the second and first depth maps.

[0089] In a step 36, a first consistency error is determined by comparing the first and fourth depth maps and a second consistency error is determined by comparing the second and third depth maps.

[0090] According to a particular exemplary embodiment, the first consistency error and the second consistency error are determined respectively by the following function: [Math.3] (Dt(p)-D't(p)Ÿ

[0091] With: • Llj)) corresponding to the first consistency error L^pY respectively the second consistency error I^p), for a pixel P, • D(p) a depth of a pixel Pt obtained from the first depth map, respectively from the second depth map, for a pixel P, and • D'(p) a depth of a pixel Pt obtained from the fourth depth map, respectively from the third depth map, for the pixel P.

[0092] Other functions can be used to perform this comparison, for example by taking the absolute value of the difference in depths. Indeed, it is important that the result of the comparison of depths is, for each pixel, a positive value.

[0093] In a step 37, the depth prediction model is learned by minimizing a loss error determined from a result of a comparison of the first and second consistency errors.

[0094] According to a first particular exemplary embodiment, the learning method 3 further comprises the determination of a first photometric error determined by comparing values ​​of the pixels of the first and fourth images and of a second photometric error determined by comparing values ​​of the pixels of the second and third images, the loss error being determined from the first and second photometric errors.

[0095] A photometric error is, for example, presented in the document “Digging Into Self-Supervised Monocular Depth Estimation” by Clément Godard, Oisin Mac Aodha, Michael Firman and Gabriel Brostow published in August 2019 and is determined by the following function:

[0096] [Math.4] Lp*(p) = (l-«) ■ U(.P)-Hp)\+a-

[0097] With: • Lp*( p ) the first photometric error noted Lp] ( p ), respectively the second photometric error noted Lp2{p), P being a pixel defined by its coordinates in an image, • I(p) a value of the pixel P in the first image, respectively second image, • a value of the pixel P in the fourth image, respectively third picture, • SSIM a function that takes into account a local structure, and • has a weighting factor depending in particular on the type of environment.

[0098] According to this particular embodiment, the loss error is determined by the following function: [Math.5] L=Hp(min(Lci(p), Lc2(p))+ min(Lpl(p), Lp2(p)))

[0099] With: • The loss error, • Lpi(p) the first photometric error for a pixel P, P being a pixel defined by its coordinates in an image, • Lp2 the second photometric error for pixel P, • L(:l the first consistency error for pixel P, and • Lc2 the second consistency error for pixel P.

[0100] According to a second particular exemplary embodiment, the loss error comprises components further determined by the following function: [Math.6]

[0101] With: • The nworth(p) component p) for a pixel p of the third image, respectively the component for a pixel p of the fourth image, • D(p) is a depth of a pixel P obtained from the third depth map, respectively obtained from the fourth depth map; • W is a parameter matrix; • 0 is the order of a smoothing gradient; • an L1 norm of second-order depth gradients is calculated with W =1, and ° =2 ; • x and y are the dimensions of the images; • / 3 is a hyperparameter dependent on the environment in which the vehicle operates; and • / ( nj is a value of the pixel P in the third image, respectively fourth image.

[0102] This function is generally used to deal with discontinuity at the edge of objects (in English “edge aware smoothness”).

[0103] The loss error includes, for example, consistency errors, photometric errors and the above-determined components: [Math.7] £ = ^min(Lcl(p), Lc2(p) ) + min(Lpi(p), Lp2(p)) + Ls3(p) + Ls4(p))

[0104] With: • the loss error, • Lp-^p) the first photometric error for a pixel P, P being a pixel defined by its coordinates in an image, • Ap2 the second photometric error for pixel P, • Lcl the first consistency error for pixel P, • Lc2 the second consistency error for pixel P, • Ls3 the component in the third image for pixel P, • Ls4 the component in the fourth image for pixel P.

[0105] It should be noted that the pixels compared are those with similar or equal coordinates in the acquired images as in the generated images.

[0106] Training the depth prediction model consists of adjusting input parameters of the convolutional neural network in order to minimize the previously calculated loss error.

[0107] In addition, an object visible in the field of vision of one camera and masked in the field of vision of the other camera does not impact the loss error, this method is therefore insensitive to occlusions. Thus, the depth prediction model used for the depth prediction of a pixel of an image acquired by the first camera 11 is made reliable thanks to this learning method.

[0108] This learning is carried out from data acquired by the on-board vision system and therefore does not require data annotated by another on-board system or storage of a library of learning images. In addition, the learning data is representative of the data received when the system is in operation or in production, in fact the learning data is representative of real environments in which the vehicle carrying the stereoscopic vision system, this training data is therefore particularly relevant.

[0109] [Fig.4] schematically illustrates a device 4 configured to learn a depth prediction model by a vision system embedded in a vehicle, according to a particular and non-limiting exemplary embodiment of the present invention. The device 4 corresponds for example to a device embedded in the first vehicle 10, for example a computer associated with the stereoscopic vision system.

[0110] The device 4 is for example configured for the implementation of the operations described with regard to figures 1 and 4 and / or steps described with regard to figures 2 and 3. Examples of such a device 4 include, but are not limited to, on-board electronic equipment such as an on-board computer of a vehicle, an electronic calculator such as an ECU (“Electronic Control Unit”), a smartphone, a tablet, a laptop. The elements of the device 4, individually or in combination, can be integrated in a single integrated circuit, in several integrated circuits, and / or in discrete components. The device 4 can be produced in the form of electronic circuits or software (or computer) modules or even a combination of electronic circuits and software modules.

[0111] The device 4 comprises one (or more) processor(s) 40 configured to execute instructions for carrying out the steps of the method and / or for executing the instructions of the software(s) embedded in the device 4. The processor 40 may include integrated memory, an input / output interface, and various circuits known to those skilled in the art. The device 4 further comprises at least one memory 41 corresponding for example to a volatile and / or non-volatile memory and / or comprises a memory storage device which may comprise volatile and / or non-volatile memory, such as EEPROM, ROM, PROM, RAM, DRAM, SRAM, flash, magnetic or optical disk.

[0112] The computer code of the embedded software(s) comprising the instructions to be loaded and executed by the processor is for example stored in the 4L memory.

[0113] According to various particular and non-limiting embodiments, the device 4 is coupled in communication with other similar devices or systems (for example other computers) and / or with communication devices, for example a TCU (from the English “Telematic Control Unit” or in French “Telematic Control Unit”), for example via a communication bus or through dedicated input / output ports.

[0114] According to a particular and non-limiting exemplary embodiment, the device 4 comprises a block 42 of interface elements for communicating with external devices. interface elements of block 42 include one or more of the following interfaces: - RF radio frequency interface, for example Wi-Fi® type (according to IEEE 802.11), for example in the 2.4 or 5 GHz frequency bands, or Bluetooth® type (according to IEEE 802.15.1), in the 2.4 GHz frequency band, or Sigfox type using UBN (Ultra Narrow Band) radio technology, or LoRa in the 868 MHz frequency band, LTE (Long-Term Evolution), LTE-Advanced; - USB interface (from the English “Universal Serial Bus” or “Universal Serial Bus” in French); HD MI interface (from the English “High Definition Multimedia Interface” or “High Definition Multimedia Interface” in French); - LIN interface (from the English “Local Interconnect Network”).

[0115] According to another particular and non-limiting exemplary embodiment, the device 4 comprises a communication interface 43 which makes it possible to establish communication with other devices (such as other computers of the on-board system) via a communication channel 430. The communication interface 43 corresponds for example to a transmitter configured to transmit and receive information and / or data via the communication channel 430. The communication interface 43 corresponds for example to a wired network of the CAN (Controller Area Network) type, CAN FD (Controller Area Network Flexible Data-Rate), FlexRay (standardized by the ISO 17458 standard) or Ethernet (standardized by the ISO / IEC 802-3 standard).

[0116] According to a particular and non-limiting exemplary embodiment, the device 4 can provide output signals to one or more external devices, such as a display screen 440, touch-sensitive or not, one or more speakers 450 and / or other peripherals 460 via the output interfaces 44, 45, 46 respectively. According to a variant, one or other of the external devices is integrated into the device 4.

[0117] Of course, the present invention is not limited to the exemplary embodiments described above but extends to a method for determining the depth of a pixel of an image acquired by a vision system, and / or for measuring a distance separating an object from a vehicle carrying a vision system, the depth and / or the distance being predicted and / or measured via a depth prediction model learned according to the learning method described above, which would include secondary steps without departing from the scope of the present invention. This would be the case even of a device configured for the implementation of such a method.

[0118] The present invention also relates to a vehicle, for example an automobile or more generally an autonomous land-based motor vehicle, comprising the device 4 of [Fig.4].

Claims

1. Claims Method for learning a depth prediction model implemented by a convolutional neural network associated with a stereoscopic vision system on board a vehicle (10), the vision system comprising a first camera (11) and a second camera (12) arranged so as to each acquire an image of a three-dimensional scene from a different point of view, said method being implemented by at least one processor, and being characterized in that it comprises the following steps: - reception (31) of data representative of a first image and a second image acquired by the first camera (11) and the second camera (12) respectively at the same acquisition time instant; - generating (32) a first depth map comprising depths associated with a set of pixels of the first image, the depths being predicted with the depth prediction model from the first and second images; - generation (33) of a first image returned by symmetry of the first image with respect to a central vertical axis of the first image and generation of a second image returned by symmetry of the second image with respect to a central vertical axis of the second image; - generating (34) a second depth map comprising depths associated with a set of pixels of the second image, the depths being predicted with the depth prediction model from the returned second and first images; - generation (35) of a third image and a third depth map from the first image, the first depth map and extrinsic parameters of the stereoscopic vision system, the third depth map comprising depths associated with a set of pixels of the third image, and generation of a fourth image and a fourth depth map from the second image, the second depth map and extrinsic parameters of the stereoscopic vision system, the fourth depth map comprising depths associated with a set of pixels of the fourth image; - determination (36) of a first consistency error by com- comparison of the first and fourth depth maps and a second consistency error by comparing the second and third depth maps; - training (37) the depth prediction model by minimizing a loss error determined from a result of a comparison of the first and second consistency errors.

2. Method according to claim 1, for which the first consistency error and the second consistency error are determined respectively by the following function: ^0= (A4» -Dt(p))2 With: • L / p) corresponding to the first consistency error L^p), respectively the second consistency error For a pixel P, • D(p) a depth of a pixel Pt obtained from the first depth map, respectively from the second depth map, for a pixel P, and • D'(p) a depth of a pixel Pt obtained from the fourth depth map, respectively from the third depth map, for the pixel P.

3. A method according to one of claims 1 to 2, which further comprises determining a first photometric error determined by comparing pixel values ​​of the first and fourth images and a second photometric error determined by comparing pixel values ​​of the second and third images, the loss error being determined from the first and second photometric errors.

4. Method according to claim 3, for which the loss error is determined by the following function: L = Yp(min(Lcl(p), Lc2(p) ) + min (Lpi(p), Lp2(p) )) With: • P the loss error, • Lp^p) the first photometric error for a pixel P, P being a pixel defined by its coordinates in an image, • Lp2 the second photometric error for the pixel P, • Lci the first consistency error for the pixel P, and • Lc2 the second consistency error for the pixel P.

5. Method according to claim 4, for which a photometric error is determined by the following function: L}Ap) = (l-«) ' \Ap)A(p)\+a- (l-jSSIM^I(p),l(p) ) ) With: • Lp*( p} the first photometric error noted Lp\( p), respectively the second photometric error noted Lpz(p)A being a pixel defined by its coordinates in an image, • Kp) a value of the pixel P in the first image, respectively second image, * l(p) a value of the pixel P in the fourth image, respectively third image, • SSIM a function which takes into account a local structure, and • has a weighting factor depending in particular on the type of environment.

6. Method according to one of claims 1 to 5, for which the third and fourth depth maps are generated by the following function: P s = 71 ( K [( P ÏK " '' D ( P t) ) ] ) With: • Ps the coordinates of a pixel of the third depth map, respectively of the fourth depth map, • 77 a function for going from homogeneous coordinates to pixel coordinates by removing a dimension of a vector, • K' an intrinsic matrix of the first camera (11), respectively to the second camera (12), • K an intrinsic matrix of the second camera (12), respectively to the first camera (11), • T an extrinsic matrix of the stereoscopic vision system, and • 0 a projection function in the three-dimensional scene of a pixel Pt as a function of its coordinates in the first image, respectively second image, and of the depth I)(p] which is associated with it in the first depth map,respectively second depth map.,

7. Method according to one of claims 1 to 6, for which the generation of the second depth map comprises a generation of another depth map associated with the second returned image, a depth of one pixel of the second depth map having for value the depth of another pixel in the other depth map, a correspondence of a pixel in the other depth map with a pixel in the second depth map being determined when generating the second returned image.

8. Computer program comprising instructions for implementing the method according to any one of the preceding claims, when these instructions are executed by a processor.

9. Device (4) configured to learn a depth prediction model by a vision system on board a vehicle (10), said device (4) comprising a memory (41) associated with at least one processor (40) configured to implement the steps of the method according to any one of claims 1 to 7.

10. Vehicle (10) comprising the device (4) according to claim 9.

Images