Method for characterising a camera module and associated imaging system

Abstract

The invention relates to a method (30) for characterising a camera module (4) comprising a lens (6) and a sensor (8), the lens (6) belonging to a predetermined family of lenses associated with a particular target geometry, the characterisation method (30) including the steps of: - obtaining (32) configuration data which are indicative of a current configuration of the camera module and comprise: • at least one production parameter which is representative of an actual geometry of the lens (6); and / or • at least one use parameter which is representative of current image-acquisition conditions of the camera module (4); and - calculating (34) a current optical transfer function, which is associated with the current configuration of the camera module (4), from: • obtained configuration data; and • a predetermined estimation function (16) which is associated with the family of lenses and takes, as input, configuration data, and provides, as a result, a corresponding estimated optical transfer function.

Description

Method for characterizing a camera module and associated imaging system DESCRIPTION technical field

[0001] The present invention relates to a method for characterizing a camera module comprising a lens and a sensor, the sensor being arranged to acquire an image of a scene observed through the lens.

[0002] The invention also relates to a computer program and an imaging system implementing such a method.

[0003] The invention applies to the field of optics, and in particular to camera modules incorporating a lens. State of the art

[0004] It is common practice to integrate one or more camera modules into an electronic device, such as a smartphone. Each camera module typically includes a lens and a sensor.

[0005] In general, the lens comprises a set of lenses stacked according to a nominal configuration, usually determined by simulation during a design step prior to lens assembly.

[0006] However, a real objective generally exhibits deviations, akin to defects, from the nominal configuration.

[0007] Such discrepancies are, for example, partly attributable to the lenses. Indeed, lenses are susceptible to imperfections in shape compared to the expected theoretical shape. These imperfections are due to, for example, the shrinkage of the plastic material in which the lenses are made, the machining precision of the molds, the functional clearances for guiding the mold parts, etc.

[0008] In addition, the quality of the plastic material used for molding the lenses, as well as the way it is injected (temperature, pressure, cooling cycle time) are likely to lead to deviations in refractive index from an expected refractive index.

[0009] Furthermore, the application of anti-reflective coatings to lenses is a potential source of discrepancies in the effective thickness of each lens compared to an expected thickness.

[0010] These structural defects result in optical defects, that is to say an optical response of the lens which is different from the expected one (generally worse), and which includes spot spreading defects, chromatic aberration defects and / or geometric aberration defects.

[0011] Generally, to compensate for these defects, manufacturers integrate an image correction module into the electronic device equipped with the camera module, implementing, for example, contrast enhancement.

[0012] However, such a correction does not provide complete satisfaction.

[0013] Indeed, two supposedly identical realizations of a camera module generally present different defects, which results in a different optical response from one camera module to another, even within the same series.

[0014] However, this correction is not individualized, and is generally established from a reference, either from the design or from the production.

[0015] The result is that by applying such a correction to a particular camera module, the resulting image is likely to show under-corrected or over-corrected areas.

[0016] For example, due to such inappropriate correction, bright white fringes may appear around a dark object against a lighter background, in a particular part of the image, or at a specific distance from an object in the scene, etc. Conversely, a lack of contrast may occur in certain areas of the corrected image.

[0017] According to another example, due to such inappropriate correction, chromatic aberrations, recognizable by colored fringes around objects, are likely to occur.

[0018] Furthermore, it would be particularly tedious to determine, using a test bench, the corresponding optical response for each camera module produced across its entire range of operating parameters. Such a characterization method is therefore not feasible.

[0019] Besides being tedious, such a characterization method also has the drawback of resulting in a discrete measurement within the sensor's field. However, obtaining the correction function requires determining the optical response in a quasi-continuous manner according to the operating parameters.

[0020] One object of the present invention is to remedy at least one of the drawbacks of the prior art.

[0021] Another aim of the invention is to propose a method for characterizing an objective that is less tedious.

[0022] Another aim of the invention is to propose a method for characterizing a lens that allows for better correction of the optical defects inherent in said lens. Description of the invention

[0023] To this end, the invention relates to a method of the aforementioned type, in which the objective belongs to a predetermined family of objectives associated with a respective target geometry, the characterization method being implemented by computer and comprising the following steps: - obtaining indicative configuration data for a current camera module configuration, the configuration data including: • at least one realization parameter representative of an actual lens geometry; and / or • at least one usage parameter representative of common image acquisition conditions for the camera module; and - Calculation of a current optical transfer function, associated with the current configuration of the camera module, from: • configuration data obtained; and a predetermined estimation function associated with the lens family and taking configuration data as input, and providing, as a result, a corresponding estimated optical transfer function.

[0024] Indeed, thanks to the implementation of the calculation step, based on the camera module configuration data, an optical transfer function specific to the current camera module configuration is calculated. In this way, an optical transfer function for any camera module configuration can be calculated, using the estimation function, without the need for tedious measurements.

[0025] In addition, the calculation of each current transfer function of the camera module allows for a correction of the acquired images that is specific to the configuration of the camera module during the acquisition of each image.

[0026] Advantageously, the process according to the invention has one or more of the following characteristics, taken individually or in any technically feasible combination:

[0027] at least one realization parameter includes imaging information representative of an image of a pattern predetermined by the objective of the camera module, preferably in the plane of the respective sensor;

[0028] the estimation function is an artificial intelligence model pre-trained from a database associating at least one configuration dataset with a corresponding optical transfer function, each configuration dataset being provided to the artificial intelligence model as input, the corresponding optical transfer function being provided to the artificial intelligence model as expected output;

[0029] the estimation function is a mathematical regression function previously obtained from a predetermined mathematical function and a database associating at least one configuration dataset with a corresponding optical transfer function, and determined so as to minimize a difference between, on the one hand, the value taken by the mathematical function for a given configuration dataset and, on the other hand, the optical transfer function corresponding to said configuration dataset;

[0030] Each configuration dataset includes imaging information representative of an image of a predetermined pattern by a lens of a camera module having an optical transfer function equal to the optical transfer function associated with said configuration dataset;

[0031] The process includes, prior to the calculation step, an initialization step comprising: - for at least one lens family, modeling of at least one simulated camera module comprising a virtual lens associated with a virtual sensor, the simulated camera module presenting configuration data including at least one realization parameter whose value is different from the corresponding value in the target geometry associated with said lens family; - for each virtual objective modeled, calculation of a corresponding simulated optical transfer function; storage, in a database, of each optical transfer function in association with the corresponding configuration data;

[0032] the process also includes the generation of a corrected image from an image acquired by means of the camera module, according to the current calculated optical transfer function corresponding to said camera module;

[0033] the generation of the corrected image includes a deconvolution of the acquired image by the calculated current optical transfer function;

[0034] The process also includes an update of the corrected image comprising the following phases: - variation of a value of all or part of the configuration data in the vicinity of the corresponding obtained value; - calculation, from the estimation function and the modified value of each configuration data, of a corresponding intermediate optical transfer function; - generation, from the acquired image and each intermediate optical transfer function, of a corresponding intermediate corrected image; - for each intermediate corrected image, calculation of a corresponding value of a predetermined quality indicator; - calculation of a gradient of the quality indicator as a function of each configuration data; - determination, from each calculated gradient, of a corrected value for each configuration data point; - calculation, from the estimation function and the corrected value of each configuration data point, of an updated optical transfer function; and - updating the corrected image by deconvolution of the image acquired by the updated optical transfer function;

[0035] The implementation parameters include at least one piece of geometric information comprising: - a thickness of at least one lens of the objective; - a refractive index of at least one lens of the objective; - a decentering of at least one face of at least one lens of the objective; - an angular separation between a mean plane of each face of at least one lens; and / or - at least one data point representing a shape deviation of at least one face of at least one lens of the objective compared to the corresponding part of the target geometry;

[0036] The usage parameters include at least one of the following: - a position in a plane of the sensor; - a distance between the lens and the sensor; - at least one angle between a plane of the sensor and an optical axis of the lens; - a distance between an imaged point and the objective; - a focusing distance for which the imaged point of the scene is sharp; - a change of diaphragm; - continuous adjustment of the aperture of a diaphragm; - a possible optical zoom adjustment; - a position of the center of the lens relative to the sensor; - a position of the point of intersection of the optical axis of the lens with the sensor; and - a band of wavelengths of light coming from at least one imaged point.

[0037] According to another aspect of the invention, a computer program is proposed comprising executable instructions which, when executed by computer, implement the steps of the process as defined above.

[0038] The computer program can be in any computer language, such as for example machine language, C, C++, JAVA, Python, etc.

[0039] According to another aspect of the invention, a processing module is proposed configured such that, for at least one camera module, each camera module comprising a lens and a sensor, the sensor being arranged to acquire an image of a scene observed through the lens, the lens belonging to a predetermined family of lenses associated with a respective target geometry: - obtain indicative configuration data for a common camera module configuration, including: • at least one realization parameter representative of an actual lens geometry; and / or • at least one usage parameter representative of common image acquisition conditions for the camera module; and - Calculate a current optical transfer function, associated with the current configuration of the camera module, from: • configuration data obtained; and • of a predetermined estimation function associated with the lens family and taking configuration data as input, and providing, as a result, a corresponding estimated optical transfer function.

[0040] Advantageously, the processing module according to the invention has the following characteristic:

[0041] The processing module is also configured so that, for each camera module, it generates a corrected image from an image acquired by means of the camera module, according to the current calculated optical transfer function corresponding to said camera module.

[0042] The processing module according to the invention can be any type of device such as a server, a computer, a tablet, a calculator, a processor, a computer chip, programmed to implement the process according to the invention, for example by executing the computer program according to the invention.

[0043] According to another aspect of the invention, an imaging system is proposed comprising at least one camera module and a processing module as defined above connected to each camera module to receive configuration data relating to said camera module and, preferably, at least one image acquired by said camera module.

[0044] Preferably, the imaging system is a smart mobile phone (or "smartphone" in English), also called a smartphone or multifunction mobile phone.

[0045] In this case, the invention also relates to the use of a processing module according to the invention and / or a method according to the invention, within the smart mobile phone, to calculate a current optical transfer function of a camera module (for example, included in the smart mobile phone).

[0046] Alternatively, the imaging system is a touch tablet.

[0047] In this case, the invention also relates to the use of a processing module according to the invention and / or a method according to the invention, within the touch tablet, to calculate a current optical transfer function of a camera module (for example, included in the touch tablet).

[0048] Alternatively, the imaging system is a computer.

[0049] In this case, the invention also relates to the use of a processing module according to the invention and / or a method according to the invention, within the computer, to calculate a current optical transfer function of a camera module (for example, included in the computer).

[0050] Alternatively, the imaging system is a television or a display terminal.

[0051] In this case, the invention also relates to the use of a processing module according to the invention and / or a method according to the invention, within the television or display terminal, to calculate a current optical transfer function of a camera module (for example, included in the television or display terminal).

[0052] Alternatively, the imaging system is a virtual reality headset or an augmented reality headset.

[0053] In this case, the invention also relates to the use of a processing module according to the invention and / or a method according to the invention, within the virtual reality headset, or within the augmented reality headset, to calculate a current optical transfer function of a camera module (for example, included in the virtual reality headset or the augmented reality headset).

[0054] Alternatively, the imaging system is included in a medical imaging device, for example an endoscope.

[0055] In this case, the invention also relates to the use of a processing module according to the invention and / or a method according to the invention, within the medical imaging device, to calculate a current optical transfer function of a camera module (for example, included in the medical imaging device).

[0056] Alternatively, the imaging system is included in a vehicle, autonomous or not, such as a land vehicle (e.g., a car), an aircraft (e.g., a drone, a plane, a helicopter, etc.) or a maritime vehicle (e.g., a boat, a submarine, etc.).

[0057] In this case, the invention also relates to the use of a processing module according to the invention and / or a method according to the invention, within the vehicle, to calculate a current optical transfer function of a camera module (for example, included in the vehicle). Brief description of the figures

[0058] The invention will be better understood upon reading the following description, given solely by way of non-limiting example and made with reference to the accompanying drawings in which:

[0059] Figure 1 is a schematic representation of a characterization device according to the invention; and

[0060] Figure 2 is a flowchart of a characterization process implemented by the device in Figure 1.

[0061] It is understood that the embodiments described below are by no means exhaustive. In particular, variants of the invention may be conceived comprising only a selection of the features described below, isolated from the other features described, if this selection of features is sufficient to confer a technical advantage or to differentiate the invention from the prior art. This selection includes at least one preferably functional feature without structural details, or with only a portion of the structural details if this portion alone is sufficient to confer a technical advantage or to differentiate the invention from the prior art.

[0062] In particular, all the variants and embodiments described can be combined with each other if there are no technical obstacles to this combination.

[0063] In the figures and in the rest of the description, elements common to several figures retain the same reference. Detailed description

[0064] An imaging system 2 according to the invention is illustrated by Figure 1.

[0065] The imaging system 2 includes a processing module 3 and at least one camera module 4.

[0066] Imaging system 2 is, for example, a digital camera, a camcorder, a smartphone, a tablet, a computer, etc.

[0067] More specifically, each camera module 4 includes a lens 6 associated with a sensor 8, the sensor 8 being arranged to acquire an image of a scene observed through lens 6. Lens 6 and sensor 8 are, for example, mounted in a housing 10 to form said camera module 4.

[0068] For each camera module 4, the corresponding lens 6 belongs to a predetermined lens family, associated with a respective target geometry. Consequently, for each camera module 4, the corresponding lens 6 is associated with a predetermined target geometry (namely, the target geometry associated with the lens family to which said lens 6 belongs). Such a target geometry depends, in particular, on the desired performance and / or behavior of the lens 6 and, by extension, of the camera module 4.

[0069] The processing module 3 is connected to each camera module 4 to receive the images acquired by said camera module 4.

[0070] In addition, the processing module 3 is configured to determine a current optical transfer function of each camera module 4, associated with a current configuration of the camera module 4, from configuration data relating to said camera module 4 and a predetermined optical transfer function estimation function (hereafter referred to as the "estimation function").

[0071] Preferably, the processing module 3 is also configured to process at least one image acquired by an associated camera module, to generate a corresponding corrected image. Determination of estimation functions

[0072] The determination of each estimation function, using a characterization device 11, will now be described.

[0073] As illustrated by Figure 1, the characterization device 11 comprises a memory 12 and a computing unit 14 connected together. Memory 12

[0074] Memory 12 is configured to store at least one optical transfer function estimation function 16.

[0075] For the purposes of this invention, the term "optical transfer function" of a camera module can refer to the optical transfer function of its lens, which can be defined as its Point Spread Function (PSF). The term "optical transfer function" of a camera module can also refer to the spatial Fourier transform of the point spread function of its lens, or its Modulation Transfer Function (MTF), that is, a ratio of the intensity of the optical contrast obtained depending on the repetition rate of the image patterns captured by the lens.

[0076] Furthermore, the optical transfer function of a camera module 4 is understood as the value of the optical transfer function of the lens 6 of said camera module 4 at at least a set of predetermined points of the corresponding sensor 8.

[0077] Preferably, memory 12 is also configured to store simulation software 18.

[0078] Preferably, memory 12 is configured to store a database 20. Estimation function 16

[0079] Each estimation function 16 has been previously determined to associate, to a given configuration data set, indicative of a configuration of a camera module 4, a corresponding optical transfer function.

[0080] Such configuration data includes camera module 4 usage parameters and camera module 4 realization parameters.

[0081] For the purposes of this invention, a "operating parameter" of a camera module is understood to mean a parameter representative of the image acquisition conditions of said camera module. The operating parameters of a camera modules are, for example, and without limitation, conditioned by settings selected by an operator using said camera module.

[0082] For example, the usage parameters of a camera module include at least one of the following: - a position in a plane of the corresponding sensor; - a distance between the lens and the corresponding sensor; - at least one angle between a plane of the sensor and an optical axis of the lens; - a distance between an imaged point and the objective; - a focusing distance for which the imaged point of the scene is sharp; - a change of diaphragm; - continuous adjustment of the aperture of a diaphragm; - a possible optical zoom adjustment; - a position of the center of the lens relative to the sensor; - a position of the point of intersection of the optical axis of the lens with the sensor; and - a band of wavelength of light coming from an imaged point and / or an imaged scene.

[0083] The case of possible optical zoom adjustment concerns in particular certain optical modules with two movable optical groups, to adjust both the focusing distance and the optical zoom (in particular periscopes which allow this thanks to their length).

[0084] Furthermore, by "realization parameter" of a camera module, it is understood, within the meaning of the present invention, to be a parameter representative of a geometry of the respective lens.

[0085] Such realization parameters depend on the geometry of the objective, in particular the shape, refractive indices and positions of the optical elements (lenses, structured phase objects, etc.) included in said objective.

[0086] Preferably, the realization parameters of a camera module include imaging information and / or geometric information relating to said camera module.

[0087] The imaging information of a camera module is representative of an image of a predetermined pattern (called a "target") by the lens of said camera module, in particular in the plane of the respective sensor.

[0088] The way light propagates through a lens depends on the geometry of that lens. Therefore, imaging information is indeed dependent on the geometry of the lens, and is indirectly representative of that geometry.

[0089] Furthermore, the geometric information of a camera module is directly representative of the geometry of the respective lens.

[0090] For example, such geometric information includes at least one of the following: - a thickness of at least one phase object (for example, a lens) of the objective; - a refractive index of at least one phase object of the objective; - a decentering of at least one face of at least one phase object of the objective; - an angle between the respective axes of rotation of the generating curves of each face of at least one phase object of the lens; and - at least one data point representing a shape deviation of at least one face of at least one phase object of the objective relative to the corresponding part of the associated target geometry.

[0091] Examples of targets leading to the acquisition of the previously mentioned imaging information will now be described. First example

[0092] According to a first example, the target comprises a series of light and dark bands juxtaposed and parallel to each other.

[0093] In this case, the imaging information includes: - the image of the target through the lens; and / or - a contrast determined from said image. Second example

[0094] According to a second example, the test pattern comprises several sets of luminous lines, at least two sets of lines with different colors (preferably green, red, or blue). For example, the lines in each set are parallel to each other, forming a grid, preferably a regular grid. For example, the same grid appears in several positions within the test pattern, possibly with different orientations.

[0095] In this case, the imaging information includes: - the image of the target through the lens; - a specific contrast, for each color, based on said image; and / or - for each color, a positional shift between lines displaying that color.

[0096] In this example (as in the following ones), a shift is likely to be understood as a relative shift, so as to encode only the chromatic dispersion. Alternatively, such a shift is likely to be an absolute shift with respect to an ideal projection grid, so as to also characterize the geometric distortion of the lens, such as barrel distortion (in which a square in a plane perpendicular to the lens axis is projected with a shape whose sides are curved outwards), pincushion distortion (in which the same square is projected with a shape whose sides are curved inwards), or mustache distortion.

[0097] Such a target is advantageous, insofar as it gives access, for each colour, to information relating to chromatic aberrations of the lens, in particular at several positions in the plane of the sensor 8.

[0098] In addition, regularly spaced lines provide indications of geometric distortions in each detected wavelength band (red, green and / or blue).

[0099] Alternatively, the lines exhibit a color that varies over time. Third example

[0100] According to a third example, the target comprises a plurality of luminous points, for example regularly spaced.

[0101] According to one variant, the color of the dots is constant over time. According to another variant, the color of the dots varies over time.

[0102] In this case, the imaging information includes: - the image of the target through the lens; - a specific contrast, for each color, based on said image; and / or - for each color, a positional shift between points displaying said color.

[0103] Preferably, the points are of a size such that their image by the lens, on the associated sensor, is smaller than or equal to the size of the pixels of said sensor.

[0104] Such a target highlights, in particular, the geometric aberrations of the lens. Fourth example

[0105] According to a fourth example, the target comprises a plurality of lines forming a lattice.

[0106] In this case, the imaging information includes: - the image of the target through the lens; - a contrast determined from said image; and / or - a shift in position and / or a deformation of the lattice lines.

[0107] Such a target is advantageous, insofar as it highlights the geometric aberrations of the lens.

[0108] The color of the lattice lines is likely to change over time. In this case, the imaging information described previously is determined for each color.

[0109] Preferably, the estimation function 16 is an artificial intelligence model (hereafter referred to as a "model"), such as a convolutional neural network.

[0110] In this case, the model has been previously trained to produce, as output, an optical transfer function, when a configuration dataset (preferably including usage parameters and imaging information) is provided as input to the model.

[0111] For example, the model was previously trained using database 20, the structure of which will be described later. In this case, during model training, each configuration dataset from database 20 is provided as input to the model, with the corresponding optical transfer function forming an expected output for said configuration dataset.

[0112] Alternatively, or additionally, each estimation function 16 is a mathematical regression function, for example a polynomial, which takes as input the variables from the configuration dataset relating to a camera module (preferably the usage parameters and imaging information relating to the camera module), and produces as output the corresponding optical transfer function.

[0113] In this case, each estimation function 16 has been previously optimized (according to any appropriate optimization method) to minimize a difference between, on the one hand, the value taken by said mathematical regression function for a given configuration dataset and, on the other hand, the optical transfer function corresponding to said dataset of configuration. For example, each estimation function 16 was optimized from the database 20.

[0114] Preferably, each estimation function 16 is associated with a respective family of objectives. In this case, for each family of objectives, the data used to obtain the respective estimation function 16 relate to said family of objectives. Simulation software 18

[0115] The simulation software 18 is configured to calculate, during its execution, an optical transfer function of a predetermined simulated camera module, defined by corresponding configuration data.

[0116] Each simulated camera module includes a virtual lens associated with a virtual sensor. In this case, the simulation software 18 is configured to calculate, during its execution, the optical transfer function of the simulated camera module at at least one set of predetermined points on the virtual sensor.

[0117] More specifically, for each simulated camera module, the simulation software 18 is configured to calculate the associated optical transfer function from corresponding configuration data.

[0118] Preferably, for each simulated camera module, the simulation software 18 is configured to calculate the corresponding optical transfer function from relevant operating parameters and geometric information. In particular, when the simulated camera module is associated with a predetermined lens family, the geometric information depends on the respective target geometry.

[0119] The calculated optical transfer functions are advantageously used to populate database 20.

[0120] Advantageously, the simulation software 18 is also configured to calculate, during its execution, the imaging information for at least one simulated camera module. In particular, the simulation software 18 is configured to calculate the imaging information for said simulated camera module for at least one test pattern.

[0121] In this case, the calculated imaging information is also used to populate database 20. Database 20

[0122] Database 20 is configured to store at least one configuration dataset, representative of a camera module, in association with the corresponding optical transfer function.

[0123] Preferably, each configuration dataset includes: - realization parameters (advantageously imaging information) of the respective camera module; and - usage parameters of the respective camera module.

[0124] In addition, each configuration dataset is preferably associated, in database 20, with the class of lenses to which the lens of the camera module corresponding to said configuration dataset belongs.

[0125] Preferably, at least one camera module in database 20 is a simulated camera module as described previously. Obtaining such a simulated camera module will be described later. Calculation unit 14

[0126] The computing unit 14 is configured to simulate at least one camera module, and to calculate the optical transfer function corresponding to each simulated camera module.

[0127] Each simulated camera module is associated with respective configuration data, including realization parameters and usage parameters.

[0128] In particular, computing unit 14 is configured to simulate at least one camera module associated with a respective class of lenses.

[0129] In this case, for each lens class, the computing unit 14 is configured to simulate at least one associated camera module by selecting: - for at least one geometric information of the simulated camera module, a value different from the corresponding value in the target geometry associated with said lens class; and / or - for at least one usage parameter, a value likely to be reached when using an actual camera module associated with said class of lenses.

[0130] Preferably, in order to generate each simulated camera module, the computing unit 14 is configured to select the value of each geometric information within a range compatible with the expected deviations in the geometry of the actual lenses from the corresponding target geometry.

[0131] For example, for a given class of lenses, the computing unit 14 is configured to simulate each corresponding camera module by successively varying the value of at least one geometric information within the associated range.

[0132] In addition, as previously stated, the computing unit 14 is configured to select, for at least one usage parameter, a value likely to be reached when using the corresponding actual camera module.

[0133] The computing unit 14 is also configured to implement the simulation software 18 in order to obtain, for each simulated camera module, a respective optical transfer function.

[0134] The computing unit 14 is also configured to write, in database 20, each calculated optical transfer function, in correspondence with the corresponding configuration data, to form a configuration dataset.

[0135] Preferably, for each simulated camera module, the computing unit 14 is further configured to calculate corresponding imaging information, for example by implementing the simulation software 18.

[0136] In this case, for each simulated camera module, the corresponding configuration dataset written to the database 20 by the computing unit 14 advantageously includes imaging information of said camera module obtained by implementing the simulation software 18.

[0137] This characteristic is advantageous because the estimation function obtained from such configuration data depends directly on the imaging information. Since determining this imaging information for a camera module to be characterized is straightforward, calculating the corresponding optical transfer function is also easy.

[0138] The computing unit 14 is, moreover, configured to calculate each estimation function 16, in particular from the database 20. Characterization of camera modules

[0139] To characterize each camera module 4 of the imaging system 2, the respective processing module 3 is configured to implement a characterization process 30, illustrated by Figure 2.

[0140] The characterization process 30 includes a step 32 of obtaining configuration data (called the "obtaining step") and a calculation step 34.

[0141] Preferably, the characterization process 30 further includes a correction step 36.

[0142] Preferably, the implementation of the characterization process 30 is preceded by an initialization step, during which, for each camera module 4 of the imaging system 2, the corresponding estimation function 16 is stored in a memory of the imaging system 2.

[0143] Preferably, the imaging information from each camera module 4 is stored in the memory of the imaging system 2. Obtaining step 32

[0144] The processing module 3 is configured to, during the acquisition step 32, obtain configuration data indicative of a current configuration of a camera module 4 of the imaging system 2 to be characterized.

[0145] The configuration data obtained includes: - at least one realization parameter representative of an actual geometry of the lens 6 of the camera module 4; and - at least one usage parameter representative of image acquisition conditions of camera module 4.

[0146] In particular, the computing unit 14 is capable of directly obtaining each usage parameter from the settings of the imaging system 2, for example the settings of the imaging system 2 for the acquisition of an image using the camera module 4.

[0147] In addition, each realization parameter of the camera module 4 preferably belongs to a predetermined set of camera module realization parameters, previously measured (for example, during its manufacture), and stored in a memory accessible for reading by the processing module 3, for example a memory of the imaging system 2.

[0148] In particular, the realization parameters are imaging information from camera module 4, determined from at least one imaged target using camera module 4.

[0149] For example, the implementation of acquisition step 32, for a camera module 4 of the imaging system, is triggered by the acquisition of an image using said camera module 4.

[0150] In addition, the processing module 3 is configured to calculate, during calculation step 34, a current optical transfer function associated with the camera module 4, from the current configuration of the camera module 4.

[0151] In particular, processing module 3 is configured to calculate the current optical transfer function from: - configuration data obtained during retrieval step 32; and - of at least one estimation function 16 stored in the memory of the imaging system 2.

[0152] In particular, to perform such a calculation, the processing module 3 is configured to load the estimation function 16 associated with the lens family to which the lens 6 of the camera module 4 to be characterized belongs.

[0153] As previously stated, the estimation function is, in particular, a mathematical regression function or an artificial intelligence model.

[0154] Furthermore, the processing module 3 is configured to provide, as input to the estimation function 16, the previously obtained configuration data. This results in an estimated current optical transfer function for the camera module 4. Correction step 36

[0155] In addition, the processing module 3 is configured to, during the correction step 36, correct an image acquired by the camera module 4 from the corresponding estimated current optical transfer function.

[0156] In particular, the processing module 3 is configured to generate a corrected image by deconvolving the acquired image using the estimated optical transfer function.

[0157] Optionally, the processing module 3 is also configured to correct, during correction step 36, all or part of the operating parameter values ​​obtained during acquisition step 32. This is because errors may affect such values. For example, the obtained value for the focal length of lens 6 may deviate slightly from its actual value. Similarly, the obtained value for the distance between an imaged point and lens 6 may differ from the actual value.

[0158] In particular, the processing module 3 is configured to perform such a correction based on a corrected image quality. In this case, the image quality is quantified using a predetermined quality indicator, such as image sharpness or image chromatic aberration.

[0159] Preferably, to perform such a correction, the processing module 3 is configured to calculate a variation of the corrected image quality indicator as a function of variations of at least one usage parameter in the vicinity of its value obtained during the acquisition step 32.

[0160] In other words, processing module 3 is configured to: - vary the value of each usage parameter in the vicinity of its value obtained during the obtaining step 32; - calculate, from the estimation function 16 and the modified value of each usage parameter, a new optical transfer function, called the "intermediate optical transfer function"; - to generate, from the acquired image and each intermediate optical transfer function, a corresponding intermediate corrected image; and - for each intermediate corrected image, calculate the corresponding value of the quality indicator.

[0161] Furthermore, the processing module 3 is configured to deduce, from the aforementioned variation in the quality indicator, a gradient of the quality indicator as a function of each modified usage parameter. This results in a gradient field for each area of ​​the acquired image.

[0162] The processing module 3 is also configured to implement an optimization method (e.g., the Levenberg-Marquardt algorithm) in order to calculate, from the gradient field, a corrected value of each usage parameter.

[0163] Furthermore, the processing module 3 is configured to calculate an updated optical transfer function from the estimation function 16 and the corrected value of each usage parameter. In this case, the processing module 3 is further configured to update the corrected image by deconvolving the acquired image with the updated optical transfer function.

[0164] Preferably, in addition to the usage parameters, the processing module 3 is also preferably configured to vary all or part of the realization parameters in order to obtain the updated corrected image.

[0165] Such a correction is advantageous, insofar as it gives the characterization process 30 greater robustness to biases or deviations during the acquisition of configuration data (which include, as previously indicated, usage parameters and realization parameters), such as errors in estimating the depth of objects or areas of the scene, errors in estimating the focal length of the camera module (determined from the lens-sensor distance, the measurement of which is likely to exhibit drifts or imperfections between factory output and use), etc.

[0166] Alternatively, the estimation function 16 associated with each camera module 4 and / or the imaging information of each camera module 4 are not stored in the memory of the imaging system 2, but in a remote memory.

[0167] In this case, the processing module 3 includes a telecommunications device configured to query the remote memory and load each estimation function 16 and / or each imaging information required for the implementation of the characterization process 30. Functioning

[0168] During a preliminary calculation step, the characterization device 11 is implemented to determine the estimation function associated with each family of objectives.

[0169] Then, during the initialization step, for each imaging system 2, and for each corresponding camera module 4, the corresponding estimation function 16 (i.e., associated with the family to which the respective objective belongs) is recorded in the imaging system 2.

[0170] Then, during the acquisition step 32, the processing module 3 obtains the configuration data relating to a camera module 4 to be characterized.

[0171] Such a camera module to be characterized is, for example, a camera module 4 used to acquire a given image. In this case, the configuration data is indicative of the configuration of the camera module 4 during the acquisition of said image.

[0172] Then, during calculation step 34, the processing module 3 calculates the current optical transfer function associated with the camera module 4, from the current configuration of said camera module 4.

[0173] In particular, to carry out such a calculation, the processing module 3 loads the estimation function 16 associated with the lens family to which the lens 6 of the camera module 4 to be characterized belongs, and provides, as input to said estimation function 16, the configuration data previously obtained during the acquisition step 32.

[0174] This results in an estimated current optical transfer function for camera module 4.

[0175] Then, during the correction step 36, the processing module 3 corrects the image acquired by the camera module 4 from the corresponding estimated current optical transfer function.

[0176] In particular, the processing module 3 generates a corrected image by deconvolving the acquired image using the estimated optical transfer function.

[0177] In the case where the estimation function 16 is not stored in the imaging system 2, but in a remote memory, then, during the calculation step 34, the processing module 3 queries the remote memory and loads the estimation function 16 required for the implementation of the calculation step 34 and / or the correction step 36.

[0178] Preferably, during correction step 36, processing module 3 updates the resulting corrected image. To do this, processing module 3: - varies a value of all or part of the configuration data in the vicinity of the corresponding obtained value; - calculates, from the estimation function 16 and the modified value of each configuration data, a corresponding intermediate optical transfer function; - generates, from the acquired image and each intermediate optical transfer function, a corresponding intermediate corrected image; - calculates, for each intermediate corrected image, a corresponding value of the predetermined quality indicator; - calculates a gradient of the quality indicator as a function of each configuration data point; - determines, from each calculated gradient, a corrected value for each configuration data point; - calculates an updated optical transfer function from the estimation function 16 and the corrected value of each configuration data point; and - updates the deconvolution-corrected image of the image acquired by the updated optical transfer function.

[0179] Of course, the invention is not limited to the examples just described.

Claims

DEMANDS 1. Method (30) for characterizing a camera module (4) comprising a lens (6) and a sensor (8), the sensor (8) being arranged to acquire an image of a scene observed through the lens (6), the lens (6) belonging to a predetermined family of lenses associated with a respective target geometry, the characterization method (30) being implemented by computer and comprising the steps: - obtaining (32) indicative configuration data of a current camera module configuration, the configuration data including: • at least one realization parameter representative of an actual geometry of the objective (6); and / or • at least one usage parameter representative of common image acquisition conditions of the camera module (4); and - calculation (34) of a current optical transfer function, associated with the current configuration of the camera module (4), from: • configuration data obtained; and • of a predetermined estimation function (16) associated with the family of objectives and taking, as input, configuration data, and providing, as a result, a corresponding estimated optical transfer function.

2. Method according to claim 1, wherein at least one embodiment parameter comprises imaging information representative of an image of a pattern predetermined by the lens (6) of the camera module (4), preferably in the plane of the respective sensor (8).

3. A method according to claim 1 or 2, wherein the estimation function (16) is an artificial intelligence model previously trained from a database associating at least one configuration dataset with a corresponding optical transfer function, each dataset configuration being provided to the artificial intelligence model as input, the corresponding optical transfer function being provided to the artificial intelligence model as expected output.

4. Method according to claim 1 or 2, wherein the estimation function (16) is a mathematical regression function previously obtained from a predetermined mathematical function and a database associating at least one configuration dataset with a corresponding optical transfer function, and determined so as to minimize a difference between, on the one hand, the value taken by the mathematical function for a given configuration dataset and, on the other hand, the optical transfer function corresponding to said configuration dataset.

5. Method according to claim 3 or 4, wherein each configuration data set includes imaging information representative of an image of a pattern predetermined by a lens of a camera module having an optical transfer function equal to the optical transfer function associated with said configuration data set.

6. A method according to any one of claims 1 to 5, comprising, prior to the calculation step, an initialization step including: - for at least one lens family, modeling of at least one simulated camera module comprising a virtual lens associated with a virtual sensor, the simulated camera module presenting configuration data including at least one realization parameter whose value is different from the corresponding value in the target geometry associated with said lens family; - for each virtual objective modeled, calculation of a corresponding simulated optical transfer function; - storage, in a database (20), of each optical transfer function in association with the corresponding configuration data.

7. A method according to any one of claims 1 to 6, further comprising generating a corrected image from an image acquired by means of the camera module (4), as a function of the current calculated optical transfer function corresponding to said camera module (4).

8. A method according to claim 7, wherein the generation of the corrected image includes a deconvolution of the image acquired by the calculated current optical transfer function.

9. A method according to any one of claims 1 to 8, further comprising an update of the corrected image comprising the following steps: - variation of a value of all or part of the configuration data in the vicinity of the corresponding obtained value; - calculation, from the estimation function (16) and the modified value of each configuration data, of a corresponding intermediate optical transfer function; - generation, from the acquired image and each intermediate optical transfer function, of a corresponding intermediate corrected image; - for each intermediate corrected image, calculation of a corresponding value of a predetermined quality indicator; - calculation of a gradient of the quality indicator as a function of each configuration data; - determination, from each calculated gradient, of a corrected value for each configuration data point; - calculation, from the estimation function (16) and the corrected value of each configuration data point, of an updated optical transfer function; and - updating the corrected image by deconvolution of the image acquired by the updated optical transfer function.

10. A method according to any one of claims 1 to 9, wherein the embodiment parameters comprise at least one geometric information including: - a thickness of at least one lens of the objective (6); - a refractive index of at least one lens of the objective (6); - a decentering of at least one face of at least one lens of the objective (6); - an angular separation between a mean plane of each face of at least one lens; and / or - at least one data point representing a shape deviation of at least one face of at least one lens of the objective (6) with respect to the corresponding part of the target geometry.

11. A method according to any one of claims 1 to 10, wherein the operating parameters comprise at least one of the following: - a position in a plane of the sensor (8); - a distance between the lens (6) and the sensor (8); - at least one angle between a plane of the sensor (8) and an optical axis of the lens (6); - a distance between an imaged point and the objective (6); - a focusing distance for which the imaged point of the scene is sharp; - a change of diaphragm; - continuous adjustment of the aperture of a diaphragm; - a possible optical zoom adjustment; - a position of a center of the lens (6) relative to the sensor (8); - a position of a point of intersection of the optical axis of the lens (6) with the sensor (8); and - a band of wavelengths of light coming from at least one imaged point.

12. Computer program comprising executable instructions which, when executed by computer, implement the steps of the process (30) according to any one of claims 1 to 11.

13. Processing module (3) configured such that, for at least one camera module (4), each camera module (4) comprising a lens (6) and a sensor (8), the sensor (8) being arranged to acquire an image of a scene observed through the lens (6), the lens (6) belonging to a predetermined lens family associated with a respective target geometry: - obtain indicative configuration data for a current configuration of the camera module (4), the configuration data including: • at least one realization parameter representative of an actual geometry of the objective (6); and / or • at least one usage parameter representative of common image acquisition conditions of the camera module (4); and - calculate a current optical transfer function, associated with the current configuration of the camera module (4), from: • configuration data obtained; and • of a predetermined estimation function (16) associated with the family of objectives and taking, as input, configuration data, and providing, as a result, a corresponding estimated optical transfer function.

14. Processing module (3) according to claim 13, also configured so as to, for each camera module (4), generate a corrected image from an image acquired by means of the camera module (4), according to the current calculated optical transfer function corresponding to said camera module (4).

15. Imaging system comprising at least one camera module (4) and one processing module (3) according to claim 13 or 14 connected to each camera module (4) to receive configuration data relating to said camera module (4) and, preferably, at least one image acquired by said camera module (4).

Images