Method and system for generating a two-way texture function for an object

Abstract

The present invention relates to a method for generating a bidirectional texture function (BTF) for an object, said method comprising at least the following steps: measuring an initial BTF (103) for the object using a camera-based measuring device, - obtaining spectral reflection data (105) for the object using a spectrophotometer for a pre-given number of different measurement geometries, - adapting the initial BTF (103) to the obtained spectral reflection data (105) so as to obtain an optimized BTF (107). The present invention also relates to a corresponding system for generating a bidirectional texture function for an object.

Description

Technical Field

[0001] The present disclosure relates to a method for generating a bidirectional texture function (BTF) of an object, in particular a physical car paint sample. The present disclosure also relates to a corresponding system and a corresponding computer system. Background Art

[0002] The current automotive paint color design process is based on physical samples of automotive paint applied to the most common small flat panels. Working only with physical samples has several disadvantages. Painting samples is expensive and time consuming. In addition, due to cost, only small flat panels are painted, and it is difficult to infer from small samples how the coating will look in different shapes (e.g., car bodies) or different lighting settings. Automotive paints are often selected as effect colors with angle-dependent color effects, especially effects caused by interference and / or metallic pigments (such as metallic flake pigments) or special effect flake pigments (such as pearlescent flake pigments).

[0003] Using a digital model of the appearance of car paint, it is possible to computer generate images of car paint applied to arbitrary shapes under arbitrary lighting conditions. The bidirectional texture function (BTF) represents such a digital model, which can also capture the spatially varying appearance of car paint, such as shimmer. Based on a computer-generated image of car paint applied to an object, it is possible to virtually assess the characteristics of the color of the car paint.

[0004] The BTF is a representation of the appearance of a texture as a function of viewing and illumination directions (i.e., viewing angle and illumination angle). It is an image-based representation because the geometry of the surface of the object under consideration is unknown and unmeasured. The BTF is typically acquired by imaging the surface at a sampling of a hemisphere of possible viewing and illumination directions. The BTF measurement result is a collection of images. The BTF is a 6-dimensional function. (Dana, Kristin J., Bram van Ginneken, Shree K. Nayar, and Jan J. Koenderink. "Reflectance and Texture of Real-World Surfaces". ACM Transactions on Graphics 18, No. 1 (January 1, 1999): 1–34. https: / / doi.org / 10.1145 / 300776.300778 .) Summary of the invention

[0005] Until now, BTF for colors (paints) has been measured using camera-based measurement devices. The camera-based measurement devices used are configured to quickly acquire reflectance data for many measurement geometries. The devices used are also able to acquire spatially varying aspects of automotive paints, such as the sparkle of effect pigments or the texture of structured clear coats. However, it has been found that the color information is not accurate and reliable enough for color design review use cases.

[0006] It is therefore an object of the present disclosure to provide color information more accurately and in particular to provide the possibility to further optimize the measured BTF.

[0007] A method, a system and a computer system for generating a bidirectional texture function for an object are provided by the features of the independent claims, respectively. Further features and embodiments of the claimed method and system are described in the dependent claims and in the description.

[0008] According to claim 1, there is provided a method for generating a bidirectional texture function (BTF) of an object, the method comprising at least the following steps:

[0009] - measuring an initial BTF for the object using a camera-based measurement device,

[0010] - acquiring spectral reflectance data for the object using a spectrophotometer for a predetermined number (i.e. a finite number) of different measurement geometries,

[0011] - Adapting the initial BTF to the acquired spectral reflectance data to obtain an optimized BTF.

[0012] In order to solve the problem of insufficient color accuracy considering the above mentioned, according to the claimed method, it is proposed to obtain an initial BTF of an object, in particular of a physical automotive paint sample, using a camera-based measurement device in a first step. Then, in a second step, a second spectral measurement of the same sample is performed using a spectrophotometer, in particular a handheld spectrophotometer. Thus, additional more accurate spectral reflectance data for a small number (e.g. <25) of measurement geometries is obtained. The initial BTF is then enhanced with more accurate but sparse spectral reflectance data. The result is a BTF that obtains color and spatially varying appearances, such as the sparkle of an automotive paint sample, and is still sufficiently accurate.

[0013] According to one embodiment of the claimed method, a camera-based measurement device creates multiple images (photos) of an object / sample at different viewing angles, different illumination angles, different illumination colors, and / or for different exposure times, thereby providing multiple measurement data taking into account multiple combinations of illumination angles, viewing angles, illumination colors, and / or exposure times. The camera-based measurement device may be a commercially available measurement device such as, for example, an X-Rite A small plate coated with the car paint sample and the clear coat was inserted into the measuring device and the measuring process started. From the measurement results and subsequent post-processing, an initial BTF was obtained.

[0014] During the post-processing, images / photos with different illumination colors and different exposure times but with equal illumination angles and viewing angles are combined into images with high dynamic range, respectively. Further, the viewing angle of the photo on the sample is corrected. Based on the data obtained from the photo and post-processing, the parameters of the initial BTF are determined.

[0015] According to a further embodiment of the claimed method, adapting the initial BTF to the acquired spectral reflectance data to obtain an optimized BTF comprises splitting the initial BTF into different terms, each term comprising a set of parameters, and optimizing the parameters of each term individually using the acquired spectral reflectance data.

[0016] Thus, the initial BTF is split (divided) into two main terms, the first being a uniform bidirectional reflectance distribution function (BRDF) which describes the reflectance properties of an object (e.g., a car paint sample) depending only on the measurement geometry; and the second being a texture function which accounts for the spatially varying appearance of the object, i.e., it adds a viewing and illumination dependent texture image. The texture image stored in the model has the property that, on average, the sum of the intensities in each of the RGB channels across all pixels is zero. When viewed from a distance, the overall color impression of the car paint is not determined by the color at a single point, but by the average color of a larger area. Due to the properties mentioned above, it is assumed that the average color across a larger area of ​​the texture image is zero or close to zero. This allows the texture image to be overlaid without changing the overall color. This also means that the texture image can be ignored when optimizing the BTF.

[0017] For the representation of BTF, we use the color model first introduced by Rump et al. (Rump, Martin, Ralf Sarlette, und Reinhard Klein., Efficient Resampling, Compression and Rendering of Metallic and Pearlescent Paint. InVision, Modeling, and Visualization, 11–18, 2009):

[0018]

[0019] x: surface coordinate of the sample / object

[0020] Illumination and observation / viewing direction at the base coating of the sample

[0021] Color table depending on illumination and viewing direction

[0022] a: Albedo or diffuse reflectivity

[0023] The kth Cook-Torrance lobe; the Cook-Torrance lobe is a commonly used BRDF to describe the glossiness of microfacet surfaces.

[0024] S k : Weight for the kth Cook-Torrance lobe

[0025] α k : Parameters of the Beckmann distribution for the kth Cook-Torrance lobe

[0026] F 0,k : Fresnel reflectivity for the kth Cook-Torrance lobe

[0027] Image table of spatial textures depending on illumination and viewing direction

[0028] In general, a bidirectional reflectance distribution function (BRDF) is a function of four real variables that defines how light reflects off an opaque surface. The function takes the incident light direction and the emission direction And return along The outgoing reflected radiation is related to the direction The ratio of the irradiance incident on a surface. BRDF refers to a collection of photometric data for any material (meaning an object in this context, i.e., a paint sample) that describes the photometric reflective light scattering properties of the material (object) as a function of the illumination angle and the reflective scattering angle. BRDF describes the spectral and spatial reflective scattering properties of an object (particularly the angle-dependent chromatic material comprised by the object), and provides a description of the appearance of the material, and many other appearance properties such as gloss, haze, and color can be easily derived from BRDF.

[0029] Typically, a BRDF includes three color coordinates as a function of the scattering geometry. When processing a BRDF, the specific light source and color system (such as CIELAB) must be specified and included with any data.

[0030] The data contained in the BTF generated by the method proposed by the present disclosure can be used for various purposes. Absolute color or reflectance data can be used in conjunction with pigment mixture models to help formulate paints containing effect flake pigments to evaluate and ensure color matching under various lighting and viewing conditions, for example, between a car body and a bumper.

[0031] Effect flake pigments include metal flake pigments, such as aluminum flakes, coated aluminum flakes, copper flakes, etc. Effect flake pigments also include special effect flake pigments that cause hue shifts, such as pearlescent pigments, such as mica flakes, glass flakes, etc.

[0032] As can be appreciated from equation (1), the first term, ie, the BRDF is divided into corresponding to the color map The first subterm and corresponds to the intensity function The parameters of the initial BTF are optimized to minimize the color difference between the spectral reflectance data and the initial BTF by optimizing the parameters of the color table in a first optimization step while the parameters of the intensity function are kept constant and by optimizing the parameters of the intensity function in a second optimization step while the parameters of the color table are kept constant.

[0033] Spectral reflectance data, i.e., spectral reflectance curves, are acquired only for a limited number of measurement geometries. Each such measurement geometry is defined by a specific illumination angle / direction and a specific viewing angle / direction. Spectral reflectance measurements are performed, for example, by a handheld spectrophotometer, such as, for example, a Byk-Mac with six measurement geometries (fixed illumination angles and viewing / measurement angles of -15°, 15°, 25°, 45°, 75°, 110°). X-Rite with twelve measurement geometries (two illumination angles and six measurement angles) or X-Rite MA (Two illumination angles and up to eleven measurement angles.) The spectral reflectance data obtained from these measurement devices is more accurate than the color information obtained from camera-based measurement devices.

[0034] According to an embodiment of the claimed method, for optimization of the color table in a first optimization step for each spectral measurement geometry, a first CIEL*a*b* value is calculated based on the spectral reflectance data (curve), and a second CIEL*a*b* value is calculated based on the initial BTF, and a correction vector in the a* and b* coordinates is calculated by subtracting the second CIEa*b* value from the first CIEa*b* value, and the correction vectors are interpolated and extrapolated component by component for the complete viewing and illumination angle range stored in the color table, the interpolated correction vectors are applied to the initial BTF CIEL*a*b* values ​​for each spectral measurement geometry stored in the color table, and the corrected BTF CIEL*a*b* values ​​are transformed into linear sRGB coordinates that are normalized (so that their sum is, for example, equal to 3) and finally stored in the color table.

[0035] A multilevel B-spline interpolation algorithm (see Lee, Seungyong, George Wolberg, und Sung Yong Shin., "Scattered data interpolation with multilevel B-splines". IEEE transactions on visualization and computer graphics 3, No. 3 (1997): 228–244.) can be used for component-by-component interpolation and extrapolation of the correction vector.

[0036] According to a further embodiment of the claimed method, for optimizing the parameters of the intensity function in the second optimization step, a cost function is defined based on the sum of the color differences across all spectral reflectance measurement geometries. 0 ,a) is defined across all reflection measurement geometries according to the following equation:

[0037]

[0038] G: A set of measurement geometries for which spectral reflectance data is available

[0039] g: one of a set of measured geometries

[0040] ΔE(f Test ,f Ref ):Measure color f Test With f Ref The weighted color difference formula of the difference between

[0041] Reference colors derived from spectral measurements

[0042] Test color calculated from initial BTF for given illumination and viewing direction

[0043] α=(α 1 ,α 2 ,α 3 ): parameter vector of the Beckmann distribution for the three Cook-Torrance lobes

[0044] S=(S 1 ,S 2 ,S 3 ): Weight vector for the three Cook-Torrance lobes

[0045] F 0 =(F 0,1 ,F 0,2 ,F 0,3 ): Fresnel reflection vectors for the three Cook-Torrance lobes

[0046] P(α,S,F 0 ,a): Penalty function

[0047] As indicated in equation (2), the cost function may be supplemented by a penalty function designed to take into account certain constraints, such constraints preferably including keeping parameter values ​​within a valid range.

[0048] To calculate the color difference, the initial BTF is evaluated at different spectral reflectance measurement geometries, and the resulting CIEL*a*b* is compared with the CIEL*a*b* values ​​from the spectral reflectance measurement using a weighted color difference formula (such as, for example, the formula defined in DIN 6157 / 2), and the parameters of the intensity function are optimized using a nonlinear optimization method, such as the Nelder-Mead-Downhill-Simplex method, so that the cost function is minimized.

[0049] According to yet another embodiment, the first and the second optimization steps are repeatedly / iteratively run to further improve the accuracy of the optimized BTF. The number of iterations can be specified and predefined. It has been found that three iterations can already generate reliably good results.

[0050] It has been found that the optimized BTF is more accurate than the initial BTF obtained directly from the camera-based device. This is true not only for the few (limited number) spectral reflectance geometries that provide additional spectral reflectance data, but also for the full range of illumination and viewing directions.

[0051] The claimed method and system are applicable not only to the automotive paint color design process, but also to comparable processes, for example, in cosmetics and electronics.

[0052] The present disclosure also relates to a system for generating a bidirectional texture function (BTF) of an object, the system comprising:

[0053] - a camera-based measurement device configured to measure an initial BTF for said object,

[0054] a spectrophotometer configured to acquire spectral reflectance data for the object for a predetermined number of different measurement geometries,

[0055] a computing device communicatively connected to the camera-based measurement device and the spectrophotometer, respectively, and configured to receive the initial BTF and the acquired spectral reflectance data for the object via the respective communication connections, and to adapt the initial BTF to the acquired reflectance data, thereby obtaining an optimized BTF.

[0056] The camera-based measurement device may be a commercially available device such as, for example, an X-Rite (Total Appearance Capture). The camera-based measurement device is configured to acquire images / photographs of the object / sample at different illumination angles and different viewing angles as well as different illumination colors and different exposure times. Images with different illumination colors and exposure times can be combined into an image with a high dynamic range (HDR image). The image can be corrected for the viewing angle relative to the object / sample.

[0057] The spectrophotometer may be selected as a handheld spectrophotometer. The spectrophotometer is a multi-angle spectrophotometer.

[0058] The object may be a sample of automotive paint applied to a panel or any other paint, particularly a paint including angle-dependent materials such as effect flake pigments.

[0059] The system may also include a database configured to store an initial BTF, spectral reflectance data for an object for a predetermined number of different measurement geometries, and an optimized BTF. The computing device may be communicatively connected to the database in order to retrieve the initial BTF and the spectral reflectance data for the object for a predetermined number of different measurement geometries and store the optimized BTF. That means that before the computing device retrieves the initial BTF and the spectral reflectance data in order to adapt the initial BTF to the acquired reflectance data, thereby obtaining the optimized BTF, the initial BTF obtained from the camera-based measurement device and the spectral reflectance data acquired by the spectrophotometer may be first stored in the database. In such a scenario, the camera-based measurement device and the spectrophotometer may also be communicatively connected to the database. Therefore, the communication connection between the computing device and the camera-based measurement device and the communication connection between the computing device and the spectrophotometer may be a direct connection or an indirect connection via a database, respectively. Each communication connection may be a wired or wireless connection. Each suitable communication technology may be used. The computing device, the camera-based measurement device, and the spectrophotometer may each include one or more communication interfaces for communicating with each other. Such communication can be performed using a wired data transmission protocol, such as Fiber Distributed Data Interface (FDDI), Digital Subscriber Line (DSL), Ethernet, Asynchronous Transfer Mode (ATM), or any other wired transmission protocol. Alternatively, the communication can be wireless via a wireless communication network using any of a variety of protocols, such as General Packet Radio Service (GPRS), Universal Mobile Telecommunications System (UMTS), Code Division Multiple Access (CDMA), Long Term Evolution (LTE), Wireless Universal Serial Bus (USB), and / or any other wireless protocol. The corresponding communication can be a combination of wireless and wired communication.

[0060] The computing device may include or communicate with one or more input units, such as a touch screen, audio input, motion input, mouse, keyboard input, and / or the like. Further, the computing device may include or communicate with one or more output units, such as audio output, video output, screen / display output, and / or the like.

[0061] Embodiments of the present invention may be used with or incorporated into a computer system, which may be a stand-alone unit or include one or more remote terminals or devices that communicate with a central computer, such as in a cloud, via a network (such as, for example, the Internet or an intranet). As such, the computing devices and related components described herein may be part of a local computer system or a remote computer or online system, or a combination thereof. The databases and software described herein may be stored in a computer's internal memory or in a non-transitory computer-readable medium.

[0062] When optimizing a color table, for each spectral reflectance measurement geometry of the spectrophotometer, a correction vector is determined. The correction vector results are respectively the difference of the reflected radiance in the RGB channels from the BRDF portion of the initial BTF and the spectral reflectance data for the same geometry. The calculation of the correction vector is performed in the CIEL*a*b* color space. The resulting correction vector is interpolated component by component over the entire parameter range of the color table.

[0063] The claimed system is particularly configured to perform the claimed method.

[0064] To correctly reflect the texture of the car paint, the BTF includes a spatial texture image table that depends on the illumination and viewing angle / direction.

[0065] The present disclosure also relates to a computer system, comprising:

[0066] - computer unit;

[0067] - a computer-readable program having a program code stored in a non-transitory computer-readable storage medium, which, when the program is executed on the computer unit, causes the computer unit to perform the following:

[0068] - acquiring and receiving an initial BTF for an object and spectral reflectance data for the object, wherein the initial BTF is measured by a camera-based measurement device and the spectral reflectance data is acquired by a spectrophotometer for a predetermined number of different measurement geometries;

[0069] - matching the spectral reflectance data to the initial BTF by adapting the parameters of the initial BTF accordingly, thereby obtaining an optimized BTF.

[0070] The initial BTF for the object and the spectral reflectance data for the object may be obtained respectively by (a) measurement results of the object or (b) previous measurement data of the object from a database containing measurement results of the object.

[0071] Further aspects of the invention will be realized and attained by means of the elements and combinations particularly described in the appended claims.It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention as described. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 A flow chart illustrating a process that may be accomplished in accordance with an exemplary embodiment of the claimed method. DETAILED DESCRIPTION

[0073] The present disclosure provides a method and associated system for determining the BTF of a subject. Figure 1 A flow chart illustrating a process that may be performed according to various embodiments of the claimed method and system is provided. Starting at step 102, an object is placed in a camera-based measurement device for measuring an initial BTF 103 of the object. The initial BTF 103 is obtained as a result of the measurement. At step 104, the object is placed in a spectrophotometer that is configured to acquire corresponding spectral reflectance data for the object for a pre-given number of different measurement geometries. Thus, spectral reflectance data 105 (reflectance spectra for different spectral measurement geometries) are obtained for the object for a limited number of different measurement geometries of the spectrophotometer. When the reflectance spectra (reflectance data) have been acquired at each desired (pre-given) measurement geometry, at step 106, the initial BTF 103 is adapted to the acquired spectral reflectance data 105 by adapting the parameters of the initial BTF accordingly. Thus, an optimized BTF 107 is obtained.

Claims

1. A method for generating a bidirectional texture function of an object, the method at least comprises the following steps: - Measuring an initial bidirectional texture function (103) for the object using a camera-based measuring device, - Obtaining spectral reflectance data (105) for the object using a spectrophotometer for a pre-given number of different measurement geometries, - Adapting the initial bidirectional texture function (103) to the obtained spectral reflectance data (105), thereby obtaining an optimized bidirectional texture function (107), wherein the initial bidirectional texture function is represented by the following formula: - x: the surface coordinates of the object, - The irradiation and observation directions at the undercoat of the object - Color table depending on the irradiation and observation directions - a: albedo or diffuse reflectance, - the k-th Cook-Torrance lobe; the Cook-Torrance lobe is a commonly used bidirectional reflectance distribution function that describes the glossiness of a microfacet surface -S k : For the weight of the k-th Cook-Torrance lobe, -α k : Parameters for the Beckmann distribution for the k-th Cook-Torrance lobe -F 0,k : Fresnel reflectivity for the k-th Cook-Torrance lobe - Spatial texture image table depending on irradiation and viewing directions The first term of the initial bidirectional texture function is divided into a first sub-term corresponding to the color table and a second sub-term corresponding to the intensity function and the parameters of the initial bidirectional texture function (103) are optimized to minimize the color difference between the spectral reflectance data (105) and the initial bidirectional texture function (103) by optimizing the parameters of the color table in a first optimization step while the parameters of the intensity function remain constant and by optimizing the parameters of the intensity function in a second optimization step while the parameters of the color table remain constant.

2. The method according to claim 1, wherein, the camera-based measuring device creates a plurality of images of the object at different viewing angles, different illumination angles, for different illumination colors and / or for different exposure times, thereby providing a plurality of measurement data considering various combinations of illumination angle, viewing angle, illumination color and / or exposure time.

3. The method according to claim 2, wherein, images having different illumination colors and different exposure times but having equal illumination angles and viewing angles are respectively combined into images having a high dynamic range.

4. The method according to claim 1, wherein, for the optimization of the color table for each spectral measurement geometry, a first CIEL*a*b* value is calculated according to the spectral reflectance data (105), and a second CIEL*a*b* value is calculated according to the initial bidirectional texture function (103), and a correction vector in the a* and b* coordinates is calculated by subtracting the second CIE a*b* value from the first CIE a*b* value, and the correction vector is interpolated and extrapolated component by component for the full viewing and illumination angle ranges stored in the color table, the interpolated correction vector is applied to the CIEL*a*b* values of the initial bidirectional texture function (103) for each spectral measurement geometry stored in the color table, and the corrected bidirectional texture function CIEL*a*b* values are transformed into normalized and finally stored linear sRGB coordinates in the color table.

5. The method according to claim 4, wherein, a multi-level B-spline interpolation algorithm is used for the component-by-component interpolation and extrapolation of the correction vector.

6. The method according to any one of claims 1 to 5, wherein, in order to optimize the parameters of the intensity function, a cost function is defined based on the sum of color differences across all spectral reflectance measurement geometries.

7. The method according to claim 6, wherein, the cost function is supplemented by a penalty function designed to consider specific constraints, such constraints including keeping the parameter values within a valid range.

8. The method according to claim 6, wherein, The initial bidirectional texture function (103) is evaluated at the different spectral reflection measurement geometries, and the resulting CIEL*a*b* values are compared with the CIEL*a*b* values from the spectral reflection measurements using a weighted color difference formula, and the parameters of the intensity function are optimized using a non-linear optimization method to minimize the cost function.

9. The method according to any one of claims 1 to 5 and 7 to 8, wherein, the first and the second optimization steps are repeated or run iteratively to further improve the accuracy of the optimized bidirectional texture function (107), wherein the number of iterations is predefined.

10. A system for generating a bidirectional texture function of an object, the system comprising: - a camera-based measurement device configured to measure an initial bidirectional texture function (103) for the object, - a spectrophotometer configured to obtain spectral reflection data (105) for the object for a pre-given number of different measurement geometries, - a computing device communicatively connected to the camera-based measurement device and the spectrophotometer respectively, and configured to receive the initial bidirectional texture function and the obtained spectral reflection data (105) for the object via the respective communication connections, and to adapt the initial bidirectional texture function (103) to the obtained reflection data (105) to obtain an optimized bidirectional texture function (107), wherein the system is configured to perform the method according to any one of claims 1 to 9.

11. A computer system, comprising: - a computer unit; - a computer-readable program having program code which, when the program is executed on the computer unit, causes the following to be performed: - obtaining and receiving an initial bidirectional texture function (103) for an object and spectral reflection data (105) for the object, wherein the initial bidirectional texture function (103) is measured by a camera-based measurement device, and the spectral reflection data (105) is obtained by a spectrophotometer for a pre-given number of different measurement geometries; - matching the spectral reflection data (105) with the initial bidirectional texture function (103) by correspondingly adapting the parameters of the initial bidirectional texture function (103) to obtain an optimized bidirectional texture function (107), wherein the initial bidirectional texture function is represented by: - x: the surface coordinates of the object, - The irradiation and observation directions at the undercoat of the object - Color table depending on the illumination and viewing directions - a: the albedo or diffuse reflectance, - The k-th Cook-Torrance lobe; the Cook-Torrance lobe is a common bidirectional reflectance distribution function that describes the specular reflectance of a microfacet surface. -S k : For the weight of the k-th Cook-Torrance lobe, -α k : Parameters for the Beckmann distribution for the k-th Cook-Torrance lobe -F 0,k : For the Fresnel reflectivity of the k-th Cook-Torrance lobe - Spatial texture image table depending on irradiation and viewing directions The first term of the initial bidirectional texture function is divided into a first sub-term corresponding to a color table and a second sub-term corresponding to an intensity function and the parameters of the initial bidirectional texture function (103) are optimized to minimize the color difference between the spectral reflection data (105) and the initial bidirectional texture function (103) by optimizing the parameters of the color table in a first optimization step while the parameters of the intensity function remain constant and by optimizing the parameters of the intensity function in a second optimization step while the parameters of the color table remain constant.

Images