WO2023072837A1 - Procédé assisté par ordinateur et dispositif de détermination des propriétés d'au moins un objet dans une scène de prise de vue - Google Patents

Procédé assisté par ordinateur et dispositif de détermination des propriétés d'au moins un objet dans une scène de prise de vue Download PDF

Info

Publication number
WO2023072837A1
WO2023072837A1 PCT/EP2022/079604 EP2022079604W WO2023072837A1 WO 2023072837 A1 WO2023072837 A1 WO 2023072837A1 EP 2022079604 W EP2022079604 W EP 2022079604W WO 2023072837 A1 WO2023072837 A1 WO 2023072837A1
Authority
WO
WIPO (PCT)
Prior art keywords
light source
image
properties
virtual
recording scene
Prior art date
Application number
PCT/EP2022/079604
Other languages
German (de)
English (en)
Inventor
Max PARGMANN
Daniel Maldonado Quinto
Stefan Kesselheim
Jan Ebert
Original Assignee
Deutsches Zentrum für Luft- und Raumfahrt e.V.
Forschungszentrum Jülich GmbH
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from DE102021127682.9A external-priority patent/DE102021127682A1/de
Priority claimed from DE102022105771.2A external-priority patent/DE102022105771A1/de
Application filed by Deutsches Zentrum für Luft- und Raumfahrt e.V., Forschungszentrum Jülich GmbH filed Critical Deutsches Zentrum für Luft- und Raumfahrt e.V.
Priority to EP22809691.3A priority Critical patent/EP4423454A1/fr
Publication of WO2023072837A1 publication Critical patent/WO2023072837A1/fr

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
    • G01B21/02—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring length, width, or thickness
    • G01B21/04—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring length, width, or thickness by measuring coordinates of points
    • G01B21/042—Calibration or calibration artifacts
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24S—SOLAR HEAT COLLECTORS; SOLAR HEAT SYSTEMS
    • F24S20/00—Solar heat collectors specially adapted for particular uses or environments
    • F24S20/20—Solar heat collectors for receiving concentrated solar energy, e.g. receivers for solar power plants
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24S—SOLAR HEAT COLLECTORS; SOLAR HEAT SYSTEMS
    • F24S50/00—Arrangements for controlling solar heat collectors
    • F24S50/20—Arrangements for controlling solar heat collectors for tracking
    • F24S2050/25—Calibration means; Methods for initial positioning of solar concentrators or solar receivers
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24S—SOLAR HEAT COLLECTORS; SOLAR HEAT SYSTEMS
    • F24S23/00—Arrangements for concentrating solar-rays for solar heat collectors
    • F24S23/70—Arrangements for concentrating solar-rays for solar heat collectors with reflectors

Definitions

  • the invention relates to a computer-assisted method for determining object properties of an object to be measured that is illuminated by a light source in a recording scene, in particular for detecting and/or measuring the object properties, the object properties comprising at least one of the following: a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the object.
  • the invention also relates to a device for determining object properties of an object, in particular for detecting and/or measuring the object properties of the object, as well as a computer program product and a device for data processing.
  • Strip pattern deflectometry is a long-established technique for measuring the surface of objects in a wide variety of laboratory and industrial applications. For example, it is commonly used to measure mirrors and optical lenses. Parallel strips are projected onto the object to be examined and their reflection recorded. From the distortions of the reflected image, conclusions can be drawn about its surface via geometric relationships. This method is also used in many industrial processes, such as in car and aircraft production. It can also be used on solar tower power plants.
  • Solar tower power plants focus incident sunlight on a so-called receiver in order to generate electricity from the heat generated.
  • the sun's rays are deflected onto the receiver by motorized mirrors known as heliostats.
  • heliostats motorized mirrors known as heliostats.
  • Each of these heliostats has its own individual errors, such as mirror errors, which affect its focal point and alignment and thus the heat distribution on the receiver.
  • deflectometric measurements are carried out.
  • Deflectometry can only be used to a limited extent in some applications. For example, with heavily concave surfaces, these can cause lines to overlap and make the measurement unusable. Errors can also occur if the position information is imprecise. If, for example, there are multiple reflections in transparent materials, then several reflections of the stripe pattern overlap, which cannot be clearly resolved with the classic method.
  • the heliostats In order to keep the alignment error, also called tracking error, as low as possible, the heliostats must be calibrated regularly. This is usually done in such a way that the focal point of a single heliostat is moved by the receiver onto a white projection surface. The position of the focal spot on the projection surface, the associated position of the sun and the motor position are recorded and saved. Using these measuring points, the deviation of the real heliostat from a theoretically ideal heliostat can be determined using a mathematical-physical regression. Strip deflectometry has been used for a long time, especially on solar towers. The method could not be fully automated despite a need for accurate heliostat measurement.
  • the object of the invention is to specify an efficient method for determining object properties of at least one object in a recording scene.
  • a further object is to provide a device for determining object properties of at least one object in a recording scene using such an efficient method.
  • a further object is to provide a computer program product comprising instructions for executing such a method.
  • a further object is to provide a data processing device for carrying out such a method.
  • a computer-aided method for determining object properties of an object to be measured that is illuminated by a light source in a recording scene in particular for detecting and/or measuring the object properties, the object properties comprising at least one of the following: a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the object, image data of a reflection of rays of the light source on the object to be measured being recorded in the recording scene and for determining at least one of the object properties to be analyzed.
  • the deflectometry for measuring an object surface is based on geometric optics, the fringe pattern of a projector must be used for the measurement. In principle, however, all information about the mirror surface, its position and/or reflectivity and the light source is available in the form of a mathematical convolution in every reflection considered. This means that any defined light source can be used.
  • Object properties of the object to be measured such as a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the object can in principle be reconstructed with any light source, provided these are sufficiently well characterized .
  • the heliostat calibration for example, is a simple, cheap and accurate method to determine the misalignment of the heliostat. So far, however, your stored data cannot be used meaningfully for other purposes.
  • the stored data of the heliostat calibration can be used sensibly.
  • the recorded image data of the recording scene can be compared with simulated image data of the recording scene, with at least one of the object properties and/or at least one parameter of the recording scene and/or at least one parameter of the Light source can be varied.
  • the method according to the invention uses what is known as an inverse render used.
  • a three-dimensional reconstruction of the object is generated from the two-dimensional image of the reflection.
  • a classic ray tracing method also known as ray tracing method
  • a simulation of the optical conditions of a scene can be generated in the so-called forward step.
  • An ideal representation of the examined object here for example a heliostat, is assumed and a 2D image is generated from the 3D scene based on the radiation characteristics of the light source while tracking and reflecting the light rays.
  • the image generated in this way is then compared with the image measured in reality and evaluated using an evaluation function.
  • the special feature of inverse ray tracing is that the evaluation function can be used to derive free parameters from the simulated geometry.
  • object properties of the object to be measured such as a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the Object and/or position of the light source and/or properties of the light source, mathematically differentiated in order to determine new values for the simulation of the recording scene in a backward step. This is a common practice in optimization algorithms.
  • the object to be measured is not directly observed by the camera, but only the reflection of the rays of the selected, arbitrary light source.
  • An evaluation function used as a convergence criterion can favorably indicate, for example, a difference in the brightness values of the individual pixels of the measured image and of the virtual image. Starting from this point and depending on the extent of the deviation, parameters of the virtual recording scene can then be adapted backwards, for example using a mathematical gradient method.
  • an iterative optimization process is carried out that generates a new image in each step and compares it with the measured image until the evaluation function has reached an optimum as a convergence criterion.
  • object properties of the object to be measured such as a surface, a shape of the object, one or more surface properties of the object, a position of the object, an alignment of the object, at least one material property of the object can be optimized using predefined start values in the iterative method. This allows inaccurate measurements to be compensated.
  • the mapping function for example, of the three-dimensional object surface after two-dimensional reflection is not a clear mapping, so there is no bijectivity.
  • the problem is mathematically underdetermined.
  • the position and the angle of the light source can be changed, for example by an LED array, or changing sun positions.
  • the additional images can be evaluated at the same time in the evaluation function by summing up the available images. In this way, ambiguities can be resolved or reduced.
  • the light rays can be traced back using the inverse ray tracing method through a differentiable representation of the virtual recording scene.
  • object properties of the object to be measured such as a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the object, a position of the light source, properties of the light source, mathematically differentiated in order to determine new values for the simulation of the shooting scene.
  • parameters of the virtual recording scene can then be adapted backwards, for example using a mathematical gradient method.
  • neural networks or simple logic circuits can be used instead of differentiability.
  • parameters of the recording scene can include at least one of: position of at least one camera, distance of the object from a carrier of a projection surface, position of an image on the projection surface, position of the projection surface on the carrier.
  • parameters of the light source can at least include: a type of the at least one light source, a position of the at least one light source.
  • the proposed method is particularly advantageous when the environmental conditions and/or production conditions when generating the images, for example in the case of inaccurate positioning of the at least one light source, inaccurate position of the object and the like, deviate from ideal measurement conditions, for example in industrial systems.
  • the proposed method eliminates many of the limitations of conventional strip deflectometry when measuring an object surface. Instead of a projector, any defined light source is sufficient. This makes it possible to greatly reduce the complexity of a measurement setup.
  • the position and/or the orientation and/or the material property, for example the reflectivity of the examined object, does not have to be clearly defined either and can also be optimized.
  • components with a strong concave shape can be advantageously measured with the method.
  • the method can at least include: illuminating the object with the at least one light source in at least one light source position; capturing at least the image of the object with the at least one camera with the at least one light source in the at least one light source position; inputting a virtual recording scene at least with object properties of the object, parameters of the at least one light source and parameters of the virtual recording scene; calculating at least one virtual image of the object with the virtual recording scene; determining deviations of the virtual image from the measured image; if the deviations violate at least one predetermined convergence criterion, changing parameters of the recording scene and/or parameters of the at least one light source and/or object properties of the object and calculating a new virtual image; and iterating the calculation of the virtual image and the determination of the deviations of the virtual image from the measured image until the deviations meet the at least one convergence criterion.
  • the method can also be operated purely by software.
  • an optimal reflection beam distribution can be specified and then the necessary parameters, such as position, alignment, shape or material properties of the object can be optimized accordingly.
  • light beams that are emitted by the at least one light source to the object and light beams that are reflected on at least the object can be tracked on a path into the at least one camera.
  • an inverse ray tracing method can be used in the method for calculating the virtual image of the object.
  • an ideal representation of the examined object can advantageously be assumed and a 2D image can be generated from the 3D scene based on the radiation characteristics of the light source while tracking and reflecting the light rays. The image generated in this way is then compared with the image measured in reality and evaluated using an evaluation function.
  • free parameters of the virtual recording scene can be derived based on properties of the virtual image.
  • the light rays can be traced back using the inverse ray tracing method through a differentiable representation of the virtual recording scene.
  • object properties of the object to be measured such as a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the object, position of the object, position of the light source, properties of the light source, mathematically differentiated in order to determine new values for the simulation of the shooting scene.
  • parameters of the virtual recording scene can then be adapted backwards, for example using a mathematical gradient method.
  • multiple images of the object can be recorded with different positions and angles of the at least one light source and compared with corresponding virtual images.
  • the sun can be used as the light source and multiple images of the object can be recorded at different positions of the sun as positions of the light source.
  • the problem of underdetermination of the mapping from the three-dimensional geometry of the object to the two-dimensional image can advantageously be solved by taking pictures with changed positions of the light sources.
  • multiple light sources can be used to illuminate the object when recording the at least one image of the object.
  • the problem of underdetermination of the mapping from the three-dimensional geometry of the object to the two-dimensional image can be solved favorably by recording with multiple light sources.
  • an evaluation function can be used as a convergence criterion.
  • This evaluation function which can also be in the form of a loss function, can be used to evaluate convergence of an optimization method when adapting the simulated images to the measured images.
  • images can be recorded with multiple cameras.
  • the problem of underdetermination of the mapping from the three-dimensional geometry of the object to the two-dimensional image can also be solved favorably with the aid of recordings from a number of cameras.
  • an image of the object projected onto a projection surface can be recorded with the at least one camera as at least one image of the object.
  • images of the sun as a light source such as those recorded when calibrating heliostats in solar tower power plants, can be used advantageously for measuring the surface of the heliostats.
  • the projection surface can have a matt white surface.
  • An absolutely matt white surface, a so-called Lambertian surface can advantageously be used for indirect imaging of the object surface, for example via illumination with sunbeams.
  • other suitable surfaces are also conceivable.
  • a device for determining object properties of an object, in particular for detecting and/or measuring the object properties of the object, using such a method, comprising at least one light source, at least one camera and at least one computer, the at least a light source is designed to illuminate the object in at least one light source position, and wherein the at least one camera is designed to record at least one image of the object.
  • the at least one computer is designed to analyze image data recorded in a recording scene of a reflection of rays from the light source on the object to be measured and to determine at least one of the object properties.
  • the proposed device enables the determination of object properties of the object to be measured such as a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the object by recording reflections of the object at least one camera and a simulation of the recording scene with so-called inverse rendering.
  • object properties of the object such as a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the object by recording reflections of the object at least one camera and a simulation of the recording scene with so-called inverse rendering.
  • a three-dimensional reconstruction of the object is thus generated from the two-dimensional image of the reflection.
  • inverse rendering for example, a simulation of the optical conditions of a scene is generated in a so-called forward step using a classic ray tracing method, also known as ray tracing.
  • An ideal representation of the examined object is assumed and a 2D image is generated from the 3D scene based on the radiation characteristics of the light source while tracking and reflecting the light rays.
  • the image generated in this way is then compared with the image measured in reality and evaluated using an evaluation function.
  • the computer can be designed to set up a virtual recording scene at least with object properties of the object, parameters of the at least one light source, and parameters of the virtual recording scene; to calculate at least one virtual image of the object with the virtual recording scene; determine deviations of the virtual image from the measured image; if the deviations violate at least one predetermined convergence criterion, to change object properties of the object and to determine a new virtual image; iterating the calculation of the virtual image and determining the deviations of the virtual image from the measured image until the deviations meet the at least one convergence criterion.
  • the light rays can be traced back using the inverse ray tracing method through a differentiable representation of the virtual recording scene.
  • This allows free parameters of the simulated geometry to be derived.
  • object properties of the object such as a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the object, position of the light source, properties of the light source, mathematically differentiated in order to determine new values for the simulation of the recording scene in a backward step. This is a common practice in optimization algorithms.
  • An evaluation function used as a convergence criterion can favorably indicate, for example, a difference in the brightness values of the individual pixels of the measured image and of the virtual image. Starting from this point and depending on the extent of the deviation, parameters of the virtual recording scene can then be adapted backwards, for example using a mathematical gradient method.
  • An iterative optimization process is carried out here, which generates a new image in each step and compares it with the measured image until the evaluation function has reached an optimum as a convergence criterion.
  • the exact position and orientation of the object to be measured can also be optimized with the iterative method using predefined starting values. This allows inaccurate measurements to be compensated.
  • the computer can be designed to track the virtual image of the object by tracking light beams emitted by the at least one light source and light beams reflected by at least the object on a path into the at least one camera to determine.
  • the computer can be designed to determine the virtual image of the object using an inverse ray tracing method.
  • an ideal representation of the examined object can advantageously be assumed and a 2D image can be generated from the 3D scene based on the radiation characteristics of the light source while tracking and reflecting the light rays.
  • the image generated in this way is then compared with the image measured in reality and evaluated using an evaluation function.
  • object properties of the object can include at least one of the following: a surface, a shape of the object, one or more surface properties of the object, a position of the object, an alignment of the object, at least one material property of the object.
  • parameters of the recording scene can include at least one of: position of at least one camera, distance of the object from a carrier of a projection surface, position of an image on the projection surface, position of the projection surface on the carrier. A virtual scene for simulating the measurement can be set up using these parameters.
  • the computer can be designed to derive free parameters of the virtual recording scene according to the evaluation function when changing object properties of the object in order to determine a new virtual image.
  • the light rays can be traced back using the inverse ray tracing method through a differentiable representation of the virtual recording scene.
  • object properties of the object to be measured such as a surface, a shape of the object, one or more surface properties of the object, a position of the object, an orientation of the object, at least one material property of the object, position of the light source, properties of the light source, mathematically differentiated in order to determine new values for the simulation of the shooting scene.
  • parameters of the virtual recording scene can then be adapted backwards, for example using a mathematical gradient method.
  • the at least one light source can be designed to illuminate the object from different positions and angles in order to record a number of images.
  • the sun can represent the light source and the at least one camera can be designed to record multiple images of the object at different positions of the sun as positions of the light source.
  • a plurality of light sources can advantageously be provided for illuminating the object when the at least one image of the object is recorded.
  • the problem of underdetermination of the mapping from the three-dimensional geometry of the object to the two-dimensional image can be solved favorably by taking pictures with a plurality of light sources.
  • the convergence criterion can include an evaluation function. This evaluation function, which can also be in the form of a loss function, can be used to evaluate convergence of an optimization method when adapting the simulated images to the measured images.
  • the projection surface can be provided to project an image of the object onto it for recording with the at least one camera.
  • images of the sun as a light source such as those recorded when calibrating heliostats in solar tower power plants, can be used advantageously for measuring the surface of the heliostats.
  • the projection surface can have a matt white surface.
  • An absolutely matt white surface a so-called Lambertian surface, can advantageously be used for an indirect imaging of the object surface, for example via illumination with sunbeams.
  • other suitable surfaces are also conceivable.
  • the method according to the invention can be carried out by means of a correspondingly set up data processing device, for example the device according to the invention for data processing, in particular automatically or semi-automatically.
  • a correspondingly set up data processing device for example the device according to the invention for data processing, in particular automatically or semi-automatically.
  • a conventional computer, a conventional workstation, a mainframe or the like can be used as such a data processing device. Due to the possibility of parallelizing the method, the use of a graphics card can be particularly advantageous.
  • the data processing device can therefore include in particular a processor, a microchip, an integrated circuit, a hardware circuit or the like for executing a computer program or program code that encodes or represents the method. Furthermore, the data processing device can in particular have a volatile and/or non-volatile data memory connected thereto, one or more interfaces for receiving and for outputting data and/or the like. The data processing device can have one or more sub-devices for carrying out the individual method steps of the method according to the invention. These partial devices can be or include corresponding hardware circuits or hardware modules, but also corresponding program parts or program modules of the named computer program, ie for example an operating program for the data processing device.
  • a computer program product which comprises commands or control instructions which, when the program is executed by a computer, in particular a device for data processing or the data processing device mentioned in connection with the method according to the invention, cause this computer to Variant of the method according to the invention, in particular automatically or semi-automatically to perform.
  • a computer program is also proposed which comprises commands or control instructions which, when the program is executed by a computer, in particular a device for data processing or the data processing device mentioned in connection with the method according to the invention, cause this computer to carry out at least one variant of the method according to the invention, in particular automatically or semi-automatically.
  • the method according to the invention can be fully or partially computer-implemented or computer-implementable, ie encoded or represented by such a computer program or a corresponding program code.
  • the computer program product according to the invention can be a computer-readable data carrier on which a corresponding computer program is stored.
  • a device for data processing has means for carrying out at least one variant of the method according to the invention, in particular in an automated or partially automated manner. These means can in particular be the means described in connection with the named data processing device.
  • the device according to the invention can thus in particular have a corresponding processor, a data memory and at least one input and/or output interface.
  • the device according to the invention can include the computer program product according to the invention.
  • the device according to the invention can be set up in particular to carry out the method according to the invention and the method according to the invention can therefore be carried out using the device according to the invention, the device according to the invention can accordingly have some or all of the properties and/or features described in connection with the method according to the invention.
  • FIG. 1 shows a device for determining object properties of an object according to an exemplary embodiment of the invention, a camera recording light reflected directly on the object and emitted by a plurality of light sources;
  • FIG. 2 shows a device for determining object properties of an object according to a further exemplary embodiment of the invention, the camera recording light reflected onto a projection surface, which light is emitted by a plurality of light sources;
  • 3 shows a device for determining object properties of an object according to a further exemplary embodiment of the invention, the object representing a heliostat of a solar tower power plant, the sun being used in a number of positions as a light source;
  • FIG. 4 shows a flow chart of a computer-aided method for determining object properties of an object according to an embodiment of the invention.
  • Fig. 5 is a schematic representation of a
  • Data processing device for executing a computer-aided method for determining object properties of an object according to an embodiment of the invention.
  • FIG. 1 shows a device 100 for determining object properties of an object 20, in particular for detecting and/or measuring the object properties of the object 20, according to an exemplary embodiment of the invention.
  • a camera 30 records light 28 that is reflected directly on the object 20 .
  • the device 100 comprises four light sources 10 spaced apart from one another, which emit light beams 18 onto the object 20 from light source positions correspondingly spaced apart from one another.
  • the device 100 further comprises a camera 30 and at least one computer 50.
  • the camera 30 records at least one image 40 of the object 20.
  • the light sources 10 can be arranged next to one another in a linear arrangement, for example.
  • the problem of underdetermination of the mapping from the three-dimensional geometry of the object 20 to the two-dimensional image 40 can advantageously be solved by taking pictures with changed positions of light sources 10 or several light sources 10 simultaneously
  • the computer 50 is designed to analyze image data recorded in a recording scene of a reflection of beams from the light source 10 on the object 20 to be measured and to determine at least one of the object properties.
  • the computer is advantageously designed to set up a virtual recording scene at least with object properties of object 20, parameters of the at least one light source 10, and parameters of a virtual recording scene, to calculate at least one virtual image 44_n of object 20 with the virtual recording scene, and to calculate deviations from the virtual To determine the image 44_n from the measured image 40.
  • the index n denotes an nth iteration loop, where n is an integer 1, 2, 3, ....
  • a virtual scene for simulating the measurement can be set up using these parameters.
  • Object properties of the object 20 can include at least one of the following: a surface 22, a shape of the object 20, one or more surface properties of the object 20, a position of the object 20, an orientation of the object 20, at least one material property of the object 20.
  • Parameters of the at least one light source 10 may include at least a type of the at least one light source 10 and a position of the at least one light source 10 .
  • Parameters of the recording scene can include at least one of: position of at least one camera 30, distance of object 20 from a support 60 of a projection surface 32, position of an image 40 on projection surface 32, position of projection surface 32 on support 60.
  • the computer 50 is further designed to change object properties of the object 20 and to determine a new virtual image 44_n+1 if the deviations violate at least one predetermined convergence criterion.
  • the index n+1 denotes an n+1th iteration loop, where n is an integer 1, 2, 3, ....
  • the computer 50 is further designed to iterate the calculation of the virtual image 44_n, 44_n+1 and the determination of the deviations of the virtual image 44_n, 44_n+1 from the measured image 40 until the deviations meet the at least one convergence criterion.
  • the convergence criterion can include an evaluation function, for example. This evaluation function, which can also be in the form of a loss function, can be used to evaluate convergence of an optimization method when adapting the simulated images 44 to the measured images 40
  • a difference in the brightness values of the individual pixels of the measured image 40 and the virtual image 44 can be used as an evaluation function.
  • object properties of the object to be measured 20 such as a surface 22, a shape of the object 20, one or more surface properties of the object 20, a position of the object 20, an orientation of the object 20, at least one material property of the object 20, a position of the light source 10, properties of the light source 10, are mathematically differentiated in order to determine new values for simulating the recording scene.
  • parameters of the virtual recording scene can then be adapted backwards, for example via a mathematical gradient method.
  • the computer 50 is designed to display the virtual image 44 of the object 20 by tracking light beams 18, which are emitted by the at least one light source 10 and light beams 28, which are reflected by at least the object 20, on a path into the camera 30 to determine.
  • the virtual image 44 of the object 20 can thus advantageously be determined using an inverse beam tracing method, also known as ray tracing.
  • an ideal representation of the examined object 20 can advantageously be assumed and a 2D image 44 can be generated from the 3D scene based on the emission characteristics of the light source 10 while tracking and reflecting the light beams 18 , 28 .
  • the image 44 generated in this way is then compared with the image 40 measured in reality and evaluated using an evaluation function.
  • object properties of the object 20 include at least one of the following: a surface 22, a shape of the object 20, one or more surface properties of the object 20, a position of the object 20, an orientation of the object 20, at least one material property of the Object 20.
  • Parameters of the at least one light source 10 include at least one type of the at least one light source 10 and the position of the at least one light source 10.
  • Properties of the virtual recording scene include at least one position of a projection surface 32 and/or the camera 30.
  • the computer 50 is further designed to derive free parameters of the virtual recording scene based on properties of the virtual image when changing object properties of the object 20 to determine a new virtual image 44 .
  • the light source 10 can illuminate the object 20 from different positions and angles for taking multiple images 40 .
  • the sun can be used as light source 10 and multiple images 40 of object 20 can be recorded with camera 30 at different positions of the sun as positions 12 , 14 , 16 of light source 10 .
  • a plurality of cameras 30 for recording images 40 of the object 20 can also advantageously be provided.
  • the problem of underdetermination of the mapping from the three-dimensional geometry of the object 20 to the two-dimensional image 40 can advantageously be solved by recordings with changed positions of the light sources 10 or multiple light sources 10 simultaneously or recordings with multiple cameras 30 .
  • the device 100 shown in FIG. 1 for determining object properties of an object 20 can be used advantageously in industrial production, for example.
  • the conditions are not always ideal for deflectometry to be useful for assessing reflective surfaces.
  • any directed light source 10 such as an LED array or a flashlight, can be directed onto the location to be examined (object 20). This represents a cost-effective and easy-to-handle option for measuring the surface.
  • the inverse rendering method also requires only rough information about where the object 20 is located exactly. If the object 20 is not quite at the defined position, the position can also be determined in the optimization process.
  • a measurement of the surfaces of both components can advantageously be carried out simultaneously, for example at the interface between the fuselage of an aircraft and its wings, at which there is a concave transition. This represents a significant advantage over the previous deflectometric method, where lines can overlap in the worst case and thus make the measurement unusable.
  • FIG. 2 shows a device 100 for measuring at least one surface 22 of an object 20 according to a further exemplary embodiment of the invention.
  • the camera 30 records light reflected on a projection surface 32 .
  • an image 40 of the object 20 is projected onto the projection surface 32 for recording with the camera 30 .
  • the projection surface 32 can advantageously have a matt white surface.
  • the object 20 is illuminated with light beams 18 from a plurality of light sources 10 .
  • the rays 28 reflected on the object 20 fall on the projection surface 32.
  • the image 42 on the projection surface 32 is recorded by the camera 30 in the camera image field 29.
  • the iteration process in the computer 50 for measuring the surface 22 of the object 20 runs as in the previous exemplary embodiment.
  • the virtual recording scene only includes additional parameters of the projection surface 32 for calculating the virtual image 44.
  • FIG. 3 shows a device 100 for measuring at least one surface 22 of an object 20 according to a further exemplary embodiment of the invention.
  • the object 20 represents a mirror 26 of a heliostat 24 of a solar tower power plant with a carrier 60 of a projection surface, for example a solar tower 64.
  • an image of the sun is generated as light source 10 on a projection surface 32 .
  • the focal spot images 42 used here are suitable for use in surface measurement.
  • a camera 30 takes pictures 40 at several positions of the sun, ie light source positions 12, 14, 16.
  • the already described evaluation of the recorded images 40 takes place in the computer 50 by inverse ray tracing.
  • the recordings generated during operation can be used to set up an inverse rendering of the images 44 and to derive the object properties of the heliostat 24 from this. For example, a regular surface measurement is possible in this way.
  • the dashed arrows are intended to identify different parameters to be varied in the simulation of the recording scene in order to determine the object properties of the object 20 .
  • the proposed method has significant advantages over the previously used deflectometry. Measurements of this type can be carried out during the day, so that the sun can be used as the light source 10 .
  • the heliostats 24 can thus be calibrated fully automatically using the computer 50 .
  • the process can be used in existing and future commercial solar tower power plants. No additional hardware is required for existing power plants, so retrofitting of existing power plants is possible.
  • FIG. 4 shows a flow chart of a computer-assisted method for determining object properties of an object 20 according to an exemplary embodiment of the invention.
  • the object 20 is illuminated with at least one light source 10, which is located in at least one light source position.
  • At least one image 40 of the object 20 is recorded with a camera 30 with the at least one light source 10 in its light source position.
  • Image data of a reflection of beams of the light source 10 on the object 20 to be measured are recorded in a recording scene. This image data should be analyzed to determine at least one of the object properties.
  • the recorded image data of the recording scene are compared with simulated image data of the recording scene. At least one of the object properties and/or at least one parameter of the recording scene and/or at least one parameter of the light source 10 is varied in the simulated image data.
  • step S120 of the flow chart shown in FIG. 4 object properties of the object 20, parameters of the at least one light source 10 and parameters of the virtual recording scene are defined.
  • Object properties of object 20 can include at least one of the following: a surface 22, a shape of object 20, one or more surface properties of object 20, a position of object 20, an orientation of object 20, at least one material property of object 20.
  • Parameters of the recording scene can include at least the position of at least one camera 30, the distance of the object 20 from a carrier 60 of a projection surface 32, the position of an image 40 on the projection surface 32, and the position of the projection surface 32 on the carrier 60.
  • Parameters of the light source 10 can include at least one type of the at least one light source 10, a position 12, 14, 16 of the at least one light source 10.
  • the parameters defined in step S120 are added to a virtual recording scene in step S100.
  • step S102 at least one virtual image 44, 44_n of the object 20 is calculated with the virtual recording scene.
  • Calculating the virtual image 44 of the object 20 includes tracking light rays 18, which are emitted from the light source to the object 20 and light rays 28, which are reflected at least on the object 20, on a path into the camera 30. In particular, this includes Calculating the virtual image 44 of the object 20 an inverse ray tracing method.
  • the image 40 measured in the same light source position is loaded into the computer 50 in step S104.
  • an image 44 that has already been pre-simulated can also be loaded into the computer 50 .
  • a plurality of images 40 of the object 20 with different positions and angles of the at least one light source 10 can also be recorded and compared with corresponding virtual images 44 .
  • the sun can be used as the light source 10 and multiple images 40 of the object 20 can be recorded at different positions of the sun as positions 12 , 14 , 16 of the light source 10 .
  • a plurality of light sources 10 can also be used to illuminate the object 20 when the at least one image 40 of the object 20 is recorded.
  • multiple images 40 can be recorded with multiple cameras 30 .
  • an image 42 of the object 20 projected onto a projection surface 32 can be recorded with the camera 30 as at least one image 40 of the object 20 .
  • the image 42 corresponds to the focal spot 62.
  • the projection surface 32 can advantageously have a matt white surface.
  • Deviations of the virtual image 44, 44_n from the measured image 40 are determined when comparing the virtual image 44, 44_n with the measured image 40 and/or with the pre-simulated image 44 in step S106. If the deviations violate at least one predetermined convergence criterion, parameters of the recording scene and/or parameters of the at least one light source 10 and/or object properties of objects 20 are changed and a new virtual image 44, 44_n+1 is calculated.
  • An evaluation function can advantageously be used as a convergence criterion.
  • step S108 it is checked whether the specified convergence criterion is met. If this is the case, the object properties of the object 20 are saved in step S110.
  • step S112 object properties such as parameters of the surface 22 of the object 20 are changed, for example using a gradient method known per se, and the calculation of the virtual image 44, 44_n+1 is repeated in step S102.
  • object properties of the object 20 to determine a new virtual image 44, 44_n+1 properties of the virtual image 44 can be derived according to free parameters of the virtual recording scene.
  • object properties of the object 20 to be measured such as a surface 22, a shape of the object 20, one or more surface properties of the object 20, a position of the object 20, an orientation of the Object 20, at least one material property of object 20, position of light source 10, properties of light source 10, mathematically differentiated in order to determine new values for simulating the recording scene. This is a common practice in optimization algorithms.
  • the evaluation function used as a convergence criterion can favorably indicate, for example, a difference in the brightness values of the individual pixels of the measured image 40 and of the virtual image 44 .
  • parameters of the virtual recording scene can then be adapted backwards, for example using a mathematical gradient method.
  • the calculation of the virtual image 44, 44_n and the determination of the deviations of the virtual image 44, 44_n, 44_n+1 from the measured image 40 is iterated until the deviations meet the at least one convergence criterion.
  • ReflectingObject ⁇ Load ideally shaped discretized object
  • Bitmap ⁇ Create array with same height, width, resolution as CameraPlane
  • Scheduler ⁇ ChooseScheduler; for Epochs do for (Image, LightPosition) in (Images, LightPositions) do for (NormalVector, DiscretePoint) on ReflectingObject do Rays ⁇ — GenerateRays(NumRays, LightPosition,
  • Bitmap ⁇ UpdateBitmap(intersection); end end end end
  • the pseudo-code describes a function for optimizing object properties of an object (OptimizeObject), which takes as input parameters a number of image scenes (Epochs), images (Images), light source positions (LightPositions), light distribution (LightDistribution), a number of light rays ( NumRays) and a camera plane (CameraPlane).
  • OptimizeObject takes as input parameters a number of image scenes (Epochs), images (Images), light source positions (LightPositions), light distribution (LightDistribution), a number of light rays ( NumRays) and a camera plane (CameraPlane).
  • An ideally formed discretized object is loaded (ReflectingObject).
  • parameters of the recording scene and/or the light source (environment) are loaded.
  • a matrix with the same height, width, resolution as the camera plane (bitmap) is created.
  • An optimizer function is selected, a scheduler is selected.
  • the calculations are iterated over the number of image scenes, images and light source positions, light beam parameters. In each case, an evaluation function (loss) is calculated as a convergence criterion. After a gradient method (gradient), new changed object properties of the object are determined. The object properties are updated iteratively (UpdateObjectProperties).
  • Figure 6 shows a schematic representation of a data processing device 500, for example a computer 50, for executing a method for determining object properties of an object 20.
  • the data processing device 500 has a data memory 514, a processor 512 connected thereto and an interface 516 connected thereto receiving and outputting data. Furthermore, the data processing device 500 has an interface to external devices 518, such as a network or an external storage system.
  • a computer program 520 that implements the method is stored here on the data memory 514 .
  • Corresponding program modules 522 represent the method steps as shown in FIG.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Length Measuring Devices By Optical Means (AREA)

Abstract

L'invention concerne un procédé - assisté par ordinateur - de détermination des propriétés d'un objet (20) à mesurer et éclairé par une source lumineuse (10), en particulier de détection et/ou de mesure des propriétés de l'objet, les propriétés de l'objet comprenant au moins l'une des propriétés suivantes : une surface (22), une forme de l'objet (20), au moins une propriété de surface de l'objet (20), une position de l'objet (20), une orientation de l'objet (20), au moins une propriété de matériau de l'objet (20); dans la scène de prise de vue, des données d'image d'une réflexion de rayons de la source lumineuse (10) sont reçues sur l'objet (20) à mesurer et sont analysées pour déterminer au moins une des propriétés de l'objet. L'invention concerne en outre un dispositif (100) de détermination des propriétés d'un objet.
PCT/EP2022/079604 2021-10-25 2022-10-24 Procédé assisté par ordinateur et dispositif de détermination des propriétés d'au moins un objet dans une scène de prise de vue WO2023072837A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP22809691.3A EP4423454A1 (fr) 2021-10-25 2022-10-24 Procédé assisté par ordinateur et dispositif de détermination des propriétés d'au moins un objet dans une scène de prise de vue

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
DE102021127682.9 2021-10-25
DE102021127682.9A DE102021127682A1 (de) 2021-10-25 2021-10-25 Computergestütztes Verfahren und Vorrichtung zum Vermessen einer Oberfläche eines Objekts
DE102022105771.2 2022-03-11
DE102022105771.2A DE102022105771A1 (de) 2022-03-11 2022-03-11 Computergestütztes Verfahren und Vorrichtung zum Bestimmen von Objekteigenschaften wenigstens eines Objekts in einer Aufnahmeszene

Publications (1)

Publication Number Publication Date
WO2023072837A1 true WO2023072837A1 (fr) 2023-05-04

Family

ID=84361576

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2022/079604 WO2023072837A1 (fr) 2021-10-25 2022-10-24 Procédé assisté par ordinateur et dispositif de détermination des propriétés d'au moins un objet dans une scène de prise de vue

Country Status (2)

Country Link
EP (1) EP4423454A1 (fr)
WO (1) WO2023072837A1 (fr)

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
JUN XIAO ET AL: "A review of available methods for surface shape measurement of solar concentrator in solar thermal power applications", RENEWABLE AND SUSTAINABLE ENERGY REVIEWS, ELSEVIERS SCIENCE, NEW YORK, NY, US, vol. 16, no. 5, 29 January 2012 (2012-01-29), pages 2539 - 2544, XP028416062, ISSN: 1364-0321, [retrieved on 20120204], DOI: 10.1016/J.RSER.2012.01.063 *
SÁNCHEZ-GONZÁLEZ ALBERTO ET AL: "Reflections between heliostats: Model to detect alignment errors", SOLAR ENERGY, ELSEVIER, AMSTERDAM, NL, vol. 201, 13 March 2020 (2020-03-13), pages 373 - 386, XP086123587, ISSN: 0038-092X, [retrieved on 20200313], DOI: 10.1016/J.SOLENER.2020.03.005 *
SATTLER JOHANNES CHRISTOPH ET AL: "Review of heliostat calibration and tracking control methods", SOLAR ENERGY, ELSEVIER, AMSTERDAM, NL, vol. 207, 30 June 2020 (2020-06-30), pages 110 - 132, XP086263028, ISSN: 0038-092X, [retrieved on 20200630], DOI: 10.1016/J.SOLENER.2020.06.030 *
STEFFEN ULMER ET AL: "Automated high resolution measurement of heliostat slope errors", SOLAR ENERGY, ELSEVIER, AMSTERDAM, NL, vol. 85, no. 4, 6 January 2010 (2010-01-06), pages 681 - 687, XP028173221, ISSN: 0038-092X, [retrieved on 20100111], DOI: 10.1016/J.SOLENER.2010.01.010 *

Also Published As

Publication number Publication date
EP4423454A1 (fr) 2024-09-04

Similar Documents

Publication Publication Date Title
DE102015116047B4 (de) Prüfvorrichtung und Steuerverfahren für eine Prüfvorrichtung
EP3807838A2 (fr) Contrôle de matériau d'échantillons optiques
DE102012108567A1 (de) Verfahren und Vorrichtung zum Erlangen von Tiefeninformationen unter Verwendung eines Lichtmusters
EP2598836B1 (fr) Détermination de cartes de réflexion défocalisées pour la détermination stable de « shape from focus » dans des images microscopiques
DE102013204375A1 (de) Partikelbild-Geschwindigkeitsmesssystem für einen dreidimensionalen Raum
EP2353146A1 (fr) Procédé pour mesurer la croissance de disques de feuille de plantes ainsi qu'un dispositif approprié à cet effet
DE102013013791A1 (de) Verfahren und Vorrichtung zum berührungslosen Vermessen von Oberflächenkonturen
DE112018002357T5 (de) Dreidimensionales Messverfahren unter Verwendung von Merkmalsgrössen, und Vorrichtung, die das Verfahren verwendet
DE102010011528A1 (de) Anordnung und Verfahren zur Aufnahme geometrischer und photometrischer Objektdaten im Raum
DE102017010683B4 (de) Verfahren zur automatischen Wiederherstellung eines eingemessenen Zustands eines Projektionssystems
DE112018001600T5 (de) System und verfahren zur 3d-profilbestimmung unter verwendung einer modellbasierten peak-auswahl
DE102009060580A1 (de) Verfahren zur Bestimmung eines Satzes von optischen Abbildungsfunktionen für die 3D-Strömungsmessung
DE102010045634A1 (de) Anordnung und Verfahren zur Aufnahme geometrischer und photometrischer Objektdaten im Raum
DE102011082280A1 (de) Bildmessvorrichtung und Bildmessverfahren
WO2023072837A1 (fr) Procédé assisté par ordinateur et dispositif de détermination des propriétés d'au moins un objet dans une scène de prise de vue
EP2887010A1 (fr) Procédé et dispositif de mesure optique en trois dimensions d'objets avec un procédé de mesure topométrique ainsi que programme informatique correspondant
DE102021127682A1 (de) Computergestütztes Verfahren und Vorrichtung zum Vermessen einer Oberfläche eines Objekts
DE102022105771A1 (de) Computergestütztes Verfahren und Vorrichtung zum Bestimmen von Objekteigenschaften wenigstens eines Objekts in einer Aufnahmeszene
DE19749974C2 (de) Verfahren und Apparat zur Erzeugung einer 3D-Punktwolke
EP3049757B1 (fr) Mesure de châssis en présence de lumière ambiante
WO2014114663A1 (fr) Dispositif optique et procédé pour déterminer par triangulation de deux caméras linéaires les coordonnées spatiales de surfaces d'objets macroscopiques
WO2019149888A1 (fr) Dispositif et procédé pour la génération d'images de rendu réaliste
DE102020206768B4 (de) Bilderfassungsvorrichtung und Verfahren zum Bestimmen einer plenoptischen Funktion
EP4235153A1 (fr) Procédé mis en oeuvre par ordinateur, programme informatique, système de traitement des données et dispositif de détermination du comportement de réflectance d'une surface d'un objet et support d'enregistrement non volatil lisible par ordinateur pourvu d'instructions mémorisées sur celui-ci destinées à la détermination du comportement de réflectance d'une surface d'un objet
AT524118A4 (de) Computer-implementiertes Verfahren zur Erstellung von mehrdimensionalen Gegenstands-Datenstrukturen

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22809691

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2022809691

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2022809691

Country of ref document: EP

Effective date: 20240527