Method for detecting a scene comprising a phase object
By providing the camera with the inverse function of the imaging function and the principle of triangulation, the line-of-sight deviation value is calculated and a deviation value map is constructed. This solves the problem of phase object detection in the BOS method under significant depth structures and motion backgrounds, and achieves high-precision phase object detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- LASER VISION CO LTD
- Filing Date
- 2024-11-08
- Publication Date
- 2026-06-19
Smart Images
Figure CN122249834A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method for detecting a scene including a phase object arranged in front of a textured background. Within the scope of the method, a corresponding number of camera images are simultaneously captured by multiple cameras arranged spatially offset from each other and passing through the phase object and pointing towards the background from different viewing directions. The shape of the phase object at the time of capture is then inferred by evaluating the camera images. Background Technology
[0002] This method is known from DE 100 10 045 C2.
[0003] The so-called BOS method is known in principle to those skilled in the art. BOS stands for "BackOriented Schlieren" and describes a set of methods suitable for visualizing phase objects, or more generally, for detecting phase objects. In the context of this specification, a phase object refers to a spatially extending region that differs from its immediate surroundings optically, essentially only in its refractive index, within the spectral range used for observation. Typical examples of such phase objects include air vortices that differ in pressure distribution from ambient air, foreign bubbles in the original gas, or high-temperature foams that differ in refractive index from the surrounding original gas due to differences in their chemical composition or density. Those skilled in the art will understand that the term "different refractive index" does not imply that the entire phase object has a uniform refractive index, but also includes a refractive index distribution that distinguishes it from its environment. Such phase objects cannot be detected by means of purely intensity-supported sensing devices, particularly those provided by conventional cameras.
[0004] The fundamental understanding of BOS (Browser-Oriented System) technology is that light traveling from a background object point to the camera takes different paths depending on whether it passes through a phase object positioned between the background object point and the camera, and depending on which path it takes through that phase object. The background object point is correspondingly imaged onto another image point within the camera image of the camera capturing the background. Therefore, BOS technology initially stipulated that a textured background, i.e., one with any imageable pattern, is imaged directly in a reference image at the first shooting point, i.e., without a phase object positioned between the background and the camera, and then a measurement image is captured using the same spatial position for both the background and the camera, where the phase object is now positioned between the background and the camera, i.e., the background is captured through the phase object. The reference image and the measurement image show the same background, but with slight, usually locally different, relative movements of the image points (on which the background object point is imaged). This movement can be known, for example, through cross-correlation between the reference image and the measurement image. The degree of local movement can be inferred from the deflection of light rays from the background object point to the image point due to the phase object. From this, a measure of the shape and refractive index deviation of the phase object can be inferred. This data can then be used, for example, to visualize phase objects. A drawback of this basic BOS technique is, on the one hand, the time delay between shots, and on the other hand, the frequent inability in practice to capture reference images when there are no phase objects and measurement images when there are phase objects.
[0005] A method is known from DE 100 10 045 C2, which forms the class as described at the beginning of this paper, in which two camera images are simultaneously captured by two spatially offset cameras. Here, the two cameras capture the same background through a phase object, wherein the phase object and, in particular, its edges are observed by the cameras from different angles. Of particular interest are image points at which the corresponding background object points are imaged through the phase object in one camera image and beside the phase object in the other camera image. Thus, each camera image can be used partly as a reference image for the corresponding region of the other camera image and partly as a measurement image. However, for this purpose, the camera images need to be corrected for the different viewing angles of the camera images before a true comparison can be made that reveals the influence of the phase object. For this purpose, a global, typical quadratic transformation function is fitted, by which one image should be able to be transformed into another image while excluding the influence of the phase object. The comparison of the transformed images should only disclose differences that can be attributed to the influence of the phase object. However, contrary to the claims in the aforementioned literature, this approach only works in backgrounds with relatively flat depth structures: low-order polynomial transformation functions cannot depict significant depth structures; while high-order polynomial transformation functions take into account the effects of phase objects themselves, thus making these effects indistinguishable from differences caused solely by parallax. Conversely, based on models that accurately reproduce the background, it is entirely reasonable to assume that the selection of transformation functions presupposes an accurate understanding of the background, and therefore it is unsuitable for unknown and / or changing, such as moving, backgrounds. Summary of the Invention
[0006] The objective of this invention is to provide a BOS method that enables the detection of phase objects even against a background of motion with significant depth structures.
[0007] This task is solved by combining the features of the preamble of claim 1 through the following steps: a) Provide the inverse function of its imaging function for each camera, where the imaging function indicates the unique association between each object point detected by its respective camera and the image point in the corresponding camera image, and The inverse function indicates the unique association between each image point in the camera image and the line of sight. The line of sight is defined by all object points assigned to this image point through the imaging function. b) In each camera image, identify the background object points that are imaged on each camera image, that is, the image points that are imaged by the object points belonging to the background, and calculate the corresponding line of sight using the inverse function of their respective imaging functions. c) Obtain the measure used to calculate the distance between lines of sight. d) Assign the deviation value representing this metric to the corresponding image point in at least one camera image, and e) Repeat steps b to d for multiple background object points.
[0008] Preferred embodiments of the present invention are the subject of the dependent claims.
[0009] This invention essentially relates to the evaluation of images that can also be obtained using previously known methods. The images are textured backgrounds captured simultaneously through a phase object from different viewing angles. Preferably, three or four simultaneously captured images are used. However, the method according to the invention can also be performed using only two simultaneously captured images. However, unlike prior art, the evaluation of such camera images according to the invention is not based on a direct comparison of camera images, but rather utilizes the principle of triangulation.
[0010] The imaging characteristics of a camera can be described by its so-called imaging function. This imaging function indicates the assignment of each object point to an image point in the camera image. Explicitly creating such imaging functions is necessary and common in, for example, the field of solid shape measurement, and is therefore familiar to those skilled in the art. Those skilled in the art are also aware of high-precision calibration methods that allow for the creation of imaging functions with extremely high accuracy. From each imaging function, its inverse function can be calculated; that is, the function of the object space can be derived from the image points in the camera image. However, this inverse function is not unique due to the dimensional difference between the three-dimensional object space and the two-dimensional camera image. In particular, the so-called line of sight can be calculated for each image point in the camera image using the inverse function of the imaging function. This line of sight corresponds to a straight line on which all those object points imaged according to the imaging function to their assigned image points lie. The distance between the imaged object point and the camera in practice along the line of sight cannot be determined from the inverse function assigned to a single camera. However, if the same object point is imaged by multiple cameras from different viewing angles, provided that the imaging function or the cameras are correctly calibrated, then the lines of sight must intersect at a common point in the object space, i.e., the imaged object point. This is the basic assumption of triangulation. Therefore, to determine the location of the object point imaged on multiple cameras in the object space, the inverse function assigned to each camera is applied to the image point (on which the object point is imaged) and the intersection of the resulting lines of sight is calculated.
[0011] However, if the imaging process is disturbed by a phase object between the object and the camera, the fundamental assumptions explained above, particularly the straight optical path from the object to the image, no longer hold. The imaging function created for each camera based on straight-line light propagation, without considering the phase object, no longer accurately describes the practical imaging achieved through the curved optical path passing through the phase object. Accordingly, the inverse function of the imaging function no longer accurately represents the actual optical path. The calculated line of sight and the actual optical path are therefore no longer consistent. This is particularly evident in that, for each image point imaged from the object, the lines of sight calculated using the inverse function extend outwards from each other, that is, they do not intersect. This degree of outward divergence is a measure of the deviation between the calculated lines of sight and the actual optical path, and therefore a measure of the influence of the phase object on the association between the object and the image.
[0012] Within the scope of this invention, this knowledge is utilized as follows: First, the object points imaged on all (or at least multiple) camera images are known. Then, for each corresponding image point, the assigned (ideal) line of sight is calculated using an inverse function, and the degree of skewness of the line of sight is determined. A corresponding value, referred to herein as the deviation value, can then be assigned to the corresponding image point in one or more or all camera images. If this method is performed with multiple object points in the background, then these deviation values can be assigned to multiple image points within the camera images.
[0013] The deviation value can be determined in various ways using specialized mathematical methods. One feasible approach, purely exemplified, is to determine the sum of the minimum distances between lines of sight and use this as the deviation value. Alternatively, it is conceivable to define a virtual object point characterized by having a minimum sum of its vertical distances to the lines of sight. It is also conceivable to employ existing techniques known to those skilled in the art, namely epipolar geometry. Thus, for example, it can be defined that the degree of skewness is determined by the distance from an actual image point in a camera image to a corresponding, i.e., actual image point in (at least) another camera image, that is, assigned to the same object point. The line corresponding to the imaging of the line of sight assigned to the actual image point in the second camera image (or, when the mathematical structure of "line of sight" can actually be imaged) can be understood as the epipolar line assigned to the actual image point in the first camera image. Thus, in this variation of the method, at least in the first and second camera images, a corresponding actual image point assigned to a common object point is determined as either the first or second actual image point. Then, in the second camera image, a line is determined as the epipolar line, which is generated as the image of the line of sight from the first actual image point calculated by the inverse function of its imaging function. Finally, the distance between the second actual image point and the epipolar line is measured, for example, its perpendicular distance, and further used as the degree of skewness. When calculating multiple epipolar lines from different cameras in a single camera image, the multiple distances of the observed actual image points can, of course, be combined in an appropriate manner, for example, by averaging, to obtain the desired degree of skewness.
[0014] Of course, for some points, due to their special locations, in exceptional cases, the calculated lines of sight may intersect at a single point, even though they do not accurately represent the actual optical path. In such cases, the presence of a phase object is not inferred; instead, the location of the object point is incorrectly estimated. This exception may occur in the worst-case scenario when using only two cameras. However, in practice, this situation is ruled out when using three cameras that are deliberately not arranged in a straight line. If the aforementioned calculations are performed for multiple, especially a large number of, object points, these situations are easily identified as "anomalies" and ruled out by further evaluation.
[0015] A significant advantage of the method according to the invention is that no assumptions need to be made regarding the type of background and, in particular, its depth structure. It should only be noted that the depth structure, compared to the distance to the camera, is not such that background objects occlude each other, that is, they obstruct the camera's view of the background objects. However, this is a general requirement for a reasonable measurement structure, which is self-evident to those skilled in the art and not specific to the invention.
[0016] An image representing the local effect of a phase object on the image of the background through the phase object is generated by the method according to the invention. This image is preferably extracted as a separate dataset. In other words, another step of the method according to the invention is preferably specified, namely step...
[0017] f) Construct a graph that is structurally similar to the camera image, in which each graph point corresponding to an image point that has been assigned a deviation value in one of the performed steps d) is assigned a pixel value representing that deviation value.
[0018] The term "structurally similar" here should be interpreted as meaning that the data matrix referred to herein as a graph has, at least in the relevant region, the same matrix point row and column structure as the data matrix referred to herein as a camera image. In short, the preferably generated deviation value map is a matrix in which the matrix points associated with the deviation values are located in the same positions within the matrix as the image points (referring to the corresponding camera image) originally associated with the deviation values in the camera image.
[0019] To further utilize the representation of phase objects embodied in the diagram, the most diverse feasible solutions are provided to those skilled in the art. In particular, those skilled in the art can use the diagram for the visualization of phase objects. This can be achieved particularly simply by displaying the diagram on a monitor, wherein each diagram point is assigned a hue or grayscale tone corresponding to its respective pixel value according to a predetermined color value encoding or grayscale value encoding. This is comparable to the visualization of phase objects only.
[0020] Alternatively, the deviation value map can be overlaid with the corresponding camera image. This visualization thus shows the phase object in front of a background. The background can be reproduced, for example, in grayscale values, and the phase object or deviation value can be reproduced in color-coded form. This overlaid visualization is particularly important in measurements where the spatial position of the phase object relative to its environment is significant.
[0021] So far, only the following points have been mentioned, which correspond to image points in the camera image that are explicitly assigned deviation values through calculated lines of sight. This results in the visualization of phase objects by displaying them point-by-point or with gaps, which may not be satisfactory depending on the application. However, the increased computational cost usually only yields a slightly larger gain in the shape of the phase object. This is particularly suitable when the refractive index distribution within the phase object is as uniform as possible. In particular, but not only in these cases, it is conceivable to supplement the areas between points with explicitly calculated deviation values on the image by interpolation or extrapolation. Preferably, pixel values representing the interpolated or extrapolated deviation values are assigned to the image points corresponding to image points that were not assigned deviation values in step d, wherein the image points assigned pixel values representing deviation values in step f are used as grid points for their respective interpolation or extrapolation methods.
[0022] As described, each calculated map (according to the explained so-called structural similarity) corresponds to an underlying camera image. In other words, each map represents a phase object from the viewing angle of its respective associated camera. It is, in principle, not important to those skilled in the art to choose which specific camera image to use to construct the map. Typically, they would choose the camera image that most visually reproduces the phase image and background in a particular case. However, in an improved embodiment of the invention, multiple maps are constructed based on multiple, preferably all, camera images. This dataset thus represents the phase object from different viewing angles. The three-dimensional shape of the phase object can thus be calculated knowing the relative spatial arrangement of the cameras. Essentially known algorithms for stereoscopic photography can be used for this purpose and applied to the maps. This also allows for the stereoscopic display of the phase object on a monitor.
[0023] However, it should be clearly pointed out that, in order to effectively use the method according to the invention, at least when a complete three-dimensional detection of the phase object shape is not required, it is not necessary to position the cameras with a large spatial distance between them. In a preferred embodiment, for example, it is conceivable that the cameras are fixed close to each other with a common frame or even rigidly fixed to each other in a common housing. Such a rigid arrangement can avoid errors caused, for example, due to structural vibrations in harsh environments, such as industrial environments.
[0024] A key feature of the method according to the invention is that the camera images used for the calculation of the phase object, as detailed above, are captured simultaneously. This avoids errors caused by the movement of the phase object or background between two time-staggered shots. One or more resulting images thus represent the phase object at each specific shooting time point. However, in an improved embodiment of the invention, the basic method is repeated at one or more time points, and thus the temporal development of the phase object in terms of its shape and / or movement is also detected and, if necessary, visualized, for example, in video format.
[0025] The nature of the background is not important to the basic principles of this invention. It is conceivable to utilize naturally occurring backgrounds or artificially patterned backgrounds. The latter can be achieved using printing or projection techniques.
[0026] Further details and advantages of the present invention will become apparent from the following detailed description and accompanying drawings. Attached Figure Description
[0027] In the picture: Figure 1 A highly schematic view of three calibrated cameras pointing towards the background is shown; Figure 2 It shows Figure 1 The structure has a phase object placed in front of the background; Figure 3 It is a schematic diagram used to illustrate the relationship between the line of sight, the actual optical path, and the ideal optical path; Figure 4 Is Figure 1 and 2 A diagram illustrating the key steps of the method according to the invention performed at the structural location; and Figure 5 It is aimed at Figure 4 A possible visual illustration of the method steps.
[0028] The same reference numerals in the figures denote the same or similar elements. Detailed Implementation
[0029] Figures 1 to 4 The basic steps of the method according to the invention are shown in a highly schematic view. Figure 5 The optional visualization steps that follow are shown.
[0030] exist Figure 1 The diagram illustrates a possible configuration for performing the method according to the invention, wherein three cameras 10 observe scene 20 from different viewing angles. For illustrative purposes, the spatial positioning of the cameras 10 is depicted as being far apart from each other. In practice, a significantly smaller spacing is usually sufficient.
[0031] exist Figure 1 In the illustrated scenario, scene 20 consists solely of background 21, which, in the illustrated embodiment, is selected as a random, printed dotted pattern. Other types of texturing of background 21, including using a "natural" background or a projection of a structure onto a scattering surface, are also feasible. Sufficient space is required between background 21 and camera 10 to locate the actual target object according to the method of the invention, i.e., the phase object, which will be discussed further below. The imaging optics 11 of each camera 10 is preferably configured such that both the background 21 and the space within it where the phase object of interest is located can be clearly imaged.
[0032] Each camera 10 can capture camera images 12. Here, the camera image is an ordered data matrix composed of pixels, as is known to those skilled in the art from CCD cameras, CMOS cameras, or similar cameras. Figures 1 to 4 In the foregoing, only for better illustration, camera image 12 is shown in front of camera 10.
[0033] The illustrated structure of camera 10 can be calibrated in a manner known to those skilled in the art. That is, an imaging function can be created for each camera 10, which assigns image points on its respective camera image 12 to each object point in scene 20. While the imaging function can be based on different models, such as the so-called pinhole model, it is always based on the assumption that light travels in a straight, ideally straight light path 30. Similarly, an inverse function of the imaging function can be determined for each camera 10; however, due to the dimensional transformation between the three-dimensional scene 20 and the two-dimensional camera image 12, the inverse function is not as unique as the imaging function. More precisely, the inverse function assigns a line of sight 31 to each image point in the camera image 12, which can be considered a straight line consisting of all object points imaged onto the mentioned image point according to the imaging function. Ideally, that is, in the absence of "interfering" phase objects, light extends along the actual light path 32, which corresponds both to the ideal light path 30 assumed by the imaging function and to the line of sight 31 derived from the inverse function. Figure 1This is illustrated for background object point 211. This background object point is imaged in each camera image onto the image point 121 assigned by the imaging function in the respective camera image 12. Light here propagates along the actual travel path 32, which is the same as the ideal travel path 30. The respective inverse functions of the imaging functions, as explained above, define a line of sight 31 for each image point 121, which corresponds to both the actual optical path 32 and the ideal optical path 30 from the background object point 211 to the respective image point 121. These lines of sight 31 intersect at a point within the scene, that is, intersect at the object point assigned when correctly calibrated, i.e., intersect at the background object point 211 in the illustrated case. Those skilled in the art will understand that these interrelationships apply to both each object point and each image point.
[0034] Figure 2 This illustrates a situation where a phase object 40 is positioned between the background 21 and the camera 10. The phase object can be an external bubble, a high-temperature foam, or a region of abnormal pressure. Figure 2 In the situation shown, purely for illustrative purposes, the phase object 40 is arranged such that it only obstructs the view of the two lower cameras 10 to the background 21, while... Figure 2 The camera 10, positioned above, has unobstructed view of the background 21. However, the camera is clearly larger than... Figure 2 In practices where people gather more closely together, this situation is relatively rare. Figure 2 In this illustration, the phase object 40 is drawn semi-transparently to show its obstruction of the unobstructed view of the two cameras 10 below to the background 21. This is again for purely illustrative purposes, because the phase object 40 is essentially characterized by its absorption properties, which are indistinguishable from the environment and therefore cannot be identified by pure intensity measurements within the selected spectral range.
[0035] However, the path of light from the object point, particularly from the background object point 211, to the image point 121 is affected by the phase object 40. This is in Figure 3 This is illustrated in the diagram. Light travels from background object point 211 along the actual optical path 32, shown as a solid line, to image point 121. Because this path passes through phase object 40, the actual optical path 32 deviates from the ideal optical path 30, shown as a dashed line, which would be the path from background object point 211 to the ideal or predicted image point 121' by the imaging function without phase object 40. For the actual image point 121, the inverse function of the imaging function calculates a straight line of sight 31, shown as a dotted line, which corresponds neither to the actual optical path 32 nor to the ideal optical path 30.
[0036] exist Figure 2 For clarity, only the path to the ideal image point 121' is shown (in the image). Figure 2The ideal optical path 30 (dashed line) shown in the hollow image and the path from the actual image point 121 (shown by the inverse function) Figure 2 The line of sight 31 (solid line) calculated from the image 121 is shown in the center. Therefore, the actual image point 121 is not located in the same position in the camera image 12 as the ideal image point 121' originally predicted by the imaging function. The line of sight 31 (in the image 121') is obtained by means of the inverse function applied to the actual image point 121'. Figure 2 (shown as a solid line in the middle) is also not equivalent to (in Figure 2 (shown as dashed lines in the middle) occurs in the absence of phase object 40 (see...) Figure 1 The ideal optical path is 30. Figure 2 The same applies to the situation shown for the two lower cameras 10. For the upper camera 10, whose view to the background 21 is not obstructed by the phase object 40, the calculated line of sight 31 corresponds to the line of sight expected according to calibration or the ideal optical path 30 that coincides with the actual optical path.
[0037] In any case, the distortion caused by the phase object 40 will result in the lines of sight 31, calculated according to the inverse function provided and applied to the actual image point 121, not intersecting, but extending outwards from each other. This is in Figure 4 The magnified section A is schematically shown. The degree of anisotropy caused is also a measure of the effect of the phase object 40 on the imaging of the background object point 211. According to the method of the invention, this degree of anisotropy is therefore quantified in an appropriate manner and the corresponding value, i.e., the deviation value, is assigned to the respective image point 121 in the camera image 12. Figure 4 The attachment is symbolically indicated by the arrow 50.
[0038] To determine the degree of non-planarity, various methods are available to those skilled in the art, some of which have been exemplarily mentioned in the overview section of this specification. Using the imaging equation, the associated image points on the camera image 12 can be determined, and the distances between these image points of the virtual object point and the actual image point 121 can be calculated and used as the basis for determining the deviation value. The deviation value can, for example, be limited to the sum of the mentioned distances between the actual image point 121 and the calculated image points of the virtual object point. As another particularly simple variation, it is conceivable to simply confirm whether the calculated lines of sight 31 extend non-planarly to each other, that is, whether unobstructed observation of at least one camera to the background object point is hindered by the phase object 40, and to limit the deviation value to a purely binary value. However, other methods for determining the degree of non-planarity of the lines of sight 31 are available to those skilled in the art.
[0039] For a large number of background points, the method described above for a specific background point 211 is repeated. This results in a large number of image points in the camera image 12 being assigned offset values. Figure 5 A feasible scheme for extracting this information and creating a visualization of the phase object 40 based on it is illustrated. Therefore, a structurally similar figure 14 is generated for each camera image 12, that is, an ordered matrix of pixels whose arrangement corresponds at least to the arrangement of pixels in the camera image within the specific region of interest. Thus, each (associated) image point 121 in the camera image 12 is assigned a point 141 in Figure 14, wherein the point 141 has the same relative position to each other as the assigned image point 121. In Figure 14, the corresponding deviation value is then recorded at each point 141 assigned to the image point 121 (the previously mentioned image point has been assigned a deviation value). This results in a figure that (at least in a semi-quantitative sense) represents the shape and thickness of the phase object 40. This figure can be used for visualization of the phase object 40, either by directly displaying Figure 14 on a display or by overlaying Figure 14 with the associated camera image 12 under appropriate color or grayscale encoding, thereby allowing the phase object 40 to be visualized individually or in association with the background 21.
[0040] Other possible applications of Figure 14 can also be considered. Therefore, for example, the three-dimensional shape of the phase object 40 can be reconstructed using a stereoscopic photography algorithm. When the camera 10 observes the scene 20 from the viewing angle, such as... Figures 1 to 4 The reconstruction is particularly effective when the objects are clearly different from each other, as shown in the diagram.
[0041] Of course, the embodiments discussed in the specific description and shown in the accompanying drawings are merely illustrative examples of the invention. Those skilled in the art will recognize many possible variations based on this disclosure. The number of cameras 10 is not strictly limited. Using three cameras is considered particularly advantageous. With four cameras, accuracy can be improved through the resulting redundancy. Using more cameras, while increasing redundancy, does not significantly improve measurement accuracy further. There are no fundamental limitations on the application of this invention in terms of scene selection. Applications for visualizing flow events in the aviation field are conceivable, such as visualizing vortices behind the rotating blades of an aircraft wing or helicopter, where the naturally textured surface of the aircraft can serve as a background. Another application possibility is visualizing flows in so-called flow chambers or clean rooms in the pharmaceutical or chip manufacturing industries. Visualizing conventional flow and combustion processes can also be an area of application for this invention.
[0042] List of reference numerals
[0043] 10 cameras
[0044] 11 Imaging Optical System
[0045] 12 camera images
[0046] 121 actual image points
[0047] 121' Ideal Image Point
[0048] Figure 14
[0049] 141 map points
[0050] 20 scenarios
[0051] 21 Background
[0052] 211 Background Object Points
[0053] 30 Ideal optical path
[0054] 31. Line of sight
[0055] 32 Actual optical path
[0056] 40 phase objects
[0057] 50 Attached Arrows
[0058] A magnified section
Claims
1. A method for detecting a scene (20) including phase objects (40) arranged in front of a textured background (21), Within the scope of the method, multiple cameras (10) arranged spatially offset from each other and passing through the phase object (40) and pointing towards the background (21) from different viewing directions simultaneously capture a corresponding number of camera images (12). The shape of the phase object (40) at the time of capture is then inferred by evaluating the camera images (12). Its features It has the following steps: a) Provide an inverse function of its imaging function for each camera (10), wherein the imaging function indicates a unique association between each object point detected by its respective camera and the image point of the corresponding camera image (12), and The inverse function indicates the unique association between each image point in the camera image and the line of sight, which is defined by all object points assigned to that image point through the imaging function. b) In each camera image (12), identify the background object point (221) that is imaged on each camera image (12), that is, the image point (121, 121') that is imaged by the object point belonging to the background (21), and calculate the line of sight (31) respectively by means of the inverse function of their respective imaging functions. c) Obtain the measure used to calculate the distance between the lines of sight (31), d) Assign the deviation value representing this metric to the corresponding image point (121) in at least one camera image (12), and e) Repeat steps b to d for multiple background object points (221).
2. The method according to claim 1, characterized in that... It has the following additional steps: f) Construct a graph (14) that is structurally similar to the camera image (12), in which each graph point (141) corresponding to an image point (121) assigned a deviation value in one of the performed steps d is assigned a pixel value representing that deviation value.
3. The method according to claim 2, Its features are, The pixel values representing the deviation values obtained by interpolation or extrapolation are assigned to the corresponding map points (141) of the image points (121) that were not assigned deviation values in step d. The map points (141) that were assigned pixel values representing deviation values in step f are used as grid points for their respective interpolation or extrapolation methods.
4. The method according to any one of claims 2 to 3, Its features are, The figure (14) is displayed on the monitor, wherein each point (141) is assigned a hue or gray tone corresponding to its respective pixel value according to a predetermined color value encoding or gray value encoding.
5. The method according to any one of claims 2 to 4, Its features are, The figure (14) is overlaid with the associated camera image to visualize it on the display.
6. The method according to any one of claims 2 to 4, Its features are, Step d is performed for multiple camera images (12), especially at least three camera images (12) taken simultaneously by different cameras (10), and a corresponding number of multiple images (14) are created in step f.
7. The method according to claim 6, Its features are, The three-dimensional shape of the phase object (40) is calculated from the created graph (14) and with the relative spatial arrangement of the camera (10) known.
8. The method according to claim 7, characterized in that, The three-dimensional shape of the phase object (40) is displayed on the monitor.
9. The method according to any one of the preceding claims, characterized in that, The method is repeated at one or more time points to detect the evolution of the phase object (40) over time.
10. The method according to any one of the preceding claims, characterized in that, Use at least three cameras (10) to capture a corresponding number of camera images (12) simultaneously.