WO2010058010A2

WO2010058010A2 - Stereo camera equipment, method for continuously and automatically calibrating a stereo camera apparatus, computer program, computer program product and monitoring device for wind energy systems, buildings with transparent areas, runways and/or flight corridors of airports

Info

Publication number: WO2010058010A2
Application number: PCT/EP2009/065656
Authority: WO
Inventors: Harry Schlemmer; Holger Vogel
Original assignee: Carl Zeiss Optronics Gmbh
Priority date: 2008-11-24
Filing date: 2009-11-23
Publication date: 2010-05-27
Also published as: WO2010058010A3; EP2364553A2

Abstract

The invention relates to a stereo camera apparatus (1') comprising at least two adjusted thermal imaging cameras (3a, 3b) disposed at a defined distance to and aligned with one another, said apparatus provided with a calibration apparatus (19') for the continuous automatic calibration thereof, wherein the calibration apparatus (19') comprises a radiation source (21) and a reference beam path (20') for imaging a reference image of the radiation source (21) onto the respective thermal imaging camera (3a, 3b), wherein an outgoing collimated beam bundle (22) from the radiation source (21) is divided into at least two partial beam bundles (22a, 22b). Each of the at least two partial beam bundles (22a, 22b) are redirected to the camera through at least one optical redirecting element (23a, 23b) associated with the respective thermal imaging camera (3a, 3b), wherein at least one of the optical redirecting elements (23a, 23b) is adapted such that the redirection angle of the optical redirecting element (23a, 23b) is always constant within a tolerance range independent of the angular position of the optical redirecting element (23a, 23b) in the plane occupied by the primary beams of the reference beam path (20').

Description

Stereo camera devices, methods for the continuous automatic calibration of a stereo camera device, computer program, computer program product and monitoring device for wind turbines, buildings with transparent areas, runways and / or flight corridors of airports

The invention relates to a stereo camera device with at least two adjusted, each other at a defined distance angeord- Neten and aligned thermal imaging cameras, which is provided with a calibration device for their continuous automatic calibration. Furthermore, the invention relates to a method for the continuous automatic calibration of such a stereo camera device. The invention also relates to a computer program and a computer program product with program code means for carrying out such a method. Moreover, the invention also relates to a monitoring device for wind turbines, buildings with transparent areas, runways and / or flight corridors of airports with a stereoscopic detection of approaching or existing birds or bird swarms.

When taking off and landing aircraft, there are often collisions with birds or bird swarms. Flocks of birds refer to aggregations of birds of the same size and species, which often fly in the same direction. In particular, when cruising flight corridors in particular of airports with flight paths of birds of regional and national bird migration, which are often based on landscape structures such as waters, valleys or coastlines, this risk is greatly increased. In such collisions, it may damage u. a. come on the engines of airplanes.

In the older unpublished DE 10 2008 018 880.8 there is a monitoring device for wind turbines, buildings with transparent areas, runways and / or flight corridors proposed by airports with a stereoscopic detection of approaching or existing birds or bird swarms for carrying out a monitoring method, wherein parameters such as altitude, flight direction, flight speed, type and size of the birds or the bird swarms can be determined. In the field of wind turbines, buildings with transparent areas, runways and / or flight corridors, at least one stereo camera device is provided, which has at least two mutually defined and adapted spaced synchronous cameras during the recording, in particular thermal imaging cameras, whose Recording times are at least approximately identical and their respective fields of view have an overlapping area.

If cameras, in particular thermal imaging cameras or thermal imagers, are used for measuring tasks, such as the stereoscopy mentioned above, they should have constant properties, in particular mechanical stability. Since the accuracies achievable in the stereo evaluation of the images can not be maintained by simple measures in the construction of the stereo camera device, it must be determined by means of suitable methods how the system properties change in order to possibly correct them downstream. The above-mentioned stereo camera device for determining the presence and / or the speed of flight of the birds is based on an accurate determination of the distance traveled by the animals in a relatively short time interval. This requires precise location determination at two points, which is very sensitive to a relative error between the two lines of sight of the cameras. The aforementioned stereo camera devices should as far as possible be calibrated so that highly-accurate stereoscopic measurements can be achieved over very long periods of time. For a setup of such measuring systems, a detailed adjustment strategy is necessary for each assembly step, which ensures that in the last step at the place of installation with the then only limited available resources a comprehensive calibration the system can be achieved.

Deviations from the calibration performed during setup of the stereo camera device can often not be completely avoided over the desired operating time. So it may be z. B. by a bending of the mechanical basis or basic structure of the stereo system or by thermally induced changes in the structural elements and also the individual thermal imaging cameras to change the line of sight of the thermal imaging cameras come. For the continuous recognition and correction of these deviations, measures should be conceived which can run simultaneously with the actual measurements in the context of the stereo image evaluation, without disturbing them and which, of course, must not introduce any additional errors into the stereo camera device itself.

Strong temperature fluctuations and / or mechanical stresses z. B. Vibrations can lead to changes in camera parameters of the stereo camera device, in particular changes in the position and position and the orientation of the respective camera. Therefore, it may be necessary to recalibrate the stereo camera device at frequent intervals or should the stereo camera device employed be capable of continuous self-calibration.

DE 10 2007 050 558 A1 relates to a method for the continuous self-calibration of an image recording device, in particular a stereo camera system, in which a pair of images is detected by means of a stereo method, on the basis of which a rectified image pair is determined by a pixel-by-pixel correspondence analysis along epipolar lines.

A system and a method for generating spatial image representations, in particular of spatial thermal images, is disclosed in EP 1 484 628 A1. For further prior art, reference is made to JP 01251990 A.

Based on this, the object of the present invention is to provide a stereo camera device and a method for the continuous automatic calibration of a stereo camera device and a computer program product, which enable a simple and accurate tracking calibration of the stereo camera device, wherein changes of the system parameters determined during the start calibration be determined and corrected if necessary, in particular without generating additional interference or errors by the Nachführkalibrierung.

This object is achieved by claim 1. The object is also solved by claim 15. With regard to the method for continuous automatic calibration of a stereo device, the object is achieved by claim 20. With regard to the computer program or the computer program product, the object is achieved by claim 24 or claim 25.

According to the invention, a stereo camera device with at least two adjusted, mutually spaced and aligned thermal imaging cameras is proposed, which is provided with a calibration device for the continuous automatic calibration, wherein the calibration device comprises a radiation source and a reference beam path, by means of which a reference image from the radiation source to the respective thermal imaging camera is imaged, wherein a collimated beam emanating from the radiation source is divided into at least two partial beams, wherein the at least two partial beams are each directed via at least one of the respective thermal imaging camera associated optical deflection and at least one of the optical deflection such. that regardless of the angular position of the optical deflection element in the plane spanned by the main beams of the reference beam path, the deflection angle of the optical deflection element is always constant within a tolerance range.

Furthermore, an inventive method for continuous automatic calibration of the stereo camera device according to the invention is proposed, wherein the reference image is located on the at least two thermal imaging cameras at predetermined, in particular regular intervals, after which the relative position and / or the relative orientation of the reference image to the reference images of the other thermal imaging cameras is determined, and after which, upon detection of a change, a corresponding correction of the parameters for the stereo evaluation is determined and made.

These measures make it possible to calibrate stereo camera devices, in particular stereo thermal imaging cameras, so that highly accurate stereoscopic measurements can be achieved over very long periods of time. System accuracies are achieved which can not be realized by passive measures. The proposed method is based on a start calibration as well as an automatic follow-up calibration, which determines the changes against the system parameters for the stereo evaluation or the image processing which are present during the start calibration and makes appropriate corrections if necessary.

In order to keep the line of sight of the two thermal imaging cameras, which are mounted on a base structure with possibly higher base length, aligned for a long time, a reference to the axis of a reference beam path or reference collimator is established. For this purpose, the beam emanating from the reference beam path is divided and deflected the resulting partial beams with suitable deflection elements such as mirrors or prisms in the respective entrance pupil of a thermal imaging camera. In that at least one of the optical deflecting elements Mente is formed such that regardless of the angular position of the optical deflecting element in the plane defined by the main beams of the reference beam path plane of deflection of the optical deflecting element between the entrance angle of the respective partial beam and the exit angle of the respective partial beam is always constant within a tolerance range, it is ensured that the reference beams generated by the deflection are always aligned perpendicular or parallel to the axis of the reference beam path regardless of the exact angular position of the optical deflection element. The deflection angle of the optical deflection element is always at least approximately constant regardless of the angle of incidence of the corresponding partial beam in the plane spanned by the partial beam. The optical deflection element is, as it were, invariant with respect to disturbance variables (eg tilting). Any changes in angle of the reference beam path or the radiation source or the collimator generate the same-way shelves in both thermal imaging cameras and can thus be detected and distinguished from the relevant errors or deviations. In any case, the reference beams for both thermal imagers remain perpendicular or parallel to the current axis of the reference beam path, so that the lines of sight of the two thermal imagers can be controlled electronically or tracked via the stereo evaluation in the image processing.

The tolerance range of the deflection angle can be ± 5 mrad, in particular +1 mrad, preferably ± 0.05 mrad.

The deflection angle can be between 83 degrees and 97 degrees, in particular at 90 degrees.

According to the invention, provision may be made for the optical deflecting element to have two reflection surfaces fixed to one another in a corresponding angular position for generating the deflection angle. In a constructive embodiment of the invention, the optical deflecting element may have as reflecting surfaces two mutually fixed plane mirrors.

It is very advantageous if the optical deflection element is monolithic, in particular designed as a pentaprism.

By a monolithic design of the optical deflection element tilting of the reflection surfaces is avoided to each other. The use of a pentaprism as an optical deflecting element of a thermal imaging camera ensures that the reference beams generated by the deflection are always aligned perpendicular or parallel to the axis of the reference beam path, regardless of the exact angular position of the pentaprism, since it is a peculiarity of the pentaprism that the Exit angle of the light beam is always 90 degrees to the entrance angle. Therefore, the pentaprism is invariant to interference (eg tilting of the pentaprism).

According to the invention, it may further be provided that at least one optical subelement is provided for dividing the beam into the sub-beams.

As a result, various, in particular symmetrical reference beam paths can be used.

It is advantageous if the optical divider element is designed as a 90-degree prism. Tilting of the 90 degree prism in turn results in the same direction displacement of the reference images or target images, which can also be compensated, so that no further errors or deviations are introduced.

It is very advantageous if the reference beam path is symmetrical, since then the geometric conditions for the at least two thermal imaging cameras are the same. For this purpose can a mirrored 90-degree prism for pupil division be introduced into the reference beam path. The additional 90 degree prism advantageously no new errors are introduced, which are not recognizable.

For collimating the radiation beam emanating from the radiation source, the reference beam path can have a collimator or reference collimator.

Of course, the reference beam path itself should not introduce any uncontrollable errors or deviations into the stereo camera device. Possible sources of error are primarily unwanted changes in the positioning of the optical elements of the reference beam path, d. H. pentaprism, 90 degree prism and collimator. Purely translatory displacements are unproblematic. Since the added optical components only have plane surfaces and are arranged in the collimated optical reference beam path, only a pupil shift occurs and no change in the image position.

It is advantageous if the radiation source has a thermoelectric element, in particular a Peltier element. Such a thermoelectric element or cooling element can be used for both heating and cooling by applying current. Accordingly, the target or reference image can be displayed (reference image has a temperature difference to the ambient temperature) or hidden (reference image has ambient temperature). The reference image can be designed, for example, as a crosshair or crosshair structure and can be punched out of a carrier element (for example made of sheet metal) or etched into it. Such a carrier element can then be connected to the thermoelectric element or be at least part of the thermoelectric element.

According to the invention, it may further be provided that the exit surface of the at least one optical deflecting element is arranged at least approximately in the area of the entrance pupil which is preferably located in the entrance pupil of the respective associated thermal imaging camera.

The division of the beam path in the observation and reference beam path should preferably take place in the entrance pupil plane of the thermal imaging camera in the area of the entrance pupil, so that its image quality is not impaired.

It is advantageous if the reference beam path has a tube. This minimizes unwanted emissions. In addition, the tube may have matte black inner surfaces and additional aperture rings.

According to the invention, it can further be provided that in the direction of propagation of the observation light of the respective thermal imaging camera in front of the optical deflection element of the thermal imaging camera, a defocusing lens which can be swiveled in and out of the viewing beam path of the thermal imaging camera is provided for calibrating the detector of the thermal imaging camera.

Infrared detector arrays typically have a strong inhomogeneity of their single detector elements in terms of dark current and gain. In order to still be able to achieve an acceptable image quality, a calibration of the individual detector elements should be carried out (so-called non-uniformity C_orrection, NUC). For a sufficiently good homogenization of the image background, however, this is generally not sufficient, since in the infrared range there are, in addition to the desired scene radiation, further radiation sources which undesirably contribute to the observation image information. Due to the expected strong and also inhomogeneous radiation from the reference beam path into the thermal imaging camera, the calibration of the individual detector elements should already be able to detect as many of these radiation components as possible. Since the internal temperature of the stereo Camera equipment and the scene temperature (eg cold sky) can be very different from each other, the calibration should be performed as close to the scene temperature. It is therefore very advantageous to turn a defocusing lens between the at least one optical deflection element and a viewing window of the stereo camera device during calibration in order to detect the unwanted radiation components during the calibration. In order to prevent the reference structure or the reference image from being removed in advance by the calibration, the radiation source should be switched off during the calibration or regulated correspondingly to the internal temperature of the stereo camera device.

It is very advantageous if the reference beam path has at least one, in particular two Kepler telescopes. In this case, it can further be provided according to the invention that a Kepler telescope is arranged in each case between the optical splitter element and the optical deflecting element in the reference beam path for the at least two partial beams.

By these measures, the calibration accuracy can be further increased in an advantageous manner. By using a (1: 1) Kepler telescope, much larger image fields, i. H. Reference images are transmitted. Thus, for example, image rotations can be better detected. The F-number is also significantly improved. The Kepler telescope is also insensitive to adjustment and so to speak largely invariant with respect to disturbances (eg tilting).

In claim 15, an alternative embodiment of the stereo camera device according to the invention with two adjusted, each other in a defined distance and aligned thermal imaging cameras specified, which is provided with a calibration device for the continuous automatic calibration, wherein as a calibration device, the two thermal imaging cameras each in the area of their entrance pupil an optical deflection - Have element that deflects the output from the first thermal imaging camera or its own heat radiation as a reference image in the observation beam path of the second thermal imaging camera.

In order to keep the line of sight of the two thermal imaging cameras aligned, passive, d. H. Advantageously, a direct relationship between the two axes of the thermal imaging cameras are produced without additional heat or radiation source. For this purpose, in each case a small part of the pupil of a thermal imaging camera with a deflection element, for. As a prism or a mirror, are deflected into the pupil of the other thermal imaging camera. Since the temperature of the detector surfaces is very low relative to the scene, each thermal imager captures a cold image detail of the detector surface of its counterpart embedded in the scene. The respective optical deflecting elements can preferably be designed such that regardless of the angular position of the optical deflecting element in the plane spanned by the main beams of the reference beam path, the deflection angle of the optical deflecting element between the entrance angle of the respective partial beam and the exit angle of the respective partial beam is always within a tolerance range is constant. If, for example, a pentaprism is used for the optical deflecting element, then it is ensured that the reference beams generated by double deflection, regardless of the exact angular position of the respective pentaprism, are always parallel to the output beam, and thus the reference beam path formed by the respective other thermal imaging camera is nonexistent generated another error.

Claim 19 specifies a monitoring device for wind turbines, buildings with transparent areas, runways and / or flight corridors from airports with stereoscopic detection of approaching or existing birds or bird swarms. This can advantageously be due to the simple and accurate Nachführkalibrierung their we- at least one stereo camera device according to the invention operate independently over very long periods away.

Due to a detected relative shift of the reference image to the reference images of the further thermal imaging cameras, a displacement of a detector of at least one of the thermal imaging cameras or a tilting or bending of at least one of the thermal imaging cameras parallel and / or perpendicular to the reference plane or to a baseline of the stereo camera. be recognized meraeinrichtung.

Due to a relative rotation of the reference image to the reference images of the further thermal imaging cameras, a rotation of at least one of the thermal imaging cameras or of a detector of at least one of the thermal imaging cameras can be detected around its optical axis.

It is advantageous if a start calibration of the stereo camera device is performed.

After building the thermal imaging cameras of the stereo camera device residual errors may be present, which can not be e- limited by adjustment. Such errors can then be reduced only by calibration using suitable image processing methods. For this purpose, suitable targets or targets are used in the landscape at a known position. The distance should ideally correspond to the later working distance, which is however often not guaranteed by spatial conditions. The calibration process is then performed by comparison with the known target positions in the scene. The result of the calibration is a constant assignment of the field angles to the pixels of the thermal imaging cameras. Since changes in the arrangement may result from this point in time, the procedure for the continuous automatic calibration of the stereo camera device should be started immediately in parallel with the method according to the invention. This can then be compared to the later gene changes of the arrangement show.

The inventive method for continuous automatic calibration of the stereo camera device according to the invention is preferably realized as a computer program on an image processing device of the stereo camera device according to the invention. For this purpose, the computer program is stored in a memory element of the image processing device. By processing on a microprocessor of the image processing device, the method is carried out. The computer program can be stored on a computer-readable data medium (floppy disk, CD, DVD, hard disk, USB memory stick or the like) or an Internet server as a computer program product and can be transferred from there into the memory element of the image processing device. Such a computer program or computer program product with program code means is specified in claim 24 or claim 25.

Advantageous embodiments and modifications of the invention will become apparent from the dependent claims. Below are shown in principle embodiments of the invention with reference to the drawings.

Show it:

Figure 1 is a schematic representation of a stereo camera device according to the invention in a monitoring device;

FIG. 2 shows a simplified representation of an arrangement of the stereo camera device according to the invention from FIG. 1 in FIG

Area of a flight corridor of an aircraft; FIG. 3 shows a simplified representation of a start calibration of a stereo camera device;

Figure 4 is a schematic representation of a first embodiment of the stereo camera device according to the invention;

FIG. 5 shows a simplified schematic representation of the radiation Lengangs the stereo camera device according to the invention according to the first embodiment with an angle change of a collimator;

FIG. 6 a schematic representation of a second embodiment of the stereo camera device according to the invention;

FIG. 7 shows a simplified schematic illustration of the beam path of the stereo camera device according to the invention according to the second embodiment with a tilting of the collimator;

FIG. 8 shows a simplified schematic representation of a beam path of the stereo camera device according to the invention according to the second embodiment with a tilting of a divider element; 9 shows a simplified schematic representation of a beam path of the stereo camera device according to the invention according to the second embodiment with an indication of image layers;

FIG. 10 shows a simplified schematic illustration of a visual field of a reference beam path of the stereo camera device according to the invention;

Figure 11 is a simplified representation of an entrance pupil of a thermal imaging camera of the stereo device according to the invention with a pentaprism; Figure 12 is a simplified representation of the pitch of the entrance pupil of the thermal imager;

FIG. 13 a simplified representation of the division of the field of view of a thermal imaging camera in the reference beam path;

FIG. 14 shows a schematic representation of the functional principle of a pentaprism;

Figure 15 is a simplified illustration of error propagation of manufacturing errors in a pentaprism;

FIG. 16 shows a simplified illustration of the beam offset due to positioning errors of a pentaprism; FIG. 17 shows a schematic illustration of a third embodiment of the stereo camera device according to the invention. processing;

Figure 18 is a schematic representation of a part of the stereo camera device according to the invention according to the second embodiment; Figure 19 is a schematic representation of a portion of the stereo camera device according to the invention according to the second embodiment with pivoted Defokussierlinse;

FIG. 20 a schematic representation of the stereo camera device according to the invention in accordance with the second embodiment with the defocusing lens pivoted in;

Figure 21 is a schematic representation of the stereo camera device according to the invention according to the first embodiment with plane mirrors as optical deflection elements; and

Figure 22 is a schematic representation of the stereo camera device according to the invention according to the second embodiment with additional Kepler telescopes.

The invention is described below as part of a monitoring device for wind turbines, buildings with transparent areas, runways and / or flight corridors of airports. Of course, the invention is not limited to this application.

1 shows a stereo camera device 1 according to the invention for a monitoring device 2 for runways 110 and / or flight corridors 11 (see FIG. 2) of airports with a stereoscopic detection of approaching birds 6 or swarms of birds 6 ', where parameters such as flight altitude, flight direction, Airspeed and type / size of the birds 6 or the birds are 6 'determined, shown. One or more of such stereo camera devices 1 are arranged in the area of the runways 110 and / or the flight corridors 11 and have at least two thermal imaging cameras 3a, 3b arranged synchronously to one another in a defined and adapted spacing during the recording. The recording times of the thermal imaging cameras 3a, 3b are at least approximately identical and their respective fields of view 4a, 4b have an overlapping region 5. In the overlapping area 5, a bird 6 is detected as an object. The two thermal imaging cameras 3a, 3b are adjusted to each other and calibrated. For the thermal imaging cameras 3a, 3b both thermal imaging areas such as LWIR, MWIR, VLWIR, FIR, as well as SWIR, NIR come into consideration.

The stereo camera device 1 has an image processing device 7, which is provided for processing the image data recorded with the two thermal imaging cameras 3a, 3b.

The stereo camera device 1 also has a radio station 8 as an interface, in particular network interface, for communication with further stereo camera devices 1 or with higher-level systems, in particular air traffic control systems 9 (indicated in FIG. 1 by the double arrow 8 '). The stereo camera device 1 operates autonomously. However, by means of the networking or the radio transmission via the radio station 8 further stations or stereo camera devices 1 can be connected. The information as well as the recordings are therefore available outside the individual station. Mainly these data are transmitted to the air traffic control.

A monitoring method for runways 110 and / or flight corridors 11 of airports, with which approaching birds 6 or flocks of birds 6 'are detected stereoscopically by means of the monitoring device 2 or the stereo camera device 1, runs on the image processing device 7 of the stereo camera device 1 , where parameters such as flight altitude, flight direction, airspeed and type / size of the birds 6 or the bird swarms 6 'or their swarm density are determined. The parameters are determined by means of a stereo evaluation. In this case, absolute space points of the birds 6 to be detected or swarms of birds 6 'are determined by the at least two viewing angles on the area 5 recorded by the at least two thermal imaging cameras 3a, 3b of the stereo camera device 1. The wing speed of the birds 6 or of the flocks of birds 6 'is determined by observation over a corresponding period of time. Birds 6 or flocks of birds 6 'at a greater distance can also be detected, with a correspondingly longer focal length being used for the two thermal imaging cameras 3a, 3b. In addition, flying objects such as model airplanes, stunt kites or the like can be detected by the stereo camera device 1 (not shown).

An evaluation is made on the basis of the parameters and, if appropriate, a corresponding warning message is issued.

As can be seen from FIG. 2, a stereo camera device 1 with a stereo vision region or overlapping region 5 monitors a known flight route 10 of birds 6 or swarms of birds 6 ¹ . The stereo camera device 1 is arranged so that a flight corridor 11 or an intersection region 12 of the flight corridor 11 with the known flight route 10 of the birds 6 or swarms of birds 6 ^{1 is} monitored. In flight corridor 11, an aircraft 13 is shown by way of example. An arrival time of the detected birds 6 or the detected flock of birds 6 'at the crossing area 12 with the flight corridor 11 of the aircraft 13 is also determined.

Upon detection of birds 6 or swarms of birds 6 ', a warning message for initiating countermeasures or abatement or avoidance measures to the higher-level system, in particular air traffic control system 9, or to take-off or landing aircraft 13 is output, if necessary. The stereo camera device 1 can monitor in the context of the monitoring device 2, as indicated by dashed lines in Figure 2, also runways 110, wind turbines 111 or buildings with transparent areas 112.

The stereo camera device 1 should have constant properties, in particular mechanical stability. Since the at the Stereoauswertung the image processing achievable accuracies can not be kept constant by simple measures in the mechanical arrangement of the stereo camera device 1, must be determined by means of suitable methods, as the system properties of the stereo camera device 1 temporally change in order to possibly correct them downstream.

After construction of the thermal imaging cameras 3a, 3b of the stereo camera device 1, residual errors may be present which can not be eliminated by mechanical adjustment. Such errors can only be reduced by a local calibration using suitable image processing methods on the image processing device 7. In FIG. 3, the basic structure as part of a stereo camera device 1 according to the invention is shown. The thermal imaging cameras 3a, 3b are arranged and aligned on a base structure 14 in a housing 15 at a defined distance from one another. The housing 15 has viewing windows 16 for the observation beam paths or lines of sight 17a, 17b for the visual fields 4a, 4b of the thermal imaging cameras 3a, 3b. When the thermal imaging cameras 3a, 3b are mounted on the base structure 14 in the preferably air-conditioned housing 15, all previous residual errors add up to the assembly errors and the angular errors caused by the viewing windows 16. For starting calibration, suitable targets 18 in the landscape are used at a known position. As goals 18 are, for example, posts, panels or the like. The distance should ideally correspond to the later working distance. The start calibration process is then performed by locating the targets 18 through the stereo camera device 1 with a measurement of the associated angular coordinates and a comparison with the actual known positions of the targets 18 in the landscape. The result is a constant calibration of the thermal imaging cameras 3a, 3b, ie the individual pixels are assigned field angles. Furthermore, a plausibility test is carried out with the known heat map data. Since from this point on changes in the arrangement may result, for example due to temperature changes. The first measurement of the follow-up calibration should then be carried out after that. Thus, it is possible to automatically calibrate the stereo camera device 1 so that highly accurate stereoscopic measurements are made possible over very long periods of time. As a result, system accuracies can be achieved that are not achievable through passive measures. An automatic follow-up calibration or self-calibration is performed, which detects changes compared to the system parameters available during the start calibration.

4, according to the invention a stereo camera device 1 with two adjusted to each other in a defined distance and aligned thermal imaging cameras 3a, 3b proposed, which is provided with a calibration device 19 for the continuous automatic calibration, wherein the calibration device 19 a reference beam path 20 with a As a thermoelectric element or Peltier element with a reference structure formed radiation source 21, by means of which a reference image from the radiation source 21 is imaged onto the respective thermal imaging camera 3a, 3b, wherein an emanating from the radiation source 21 collimated beam 22 in two partial beams 22a, 22b is divided, wherein the two partial beams 22a, 22b are each directed to one of the respective thermal imaging camera 3a, 3b optical deflection element 23a, 23b directed to this. The optical deflecting elements 23a, 23b are designed in such a way that, independently of the angular position of the optical deflecting element 23a, 23b, in the plane spanned by the main beams of the reference beam path 20 (intermediate plane in FIG. 4 as well as in FIGS. 5 to 9 and 21). the deflection angle 29 of the optical deflection element 23a, 23b between the entry angle of the respective partial beam 22a, 22b and the exit angle of the respective partial beam 22a, 22b is always constant within a tolerance range. The tolerance range of the deflection angle 29 is ± 5 mrad, in particular ± 1 mrad, preferably ± 0.05 mrad. The diversion angle 29 is between 83 degrees and 97 degrees, especially at 90 degrees. The optical deflecting elements 23a, 23b have two reflecting surfaces 40a, 40b, 41a, 41b arranged at a corresponding angular position to generate the deflection angle 29. The optical deflection elements are monolithically designed as pentaprisms 23a, 23b. The reference beam path 20 has a collimator 24 for collimating the radiation beams.

In FIG. 21, an alternative embodiment of an optical deflecting element 23a ', 23b' has as reflection surfaces two plane mirrors 40a ', 40b', 41a ', 41b' which are fixedly arranged relative to one another.

As can be seen further from FIG. 4, the thermal imaging cameras 3a, 3b have detectors 25a, 25b with individual detector elements (not shown). The pentaprisms 23a, 23b have inlet surfaces 26a, 26b and outlet surfaces 27a, 27b. The exit surfaces 27a, 27b of the pentaprisms 23a, 23b are arranged at least approximately in the region of the entrance pupils 28a, 28b of the respective associated thermal imaging cameras 3a, 3b.

5, the observation beam paths 17a, 17b and the reference beam path 20 of the stereo camera device 1 according to the invention are indicated in greatly simplified form. As shown by the angles 29, any angular change of the collimator 24 results in co-located shelves in both thermal imaging cameras 3a, 3b and their detectors 25a, 25b, respectively. However, the reference beams or partial beams 22a, 22b for both thermal imaging cameras 3a, 3b remain in any case perpendicular to the current axis of the collimator 24, so that the lines of sight 17a, 17b of the two thermal imaging cameras 3a, 3b can optionally be controlled and tracked electronically.

FIG. 6 shows a further embodiment of a stereo camera device 1 'according to the invention with a symmetrically designed reference beam path 20'. This is an optimal beautiful splitter element 30 for dividing the beam 22 into the partial beams 22a, 22b provided. The optical subelement is designed as a 90 degree prism 30. It is advantageous to construct the reference beam path 20 as symmetrically as possible, so that the geometric conditions for both thermal imaging cameras 3a, 3b are as equal as possible. This is possible in a simple manner with a correspondingly mirrored 90-degree prism 30 for pupil division in the reference beam path 20. The following error analysis shows that the additional 90-degree prism 30 can not give rise to any new errors which are not recognizable.

Of course, a reference beam path 20, 20 'must not itself generate uncontrollable errors. Possible sources of error in the stereo camera device 1, 1 'according to the invention are primarily unwanted changes in the positioning of the added optical elements (pentaprisms 23a, 23b, 90-degree prism 30 and collimator 24). Purely translatory shifts are rather unproblematic. Since the added optical components 23a, 23b, 24, 30 have only planar surfaces and are arranged in the collimated beam path 22, only a pupil shift and no change in the image position can occur.

For the accuracy of the position velocity calculations, all angular changes in the reference plane, which in the present case is spanned by the base structure 14 and the lines of sight 17a, 17b of the thermal imaging cameras 3a, 3b, are particularly critical. Due to the special functional principle of these components, a rotation of the pentaprisms 23a, 23b does not cause any angle changes and thus does not generate any additional errors. A tilting of the collimator 24, in turn, causes a shift of the reference images in the same direction in both thermal imaging cameras 3a, 3b (see FIG. 7 or deflection angle 29), which can be detected and compensated, and thus also produces no error.

From a tilt angle αi of the collimator 24 results in the result erroneous deviation δi of the individual reference image on the respective detector 25a, 25b with the aid of the focal length F _{w of} the thermal imaging camera 3a, 3b to:

Filing error δ _x = F _w • tan oci [mm].

As in the case of the collimator 24, a tilting of the 90-degree prism 30 results in the same-direction displacement of the reference images in both thermal imaging cameras 3a, 3b, which can also be compensated so that no further errors occur (see FIGS. there deflection angle 29).

From the tilt angle 0C _{2 of} the 90-degree prism 30, the resulting erroneous deviation 8 _{2 of} the individual reference images on the respective detector 25 a, 25 b results with the aid of the focal length F _{w of} the thermal imaging cameras 3 a, 3 b:

Storage error 8 ₂ = F _w ^• tan 2α ₂ [mm].

A rotation of the collimator 24 about its optical axis generates a synchronous rotation of the reference images in both thermal imaging cameras 3a, 3b, since the image positions in both thermal imaging cameras 3a, 3b, as indicated in Fig. 9, are the same. This deviation can also be compensated so that no new errors occur. The respective image layers are indicated in FIG. 9 by crossed arrows 31.

The effects of erroneous tilting of the optical components 23a, 23b, 24, 30 perpendicular to the reference plane are less serious since they do not directly affect the accuracy of the stereo evaluation. A tilting of the collimator 24 and / or the 90-degree prism 30 in turn acts in the same direction on both thermal imaging cameras 3a, 3b and can therefore be detected and compensated. A tilting of a single pentaprism 23a, 23b leads to a shift of the reference image on the supplied impaired thermal imaging camera 3a, 3b, which can not be distinguished from the errors of the rest of the measuring system and therefore can lead to an additional error. With a correspondingly stable construction, however, it can be assumed that such errors are small or rare and thus tolerable.

From the tilt angle β _{3 of} the pentaprism 23a, 23b, the resulting erroneous deviation 6 _{3 of} the associated reference image on the detector 25a, 25b or the thermal imaging camera 3a, 3b with the aid of the focal length F _{w of} the thermal imaging cameras 3a, 3b yields:

Filing error 6 ₃ = F _w • tan ß ₃ [mm].

The entrance pupil of the collimator 24 lies in its objective (unspecified), while the entrance pupils 28a, 28b of the objectives of the thermal imaging cameras 3a, 3b lie in a relatively small distance in front of the respective first lens vertices. Since no lens optics is otherwise provided in the reference beam path 20, 20 ', the maximum transferable reference field FOV _{REF of} the reference beam path 20, 20' results from the size of the pentaprism 23a, 23b and the base length b (b = 1.5 m in the present exemplary embodiment) ). For larger field angles, the beams are cut up to 100%. From the geometry shown in FIG. 10, it follows that only a field point from a thermal imaging camera 3a, 3b can still be imaged with finite brightness when the associated beam strikes the entrance surface 26a, 26b of the associated pentaprism 23a, 23b. From the angle condition for the marginal ray bundle which is just no longer impinging (comparison to complete section of the pixel), a contact angle α of the field of view for the reference beam path 20 'results:

Field of view edge angle α ₃ = arc tan p / b,

where b is the base length and p is the height of the entrance surface 26a, 26b of the pentaprism 23a, 23b. The full field of view FOV _REF is twice as large, ie

Reference field FOV _REF = 20C ₃ .

For b = 3 m and p = 10 mm results for CX ₃ = 0.2 degrees. The full reference field FOV _REF is thus about 0.4 degrees, which corresponds to about 24 pixels. The brightness is maximum in the middle and falls to the edge to 0%. On the detector 25a, 25b of the thermal imaging camera 3a, 3b with 640 x 512 pixels, the reference image occupies a maximum of 24 x 24 pixels.

The division of the beam path in observation beam path 17a, 17b and reference beam path 20, 20 'should be done in the entrance pupil 28a, 28b of the thermal imaging camera 3a, 3b, so that the image quality is not impaired. A realistic division ratio can be seen from FIG. In the present exemplary embodiment, the entrance pupil 28a, 28b of the thermal imaging camera 3a, 3b has a diameter of 27.5 mm. The taprism 23a, 23b has an area of 10 mm × 10 mm, with shadowing 32 of the pupil 28a, 28b of the thermal imaging camera 3a, 3b of 11 mm × 11 mm being present. The pupil of the reference beam path 20, 20 'is therefore 10 mm × 10 mm. Due to the spatial extent of the pentaprism 23a, 23b, an additional loss of area due to shadowing occurs, as shown in FIG. The resulting field of view 33 is shown again in FIG. As a result of the inserted pentaprism 23a, 23b and the shadowing 32, the area of the entrance pupil 28a, 28b of the observation beam path 17a, 17b of the thermal imaging camera 3a, 3b is reduced, as a result of which the F number deteriorates somewhat.

The entrance surface 26a, 2βb of the pentaprism 23a, 23b represents the effective pupil surface for the reference beam path 20, 20 '. As shown in FIG. 13, the thermal imaging camera 3a, 3b with the largest part 34 of its entrance pupil 28a, 28b looks at the pentaprism 23a, 23b over. The smaller part 35 of the entrance pupil 28a, 28b looks through the pentaprism 23a, 23b in the direction of the collimator 24. Due to the large focal length b and the small collimator pupil, however, the collimator 24 only fills the very small part 35 of the visual field with a diameter of 23 mm, while the much larger part 34 of the entrance pupil 28a, 28b vorbeischaut the collimator 24. The part 35 must be blocked with appropriate apertures to prevent unwanted image information (eg from hot device components or the like) from being detected.

FIG. 14 shows a pentaprism 23a, 23b with the mechanical dimensions and the functional principle. The height of the entrance surface 26a, 26b is designated by x. By two-fold 45 degree deflection, the pentaprisms 23a, 23b generate a dropping beam that is at 90 degrees to the incident beam within manufacturing tolerances. This 90 degree deflection is independent of the angle of incidence in the plane defined by these beams. In practice, this means that a pentaprism 23a, 23b, regardless of its angular position, always deflects the incident beam by 90 degrees. A faulty position of the pentaprism 23a, 23b thus produces no faulty storage of the desired 90 degree direction. The accuracy of the beam deflection is determined by the manufacturing angle tolerances of the pentaprism 23a, 23b. From the angle errors α ₄ , β ₄ and γ ₄ and the refractive index n of the prism substrate, as shown in FIG. 15, the total angle error of the emergent beam is δ ₄ = n- (3α ₄ + β ₄ +

Y4) •

In addition to the angular errors in the tangential plane, a pentaprism 23a, 23b can also have production-related pyramidal defects. The pyramidal errors can theoretically be attributed to the tilting of the reflecting surface. In the case of pentaprisms 23a, 23b, this results in a corresponding angular deflection relative to the tangential plane (drawing plane in FIG. 15). Another error is that the surfaces need not be flat, but may have a very flat curvature, which leads to a refractive power. For technical reasons, this is often positive; in the general case, it will also be cylindrical. The pentaprism 23a, 23b is then a combination of a lens with concave mirrors and forms an optical system with a corresponding refractive power. As can be seen from FIG. 16, positioning errors in the tangential plane do not produce any angle errors, but result in only one beam offset δy. As long as the pentaprism 23a, 23b is operated in the collimated beam path, this only means a pupil offset δy, which as a rule is harmless. The positioning errors are indicated by dashed lines in FIG. In the left part of the figure in Figure 16 is an angle error, in the middle part of the figure to a radial displacement and the right part of the figure to an axial displacement.

In Figure 17, a third embodiment of a stereo camera device according to the invention is l ^'1 with two aligned, mutually arranged at a defined distance and aligned heat mebildkameras 3a, 3b, which with a calibration device 19''for their continuous automatic calibration The two thermal imaging cameras 3a, 3b each have, in the region of their entrance pupils 28a, 28b, an optical deflection element formed as pentaprism 23a, 23b, which transmits the image emitted by the first thermal imaging camera 3a or its detector 25a own heat radiation as a reference image in the observation beam 17b of the second thermal imaging camera 3b deflects.

In order to keep the lines of sight 17a, 17b of the two thermal imaging cameras 3a, 3b aligned, a direct relationship between the two axes of the thermal imaging cameras 3a, 3b can also be produced passively (without additional heat source or radiation source 21). For this purpose, in each case a small part of the pupil of a thermal imaging camera 3a, 3b with prisms 23a, 23b or mirrors is in each case deflected into the pupil of the respective other thermal imaging camera 3a, 3b. Since the temperature of the detector surfaces is very low relative to the scene, each thermal imaging camera 3a, 3b sees a cold image detail of the detector surface of its counterpart embedded in the scene. If one uses pentaprisms 23a, 23b for the deflection, then again it is ensured that the reference beams generated by double deflection are always parallel to the output beam, regardless of the exact angular position of the respective pentaprism 23a, 23b and thus the reference system does not generate new errors.

Thermal image detector arrays typically exhibit strong inhomogeneity of their single detectors in terms of dark current and gain. In order nevertheless to achieve an acceptable image quality, therefore, at least one calibration of the detector elements must be carried out. This is known as Non-Uniformity Correction (NUC). However, this is generally not sufficient for a good homogenization of the image background since, in addition to the desired scene radiation, there are other radiation sources in the infrared range which contribute undesirably to the image information. As can be seen from FIG. 18 for a cutout region of the stereo camera device I ¹ according to the invention, there are the following irradiation mechanisms. The inner surfaces of the measuring structure radiate via the residual reflection of lenses, prisms 23a and viewing windows 16 on the detector surface. This effect also occurs in each thermal imaging camera 3a. The attachment parts of the pentaprism 23a irradiate the detector 25a homogeneously since they are located in the entrance pupil 28a. The unavoidable Sehfeldbeschnitt in the reference beam 20 through a tube 36, means that, conversely, the tube walls in the field of view strong and also radiate inhomogeneous. The reference image in the collimator 24 and the surrounding structures are inevitably imaged sharply on the detector 25a of the thermal imaging camera 3a. Overall, there is a high radiation background. Effective measures to reduce this effect are z. B. matt black inner surfaces of the tube 36 with additional aperture rings 36a and a rough dull black version of the surface of the reference image and the interior of the collimator 24 and introduction of a field stop 37 in an intermediate image plane of the lens of the thermal imaging camera 3a. Nevertheless, a comprehensive calibration involving as many unwanted radiation sources as possible is additionally necessary. Due to the expected strong and also inhomogeneous radiation from the reference beam path 20 'in the thermal imaging camera 3a, the calibration method should be able to detect as many of these radiation components as possible. The unwanted radiation components are indicated in FIG. 18 by dashed arrows. Since the inside temperature of the device and the scene temperature (eg, cold sky) are often extremely different, the calibration should be performed at a temperature near the scene temperature if possible. In the direction of propagation of the observation light of the thermal imaging camera 3a, a defocusing lens 38 is provided in front of the pentaprism 23a of the thermal imager 3a for calibrating the detector 25a of the thermal imager 3a, in the observation beam path 17a of the thermal imager 3a. If the defocussing lens 38 is placed between the pentaprism 23a and the viewing window 16, virtually all unwanted irradiation components are detected. The scene is provided with the reference numeral 39.

FIG. 19 shows the normal operating situation with the defocusing lens 38 swiveled out, in which the observation beam path 17a and the reference beam path 20 'are focused onto the detector 25a. The detector 25a also sees unwanted heat radiation from the tube 36 and the device interior.

In FIG. 20, the defocusing lens 38 is swiveled in for calibration, whereby the observation beam path 17a is purposefully defocused and the image of the scene 39 on the detector 25a is smeared. Overall, the detector 25a thus experiences a quasi-homogeneous illumination with thermal radiation, the intensity of which corresponds to the scene temperature. In addition, the unwanted heat radiation from the tube 36 and the device interior again hits the detector 25a. If a calibration is carried out in this configuration, then in addition to the scene-typical radiation background, all unwanted effects are recorded and can thus be eliminated by suitable image correction algorithms from the subsequently recorded thermal images. To prevent the reference mark or reference image from being burned during calibration, d. H. is included in the calibration, the heat source or radiation source 21 must be switched off behind the reference structure or even better regulated to the internal device temperature. This is done by the thermoelectric cooler or the pelletizing element of the radiation source 21.

FIG. 22 again shows the second embodiment of the stereo camera device 1 according to the invention according to FIG. 6. In this case, a Kepler is provided in the reference beam path 20 'for the at least two partial beams 22a, 22b between the optical splitter element 30 and the optical deflection elements formed as pentaprisms 23a, 23b Telescope 42a, 42b arranged. The Kepler telescopes 42a, 42b can in principle be equipped with the necessary manoeuvrable changes in all embodiments of the stereo camera device 1, 1 ', 1 according to the invention ^{? I} will be employed.

In the context of the stereo evaluation, a method for the continuous automatic calibration of the stereo camera device 1, I ¹ , I 'can now be provided on the image processing device 7 ^! wherein the reference image is located on the two thermal imaging cameras 3a, 3b at predetermined, preferably regular time intervals, after which the relative position and / or the relative orientation of the reference image to the reference images of the further thermal imaging cameras 3a, 3b is determined and from this upon detection of a change, a corresponding correction of the parameters for the stereo evaluation is determined and carried out.

Due to a detected relative displacement of the reference image, a displacement of the detector 25a, 25b of at least one of the thermal imaging cameras 3a, 3b or a tilting or bending of at least one of the thermal imaging cameras 3a, 3b parallel and / or perpendicular to the reference plane or to the baseline 14 of the Stereo camera device 1, 1 ', I' ¹ detected. In addition, due to a relative rotation of the reference image, a rotation of at least one of the thermal imaging cameras 3a, 3b or the detector 25a, 25b of at least one of the thermal imaging cameras 3a, 3b is detected about its optical axis.

The inventive method for continuous automatic calibration of the stereo camera device 1 according to the invention, 1 ', l' ¹ is preferably beitungseinrichtung as a computer program on the Bildverar- 7 of the stereo camera device according to the invention 1,1 ', I ^1' realized, other solutions also occur naturally in question. For this purpose, the computer program is stored in a memory element, not shown, of the image processing device 7. By processing on a microprocessor of the image processing device 7, the method is carried out. The computer program can be stored on a computer-readable medium (floppy disk, CD, DVD, hard disk, ÜSB). Memory stick or the like) or an internet server as a computer program product and from there to the storage element of the image processing device 7.

Claims

Claims:

1. Stereocamera device (1,1 ', I ¹ ') with at least two adjusted, mutually spaced at a defined distance and aligned thermal imaging cameras (3a, 3b), which provided with a calibration device (19,19 ') for their continuous automatic calibration wherein the calibration device (19, 19 ') has a radiation source (21) and a reference beam path (20, 20', 20 ''), by means of which a reference image from the radiation source (21) onto the respective thermal imaging camera (3a, 3b) wherein a collimated beam (22) emanating from the radiation source (21) is divided into at least two partial beams (22a, 22b), wherein the at least two partial beams (22a, 22b) are respectively connected via at least one of the respective thermal imaging camera (3a, 3b ) are directed to this, and wherein at least one of the optical deflection elements (23a, 23a ', 23b, 23b') is formed such that regardless of the angular position of the optical deflection element (23a, 23a ', 23b, 23b') in the plane defined by the main beams of the reference beam path (20, 20 ', 20''), the deflection angle (29) of the optical deflection element (23a, 23a') , 23b, 23b ') is always constant within a tolerance range.

2. Stereo camera device according to claim 1, characterized in that the tolerance range of the deflection angle ± 5 mrad, in particular ± 1 mrad, preferably ± 0.05 mrad.

3. Stereo camera device according to claim 1 or 2, character- ized in that the deflection angle (29) between 83 degrees and 97

Degrees, especially at 90 degrees.

4. Stereo camera device according to claim 1, 2 or 3, characterized in that the optical deflection element (23a, 23a ', 23b, 23b') for generating the deflection angle two in a corresponding angular position fixed to each other arranged reflection surfaces (4 Oa, 4 Ob, 4 Ia, 4 Ib, 40a ', 40b', 41a ', 41b').

5. Stereo camera device according to claim 4, characterized in that the optical Umlenkelernent (23a, 23b) as reflection surfaces two mutually fixed plane mirror (4Oa ', 40b', 41a ', 41b').

6. Stereo camera device according to one of claims 1 to 4, characterized in that the optical deflection element monolithic, in particular as a pentaprism (23a, 23b) is executed.

7. Stereo camera device according to one of claims 1 to 6, characterized in that at least one optical splitter element (29) for dividing the beam (22) in the Teilstrah- lenbündel (22a, 22b) is provided.

8. Stereo camera device according to claim 7, characterized in that the optical splitter element is formed as a 90-degree prism (29).

9. Stereo camera device according to one of claims 1 to 8, characterized in that the reference beam path (20 ', 2O ¹ ') is symmetrical.

10. Stereo camera device according to one of claims 1 to 9, characterized in that the reference beam path (20,20 ') has a collimator (24).

11. Stereo camera device according to one of claims 1 to 10, characterized in that the radiation source (21) has a thermoelectric element, in particular a Peltier element (21).

12. Stereo camera device according to one of claims 1 to 11, characterized in that the exit surface (26a, 26b) of the optical deflecting element (23a, 23a ', 23b, 23b') (23a, 23b) in the loading rich of the entrance pupil (27a, 27b), preferably in the entrance pupil (27a, 27b) of the respective associated thermal imaging camera (3a, 3b) is arranged.

13. Stereo camera device according to one of claims 1 to 12, characterized in that the reference beam path (20, 20 ', 2O ¹¹ ) has a tube (35).

14. Stereo camera device according to one of claims 1 to 13, characterized in that in the propagation direction of the observation light of the respective thermal imager (3a, 3b) in front of the optical deflecting element (23a, 23a ', 23b, 23b') of the thermal imaging camera (3a, 3b) Calibration of the detector (25a, 25b) of the thermal imaging camera (3a, 3b) is provided in the observation beam path (17) of the thermal imaging camera (3a, 3b) and can be pivoted defocusing lens (38).

15. Stereo camera device according to one of claims 1 to 14, characterized in that the reference beam path (20 ') at least one, in particular two Kepler telescopes (42a, 42b).

16. Stereocamera device according to claim 15, characterized in that in the reference beam path (20 ') for at least two partial beams (22a, 22b) between the optical splitter element (30) and the optical deflecting element (23a, 23b) a Kepler Telescope (42a, 42b) is arranged.

( ^'? L) 17 adjusted stereo camera device having two, zueinan- the arranged at a defined distance and aligned thermal imaging cameras (3a, 3b) provided with a calibration device

(19 '') is provided for their continuous automatic calibration, wherein as calibration means (19 ''), the two thermal imaging cameras (3a, 3b) each in the region of their entrance pupil (27a, 27b) an optical deflecting element (23a, 23a ', 23b, 23b '), which the from the first thermal imaging camera (3a) and / or from the detector (25a) emitted own thermal radiation, as a reference image in the observation beam path (17) of the second thermal imaging camera (3b) deflects.

18. Stereo camera device according to one of claims 1 to 17, which at least part of a monitoring device (2) for wind turbines, buildings with transparent areas, runways and / or flight corridors (11) of airports with a stereoscopic detection approaching or IN ANY - which birds (6) or swarms of birds (6 ¹ ) are to carry out a monitoring procedure, whereby parameters such as flight altitude, flight direction, airspeed, type and size of birds (6) or flocks of birds (β ^? ) can be determined, on the basis of which an assessment and, if appropriate, issuing a corresponding warning, and which is arranged in the area of wind turbines, buildings with transparent areas, runways and / or flight corridors.

19. Monitoring device (2) for wind turbines, buildings with transparent areas, runways and / or flight corridors (11) of airports with a stereoscopic detection of approaching or existing birds (6) or bird swarms (6 ') for carrying out a monitoring method , where parameters such as flight altitude, flight direction, airspeed, type and size of the birds (6) or the bird swarms (6 ') can be determined on the basis of which an assessment is carried out and, if appropriate, a warning message is issued, and wherein in the field of wind turbines, buildings with transparent areas, runways and / or the flight corridors (11) at least one stereo camera device (1,1 ', I ¹ ') according to one of claims 1 to 18 with at least two thermal imaging cameras (3a, 3b) is provided, which during the Recording synchronous run whose recording times are at least approximately identical and their respective field of view r (4a, 4b) have an overlapping region (5).

20. A method for the continuous automatic calibration of a stereo camera device (1,1 ', I ¹ ') according to one of claims 1 to 18, characterized in that the reference image on the at least two thermal imaging cameras (3a, 3b) in predetermined, in particular regular Time intervals is localized, after which the relative position and / or the relative orientation of the reference image to the reference images of the other thermal imaging cameras (3a, 3b) is determined, and then it is determined upon detection of a change, a corresponding correction of the parameters for the stereo evaluation and made.

21. The method according to claim 20, characterized in that due to a detected relative displacement of the reference image, a displacement of a detector (25a, 25b) at least one of the thermal imaging cameras (3a, 3b) or a tilting or bending of at least one of the thermal imaging cameras (3a, 3b) parallel and / or perpendicular to a reference plane or to a base structure (14) of the stereo camera device (1,1 ', 1' ') is detected.

22. The method according to claim 20 or 21, characterized in that due to a relative rotation of the reference image, a rotation of at least one of the thermal imaging cameras (3a, 3b) or a detector (25a, 25b) at least one of the thermal imaging cameras (3a, 3b) about the optical Axis is detected.

23. The method of claim 20, 21 or 22, characterized in that a start calibration of the stereo camera device

(1,1 ', 1' ') is performed.

24. Computer program with program code means for performing a method according to one of claims 20 to 23, when the program on a microprocessor of a computer, in particular on an image processing device (7) of a stereo camera device (1,1 ', I ¹ ') according to a of claims 1 to 16, is executed.

25. Computer program product with program code means which are stored on a computer-readable data carrier to perform a method according to one of claims 20 to 23, when the program on an image processing device (7) of a stereo camera device (1,1 ', I ¹ ') according to one of Claims 1 to 18, is executed.