DE102008046964A1 - Image recording unit i.e. thermographic camera, for consolidating thermographic images produced by image sensors, has signal processing unit converting signals from optical unit into signals of another optical unit using formula - Google Patents

Abstract

The recording unit (1) has an optical unit (2) with a wavelength sensitivity lying between 0.4 to 1.2 micrometers. Another optical unit (3) includes a wavelength sensitivity lying between 8 to 12 micrometers. A distance sensor (12) e.g. optical distance sensor, is arranged coplanar to image sensors (5, 8). The optical units are connected with a signal processing unit (14) such that the processing unit obtains signals processed by the optical units per pixels of images and converts the signals from one of the optical units into signals of other optical unit using a specific formula.

Description

The The invention relates to an image acquisition unit for fusion with Sensors of different wavelength sensitivity generated images according to the preamble of the claim 1.

in the Infrared range, objects have a completely different emission, Reflection and transmission behavior as in the visible wavelength range. It is therefore useful for many machine vision applications obvious advantage, besides imaging sensors for the visible and near infrared range, that is a wavelength range to about 1.2 microns, in more detail a wavelength range from about 0.4 μm to 1.2 μm, hereinafter short called image sensors, also corresponding sensors for the long-wave infrared range, that is a wavelength range to about 14 microns, in more detail a wavelength range from about 8 μm to 12 μm, to test, Measuring and monitoring purposes to use.

Latter Sensors are usually called thermal image sensors denotes, as with them mainly emitted by a target object Heat radiation is measured. The pictures are called according to thermographic images or simply thermal images.

A The weaknesses of thermal images lies in the local Resolution, that is, it is difficult, important Accurately identify and locate details in the image.

• 1. Wärmebildsensoren haben, bedingt durch den Herstellprozess, bei weitem nicht die Anzahl von Pixel, die von Bildsensoren her bekannt sind. Der Stand der Technik im Jahre 2006 in diesem Punkt ist: 640×480 Pixel.
• 2. Wärmebildsensoren sind extrem teuer. Daher besteht aus Kostengründen oft der Zwang, auf Varianten mit einer niedrigen Auflösung, zum Beispiel 160×120 Pixel, auszuweichen. Bei Technologien, die eine kostengünstige Herstellung versprechen, ist die bisher erreichte Auflösung noch geringer, nämlich maximal 32×32 Pixel.
• 3. Zusätzlich zur Emission von Strahlung wird Wärmeenergie auch mittels Wärmeleitung abgegeben. Bedingt durch diesen Wärmeausgleich mit der Nachbarschaft werden Objekte im Wärmebild nur unscharf dargestellt.
• 4. Die Wärmestrahlung realer Körper ist abhängig vom Emissionsgrad. Daher können im Wärmebild Artefakte entstehen, hervorgerufen durch unterschiedliche Arten von Oberflächen.

This limitation is due to several factors:

• 1. Thermal image sensors, due to the manufacturing process, are nowhere near the number of pixels known from image sensors. The state of the art in 2006 at this point is 640 × 480 pixels.
• 2. Thermal image sensors are extremely expensive. Therefore, for cost reasons, there is often a need to avoid variants with a low resolution, for example 160 × 120 pixels. In technologies that promise a cost-effective production, the previously achieved resolution is even lower, namely a maximum of 32 × 32 pixels.
• 3. In addition to the emission of radiation, heat energy is also dissipated by heat conduction. Due to this heat balance with the neighborhood, objects in the thermal image are only displayed blurred.
• 4. The heat radiation of real bodies depends on the emissivity. Therefore, artifacts caused by different types of surfaces can arise in the thermal image.

Because of The reasons listed today will be the evaluation of thermal images with images of image sensors in the visible Working area, combined. The image sensors working in the visible range originating images, hereinafter referred to as visual images, offer more resolution, sharpness and contrast and therefore allow precise location and size determination to the objects detected in the thermal image.

The combining of thermal image and visual image sensor, for example, in the document EP 0 973 137 A1 described without going into special methods and designs.

Of the Benefit of a combination of heat and visual image Specifically, there is an association between the different produce picture objects.

Mostly this is relatively easy for a human observer. He compares both pictures and, due to his knowledge of the shown scene corresponding assignments.

In In this case, it does not matter if image sensor and thermal imaging sensor spatially different positions and different directions of view to have.

Should however, this comparison happen automatically, it is necessary that the heat and the visual image are as congruent as possible to each other. Only then can both pictures be easier Be linked by means of local operators. Otherwise Difficult methods for model-based object recognition would have to be used become.

It So it is necessary that the two pictures, with two after different Principles of working imaging sensors are generated Cover be brought.

Coincident, for example, two-dimensional, images of any spatial Arrangement can with two different sensors for Example can be created by the position and the viewing direction the sensors at least approximately agree.

For this are next to improvised solutions, where successively a thermal imaging camera and a second, with an image sensor equipped camera on the identical field of view and identical Perspective, from company publications Thermal imaging cameras are known, which is an additional sensor module integrated for the visible area.

• 1. Fluke Switzerland GmbH: „Neue Thermografie-Technologie: Wartungsproblemen schneller auf der Spur”, polyscope 18/06, S. 28–30 [1].
• 2. Roger Schmidt of Fluke Thermography, „Benefits of IR/Visible Fusion”, 2007 [2].
• 3. K. Johnson, T. McManus und R. Schmidt, Infrared Solutions, Inc., ”Commercial fusion camera”, Proc. of SPIE Volume 6205 62050H, Thermosense XXVIII, 17 April 2006 , ISBN 9780819462619 [3].

Examples of such company publications include:

• 1. Fluke Switzerland GmbH: "New Thermography Technology: Maintaining Maintenance Problems Faster", polyscope 18/06, p. 28-30 [1].
• Second Roger Schmidt of Fluke Thermography, "Benefits of IR / Visible Fusion", 2007 [2].
• Third K. Johnson, T. McManus and R. Schmidt, Infrared Solutions, Inc., "Commercial Fusion Camera", Proc. of SPIE Volume 6205 62050H, Thermosense XXVIII, 17 April 2006 . ISBN 9780819462619 [3].

• a) die Position des optischen Zentrums (engl. Nodal Point) im Raum. Hier gibt es drei Freiheitsgrade;
• b) die Richtung der optischen Achse, der Hoch- und der Querachse. Auch hier gibt es drei Freiheitsgrade;

Since the concept of the invention also starts from a device having at least two simultaneously active sensors, the relationships with respect to the geometrical optics are explained in more detail below:
Images are created by a perspective projection of the outside world on the image plane, also called central projection. Since the central projection is not an affine image, it is not possible to match both images by any linear operations such as stretching or shifting if the so-called external camera parameters of both sensors do not match. These parameters describe the 3D pose of a camera coordinate system relative to a world coordinate system. As external camera parameters, the expert calls this

• a) the position of the optical center (English nodal point) in space. There are three degrees of freedom here;
• b) the direction of the optical axis, the vertical axis and the transverse axis. Again, there are three degrees of freedom;

The equations of the perspective projection are:

Of the Index IR indicates that the thermal image sensor is meant. f stands for the focal length, here for the thermal image sensor.

The 1 clarifies these optical conditions. There is shown a camera model that illustrates the central projection.

Analogous to corresponding textbooks, for example Hanspeter A. Mallot, "Seeing and Processing Visual Information, An Introduction," Vieweg Publishing Company, 2nd Edition, January 2000 . ISBN 3-528-15659-7, p. 32-34 , is the thermal image sensor in the 1 for a simplification of the situation as Lochkameramodell executed.

As the 1 shows, the optical axis, which passes through the optical center N →, the image plane at a distance f _IR , namely in the image main H. Here falls the orthogonal sensor coordinate system of the thermal image sensor with the origin in N →, to simplify the situation, with the world coordinate system together, in which the point P → is specified. The point P → thus has the same coordinates (x, y, z) ^T in both coordinate systems. He is on the picture point P → IR imaged in the image plane of the thermal image sensor.

If the point P → was observed in another sensor coordinate system, its pixel would be displaced, with respect to P → IR distorted in perspective.

If, for example, an image sensor, indicated below by the index VIS, was displaced at a distance a in the x direction, only a scaling in the ratio of the focal lengths would result for the y components (equation 2b), whereas the x component would would depend on the Ent distance z of the point P → be transformed (equation 2a).

Further Transformations, such as a shift in the z direction or twists, are not considered here because they are in the subject of the application not necessarily occur.

Creating congruent images is thus, for all points P → IR in the thermal image, the irradiance E _IR measured there with the point calculated in accordance with Equations 2a, b P → VIS measured illuminance E _VIS to a two-dimensional pixel summarize.

As a visualization aid, the result of this operation can be seen that in an HSI color space, the brightness is determined by the value E _VIS and the color tone by the value E _IR , and that the color saturation stands for the mixing ratio of the two values. However, this pixel representation is, as already said, only illustrative. For a machine image analysis, it is not suitable.

Of the additive term in equation 2a is in the stereometry as disparity designated.

die Gleichungen 3a, b mit den dimensionslosen Pixelkoordinaten X_VIS, Y_VIS, X_IR, Y_IR und den auf die jeweiligen Pixelrastermaße normierten Brennweiten F_VIS und F_IR:

oder aufgelöst nach X_IR, Y_IR:

If the pixel pitch d _VIS given by the structure of the image sensor is introduced into the equations 2 a, b and the pixel spacing d _{IR is} correspondingly introduced for the thermal image sensor, this results in

Equations 3a, b with the dimensionless pixel coordinates X _VIS, Y _{VIS, IR} X, _IR Y and normalized to the respective pixel pitches focal lengths F F _VIS and _IR:

or resolved to X _IR , Y _IR :

The individual pixel coordinate components X _VIS , Y _VIS , X _IR , Y _IR are integer variables. Their value range depends on the number of pixels. For example, if the image sensor has the size 640 × 480 pixels, which corresponds to the so-called VGA resolution, then there are 640 × 480 pixels P _VIS (X _VIS , Y _VIS ) with {X _VIS | - 320, ..., 319} and {Y _VIS | - 240, ..., 239].

Out The transformation equations 3a and 4a can be considered as special cases derive two simple scenarios for congruent images, but each a special problem with itself bring.

1st special case:

The Objects are practically at an infinite distance.

Examples:

thermographic images from distant objects such as buildings, silos or tanks.

In In this case, the disparity goes to zero and the equations 3a, 4a are, just like the equations 3b, 4b, only one scaling the two coordinate systems.

This Special case occurs when the pixel resolution is no longer is sufficient to measure the disparity. It usually gets that half the pixel diameter as a limit. The resolution limit is determined by the sensor that has the smaller normalized focal length has, so because of the reasons already mentioned above usually the thermal image sensor.

Different It would be if this had an extremely long focal length. In civilian use, however, this does not occur in practice.

From equation 4a this yields: z ∞ ≥ 2 · a · F IR , (Equation 5)

It is clear to the skilled person that an intentional bringing about this particular case, that the reduction of such _∞, by reducing F _IR from practical reasons, it is not brought about, as this resolution generally deteriorated.

2nd special case:

All objects of interest are very flat in nature and are in a single plane called z = z ₀ , called "frontoparallel plane".

Examples:

Measurement the heat distribution on a populated printed circuit board; monitoring of operations on the ground by a ceiling-mounted Presence sensor.

Here, the central projection becomes a parallel projection. It must be determined "only" z ₀ and used in Equation 3a or 4a.

One possibility for the determination of z _{0 is} provided by the Abbe projection law, that is, the so-called lens equation known as such in the art, resolved according to the object distance z ₀ .

In Equation 6, b _{0 is} the image width to be set on the lens when focusing on an object at the distance z ₀ . The measurement of the angle of rotation of the focus ring is thus sufficient to conclude from b ₀ to the distance z ₀ .

As is already known in circles of amateur photographers, distant objects are sharply imaged regardless of their distance. Conversely, this means that the determination of z ₀ can not be applied with the help of b ₀ outside of the near range. The border of the near zone lies at half of the so-called hyperfocal distance z _h . See also: Oliver Jennrich, "A View of the Depth of Field", Version 0.2, 22nd November 1999 ,

der Blendendurchmesser des Infrarotobjektivs.In this equation, K _{IR is} the f-number and

the aperture diameter of the infrared lens.

Werden die Gleichungen 5 und 7 zusammengefasst, ist zu erkennen, dass der Bereich DIR·FIR < z < 2·a·FIR nicht durch die in [3] angegebene Lösung abgedeckt ist. Die Disparität fällt vielmehr abstandsabhängig von dem Wert

gemäß der Gleichung 8 nach Null ab.The disparity at the boundary of the short range results from this by inserting into equation 4a:

If the equations 5 and 7 are summarized, it can be seen that the range D IR · F IR <z <2 · a · F IR not covered by the solution given in [3]. The disparity falls rather distance dependent on the value

according to the equation 8 to zero.

task The present invention is based on an image acquisition unit of the type mentioned at the outset to technically improve them in such a way that that not only the transformations according to the Equations 3a, b or 4a, b is mechanically possible, but furthermore also machine-specific an application-specific image analysis with respect to that obtained with the image capture unit Images.

These The object is achieved by an image acquisition unit, which features in the characterizing Part of claim 1.

The Image capture unit according to the invention covers the at the outset described special case 2, so it is suitable for Frontoparallel arranged objects.

The arrangement has a distance sensor coplanar with the thermal image sensor and the image sensor for measuring the distance z ₀ .

A signal processing unit, which is preferably mounted together with the image or thermal imaging sensor and its control electronics on one or more electronic components, converts according to the equations 3a, b or 4a, b all their own pixel coordinates in the respective other sensor and determines the associated intensity value. This becomes a Pi with its own intensity value xelpaar summarized.

• a) Wird gemäß den Gleichungen 4a, b vom Sichtbild ausgegangen, bleibt dessen Auflösung bei der Bildung der Pixelpaare erhalten. Es wird einfach vielen benachbarten Pixeln eine gemeinsame thermische Bestrahlungsstärke zugewiesen. Dies ist sinnvoll, wenn sich die nachfolgende anwendungsspezifische Auswertung vor allem auf Details im sichtbaren Licht stützt, beispielsweise beim Lokalisieren von schlechten Kontaktstellen.
• b) Wird gemäß den Gleichungen 3a, b vom Wärmebild ausgegangen, reduziert sich die Anzahl der Pixel auf die des Wärmebildsensors. Dies ist beispielsweise dann sinnvoll, wenn es sich bei der nachfolgenden anwendungsspezifischen Auswertung um eine einfache Präsenzkontrolle handelt.

Depending on the choice of equations 3a, b or 4a, b, two different results result according to the size difference between the thermal image sensor and the image sensor (F _VIS > F _IR ) assumed in the problem description as state of the art:

• a) If, according to the equations 4a, b, the visual image is assumed, its resolution in the formation of the pixel pairs remains. Simply assign a common thermal irradiance to many neighboring pixels. This is useful if the subsequent application-specific evaluation relies primarily on details in the visible light, for example, when locating poor contact points.
• b) If the thermal image is used according to equations 3a, b, the number of pixels is reduced to that of the thermal image sensor. This is useful, for example, if the subsequent application-specific evaluation is a simple presence check.

With the choice of the transformation equation can thus the image resolution and thus the computational effort for a downstream application-specific Image processing unit to be determined.

• a) Im Gegensatz zum Stand den Technik gemäß der Firmenveröffentlichungen [1, 2, 3] gibt es keinen systematischen Restfehler (vgl. Gleichung 8).
• b) Es ist eine Unterscheidung zwischen reflektierter Strahlung und Eigenstrahlung eines Objekts möglich, wenn eine Korrespondenzbildung zumindest für einige Punkte zwischen Wärmebild und Sichtbild möglich ist. Beispielsweise lässt sich durch Auffinden einer gemeinsamen Kante und dann durch Einsetzen in die Gleichungen 3a, b oder 4a, b feststellen, ob der vom Abstandssensor gefundene Wert z₀ „richtig” ist.

The image capture unit according to the invention has the following advantages:

• a) In contrast to the state of the art according to the company publications [1, 2, 3], there is no systematic residual error (see equation 8).
• b) A distinction between reflected radiation and own radiation of an object is possible if a correspondence formation is possible, at least for some points between the thermal image and the visual image. For example, can be achieved by finding a common edge and then by substituting into equations 3a, b and 4a, b to determine whether the value found by the distance sensor z is ₀ "correctly".

If this is not the case, it is radiation originating from a body outside the plane given by z ₀ .

All in all can with the help of the invention Image acquisition unit not only the transformations according to the Equations 3a, b or 4a, b are performed by machine, but can also machine an application-specific image analysis with respect to the image acquisition unit Forming done.

advantageous Embodiments of the invention are the subject of dependent claims.

After that the signal processing unit is either a digital signal processor or preferably one programmed with the transformation equations Logic module (FPGA), which saves space, calculates quickly and cost-effectively are.

Of the Distance sensor may, for example, an ultrasonic distance sensor or one on the so-called Photomischdetektor technique (PMD technique) be based optical distance sensor, the readily to Available. But it can also optical Distance sensors according to other techniques used.

One PMD sensor, for example, is an optical sensor whose principle of operation based on the light transit time method.

The Dimensions of the pictures may be the same or different be.

The The above-mentioned pixel pair becomes an application-specific one Image processing unit are supplied, which in the figures but not shown in detail. This image processing unit can on the basis of the pictures known to her with the corresponding pixel pairs, namely that of the visual image and that of the thermal image, a user-specific evaluation of the objects in the machine Take pictures. This is possible because of a clear assignment from locations of one image to corresponding locations the other picture is possible, so the respect specific points arising from the respective pictures Information is linked and evaluated can.

following an embodiment of the invention with reference to a Drawing explained in more detail. Show:

1 a schematic representation of a camera model for illustrating the central projection according to the prior art, and

2 a schematic representation of an image capture unit according to the invention.

The 1 has already been discussed above. It will therefore not be discussed again at this point. Rather, reference is made to the earlier description at this point.

The 2 schematically shows the image capture unit according to the invention 1 for fusion with a first optical unit 2 for visual images and a second optical unit 3 images generated for thermal images.

The first optical unit 2 has a first focus device 4 and an image sensor 5 for visual images. The image sensor 5 again has a first optical axis 6 on.

The second optical unit 3 has a second focus device 7 and a thermal image sensor 8th for thermal images. The thermal image sensor 8th in turn has a second optical axis 9 on.

The wavelength sensitivity of the image sensor 5 is in a first wavelength range of about 0.4 microns to 1.2 microns and thus just in the visible wavelength range.

The wavelength sensitivity of the thermal image sensor 8th lies in a second wavelength range of about 8 microns to 12 microns and thus just in the non-visible infrared range.

With the image sensor 5 becomes a first picture with first rays 10 in the range of the first wavelength sensitivity and thus detected at a first wavelength.

With the thermal image sensor 8th becomes a second image with second rays 11 in the range of the second wavelength sensitivity and thus detected at a second wavelength.

With the first focus device 4 become the first rays 10 an image on the first image sensor 5 the first optical unit 2 mapped with a first optical resolution.

With the second focus device 7 become the second rays 11 an image on the second image sensor 8th the second optical unit 3 mapped with a second optical resolution.

from Principle can be the first and the second optical Resolution can be the same, but it can be the same the first or the second optical resolution opposite be smaller in each case the other optical resolution.

In the rule, however, is the resolution of the thermal image smaller than the resolution of the visual image.

The image capture unit 1 further includes a distance sensor 12 for measuring the distance z ₀ (see above). The distance sensor 12 has a measuring axis 13 , coplanar with the first optical axis 6 of the first image sensor 5 and coplanar with the second optical axis 9 of the second image sensor 8th is aligned.

The distance sensor 12 can perform its distance measurement according to different measurement techniques. He measures the distance z ₀ , which is described in more detail above.

In addition, the image capture unit points 1 a signal processing unit 14 on.

The signal processing unit 14 is in a first signal processing unit 15 for the first optical unit 2 and a second signal processing unit 16 for the second optical unit 3 divided up.

The first signal processing unit 15 is on a first circuit board 17 and the second signal processing unit 16 is on a second circuit board 18 placed.

The signal processing unit 14 is with the first and the second optical unit 2 . 3 connected in such a way that the signal processing unit 14 from the first and second optical units 2 . 3 receives data processes which can be processed by data technology (X _IR , Y _IR , X _VIS , Y _VIS ) for each pixel (x, y) of the image, and this starting from one of the two optical units 2 ; 3 each in corresponding signals (X _VIS , Y _VIS , X _IR , Y _IR ) of the other optical unit 3 ; 2 according to the relevant relations according to the equations 3a, b and 4a, b, respectively, by setting z = z ₀ as from the distance sensor 12 measured converted.

According to the present embodiment, the signal processing unit 14 realized by a digital signal processor. But it can also be realized by a programmed with the transformation equations logic block.

The signal processing unit 14 the image capture unit 1 an application-specific image processing unit is connected downstream of that of the signal processing unit 14 calculated data-technically processable signals (X _IR , Y _IR , X _VIS , Y _VIS ) per respective pixel of a respective image takes over and further processed application-specific.

QUOTES INCLUDE IN THE DESCRIPTION

This list The documents listed by the applicant have been automated generated and is solely for better information recorded by the reader. The list is not part of the German Patent or utility model application. The DPMA takes over no liability for any errors or omissions.

Cited patent literature

• EP 0973137 A1 [0007]

Cited non-patent literature

• - Fluke Switzerland GmbH: "New Thermography Technology: Maintenance Problems Faster On Track", polyscope 18/06, p. 28-30 [0015]
• - Roger Schmidt of Fluke Thermography, "Benefits of IR / Visible Fusion", 2007 [0015]
• K. Johnson, T. McManus and R. Schmidt, Infrared Solutions, Inc., Commercial Fusion Camera, Proc. of SPIE Volume 6205 62050H, Thermosense XXVIII, April 17, 2006 [0015]
• - ISBN 9780819462619 [0015]
• - A. Mallot, "Seeing and Processing Visual Information, An Introduction", Vieweg Publishing Company, 2nd Edition, January 2000 [0020]
• - ISBN 3-528-15659-7, p. 32-34 [0020]
• - Oliver Jennrich, "A View of the Depth of Field ", Version 0.2, November 22, 1999 [0043]

Claims (8)

Image capture unit ( 1 ) for fusion with sensors ( 5 ; 8th ) images of different wavelength sensitivity, comprising - a first optical unit ( 2 ) having a first wavelength sensitivity in the range from about 0.4 μm to 1.2 μm for acquiring a first image with first beams ( 10 ) in the range of the first wavelength sensitivity and thus with a first wavelength, and - a second optical unit ( 3 ) having a second wavelength sensitivity in the range of about 8 μm to 12 μm for acquiring a second image with second beams ( 11 ) in the range of the second wavelength sensitivity and thus with a second wavelength, - wherein the first optical unit ( 2 ) with a first optical axis ( 6 ) a first focus device ( 4 ) for imaging the first rays ( 10 ) on a first image sensor ( 5 ) of the first optical unit ( 2 ) with a first optical resolution (x1, y1) and the second optical unit ( 3 ) with a second optical axis ( 9 ) a second focus device ( 7 ) for imaging the second beams ( 11 ) on a second image sensor ( 8th ) of the second optical unit ( 3 ) with a second optical resolution (x2, y2), characterized in that coplanar with the first image sensor ( 5 ) and to the second image sensor ( 8th ) for measuring the distance z ₀ a distance sensor ( 12 ), and that a signal processing unit ( 14 ) is provided with the first and the second optical unit ( 2 ; 3 ) are connected in such a way that the signal processing unit ( 14 ) of the first and second optical units ( 2 ; 3 ) receives data processing-capable signals (X _IR , Y _IR , X _VIS , Y _VIS ) for each respective pixel (x, y) of the image and obtains them from one of the two optical units ( 2 ; 3 ) in each case in corresponding signals (X _VIS , Y _VIS , X _IR , Y _IR ) of the other optical unit ( 3 ; 2 ) according to the respective relations:

converted by setting z = z ₀ . Image acquisition unit according to claim 1, characterized in that the signal processing unit ( 14 ) is a digital signal processor or a logic device programmed with the transformation equations. Image acquisition unit according to claim 1 or 2, characterized in that the distance sensor ( 12 ) is an ultrasonic distance sensor or an optical distance sensor. Image acquisition unit according to claim 3, characterized in that the optical distance sensor ( 12 ) is based on the PMD measurement technology or another corresponding measurement technique. Imaging unit according to one of the preceding claims, characterized in that the dimensions (x; y) or the product (x · y) of the dimensions of the first image substantially equal to the dimensions (x; y) or the product (x. y) of the Dimensions of the second picture are. Imaging unit according to one of the preceding claims, characterized in that the dimensions (x; y) or the product (x · y) the dimensions of the first image larger as the dimensions (x; y) or the product (x x y) of the dimensions of the second picture. Imaging unit according to one of the preceding claims, characterized in that the dimensions (x; y) or the product (x · y) of the dimensions of the first image smaller than that Dimensions (x; y) or the product (x x y) of the dimensions of the second picture. Image acquisition unit according to one of the preceding claims, characterized in that the signal processing unit ( 14 ) an application-specific image processing unit is connected downstream of that of the signal processing unit ( 14 ) is calculated and processed in an application-specific manner (X _IR , Y _IR , X _VIS , Y _VIS ) per respective pixel of a respective image.