DE102017123971A1 - Device and method for object observation, in particular face recognition - Google Patents

Abstract

An apparatus for observing objects (1.7) and / or persons in an observation area (1.9) has at least one illumination means (1.1), which temporally successively displays the observation area (1.9) with at least two different wavelengths of light not visible to the human eye illuminate. At least one camera (1.5, 11.5) detects the reflections arising from objects (1.7) and / or persons in the observation area (1.9) as a result of the illumination according to the different wavelengths one after the other pixel by pixel. At least one camera (11.5) for daylight detects the reflections arising from objects (1.7) and / or persons in the observation area (1.9) as a result of illumination with light visible to the human eye on the object (1.7) and / or persons, one after the other pixel by pixel. The reflection values of each image illuminated with one of the different, invisible to the human eye and visible wavelengths image are stored pixel by pixel in image memories. A monitor serves to display the observation area (1.9) detected by the camera (1.5). In an evaluation unit, the reflection values, stored in the image memories, of the pixels of the wavelengths not visible to the human eye are superimposed with the reflection values of the pixels of the wavelengths visible to the human eye and an overlay image is generated Face recognition.

Description

Field of the invention

The invention relates to a device and a method for observing objects and / or persons, in particular at night, in an observation area according to the features of claims 1 and 9, respectively.

State of the art

For vulnerable places, such as Public places, video surveillance has proved to be a suitable means of preventing criminal acts. Video surveillance is problematic during the dark, especially when suspicious people have to be tracked across the coverage area of multiple cameras, moving in groups of other people.

For visualization of objects or persons in darkness, infrared cameras, residual light amplifiers or thermal imaging cameras are generally used.

Residual light cameras use, as the name suggests, the often existing weak light and amplify it accordingly. This only works if there is still low light. With low residual light noise these devices, so that a precise identification of people is hardly possible.

Thermal imaging cameras use the fact that every body emits heat radiation above 0 degrees Kelvin. It is recorded by the camera and can be displayed in a false color mode. Again, accurate facial recognition is difficult or impossible.

In one of the most widely used infrared camera systems, an infrared radiator illuminates an observation area captured by a camera sensitive to infrared radiation. The advantage of infrared radiation lies in the invisibility for the human eye. Therefore, this technology is preferably used for video surveillance of public places at night. Another advantage is the thereby prevented light pollution.

However, infrared cameras only provide a black and white image of the observation area at night. Suppose a person behaves conspicuously in the surveillance of a public place, then it is easy for this person to mix with other people and thus easily submerge. During the day it would not be so easy as the person would be e.g. can be more easily identified by the color of the clothing.

However, problems occur when an identification of persons is to take place. This is not only the case generally in public places, but e.g. also at ATMs, where so far a video surveillance with a conventional camera already takes place, the recordings of the ATM operates the person. This is even more true when e.g. a person wears a lifelike mask at skin temperature. This can not be unmasked with conventional cameras alone.

Presentation of the invention

The object of the invention is the identification of an object or a person, in particular for face recognition.

This is achieved with a device having the features of patent claim 1 and a method having the features of patent claim 9. Advantageous developments are the subject of the dependent claims. The features listed individually in the claims can be combined in a technologically meaningful manner and can be supplemented by explanatory facts from the description and by details from the figures, with further embodiments of the invention are shown.

With the device it is possible to observe and possibly identify objects and / or persons in an observation area. For this purpose, at least one illumination means illuminates the observation area with at least two different wavelengths of light, which are not visible to the human eye, in chronological succession. With at least one first camera, the reflections arising from objects and / or persons in the observation area as a result of the illumination are detected pixel by pixel in temporal succession according to the different wavelengths. At least one further camera for daylight is additionally provided, which records the reflections arising from objects and / or persons in the observation area as a result of illumination with light visible to the human eye on the object and / or persons, one after the other pixel by pixel. The respectively detected reflections are stored in several image memories.

With particular advantage, the device and the method for face recognition can be used. It is exploited here that the reflection values of faces at different infrared wavelengths in the non-visible region represent a signature of the face. These reflection values can be advantageous for Face recognition can be used. Each skin behaves differently and also the fat layers and the tissue under the skin form different infrared images. As a result, an increased security in the face recognition can be achieved, in particular if there is also a pixel-by-pixel offset with the other values for the face recognition. Preferably, one pixel of another daylight camera may be overlaid with correspondingly associated pixels of infrared light cameras to thereby more reliably achieve face recognition. This is expediently carried out each pixelwise.

In other words, therefore, e.g. In the case of the person standing in front of the ATM, not only the image in the sense of an "identification photo" such as a passport photo are used, but rather a signature of the face can be generated by the different reflection values of the skin, skin structure, lower tissue and fat layers, for additional identification the user can be used. Due to the different reflection properties, this results in a specific signature of the recorded object, for example the recorded face of a person, as it were an additional "passport picture" under the skin. Since the reflection behavior is recognized right under the skin, this also means that the face behind a mask can still be identified. The security against abuse can therefore be significantly increased.

The spectral range has several, preferably three frequencies in the region of the light which is not visible to the human eye. The infrared range and, in particular, the range of the near infrared has proved to be advantageous, but comparable results also result from the reflection values of an object, body or face irradiated with UV light.

By superimposing the pixels with different information content, it is possible to improve the face recognition of a person. With a correspondingly professional data processing, a face can be detected relatively quickly and reliably even in bulk.

Preferably, the sub-picture is stored in a database. If the person moves further so that the recognized marking is within a few seconds of other pixels, the current reflection values can be compared in a mathematical evaluation process with the stored superimposed reflection values. Are they essentially, z. B. to 95%, the same, the software recognizes this. The corresponding pixel on the monitor, in which superimposed information from daylight and invisible area of the light can be found, can be highlighted or marked accordingly, for example in strong color, with a brightness value and or by other visual highlighting. This facilitates tracking the movement of a particular person whose face has been recognized in a crowd. This allows a person or a face to be tracked within a group of objects or people.

By means of the database, a color and / or a brightness value and / or an otherwise optical highlighting, which is e.g. were generated from the images of the daylight camera. This makes sense, if e.g. a face has already been recognized on a day picture and is now to be tracked as a supplement at night.

Good results have in practice preferably then set when using a lighting means in the cycle alternately with different wavelengths both in the non-visible to the human eye wavelength range as well as in the visible to the human eye wavelength range of the light are made. This results within a fraction of a second a significant image of the relevant observation area, in particular a significant representation of a face of a person in the observation area, which can not be obtained alone in the visible range of light. In other words, by combining or superimposing images in the visible region with images in the non-visible region, an information content can be obtained which is particularly advantageous for facial recognition and possibly also for tracking a face.

A particularly advantageous solution results if the at least one illumination means emits the light with the different wavelengths. This allows a cheap solution to be implemented quickly and easily.

Just when the illumination means comprises a plurality of light emitting diodes, each of which emits at least one light having one of the different wavelengths, such detection of reflections originating from a face or a significant signature of an observed face can be implemented extremely favorably and in a small space.

It is advantageous if the wavelengths of light which are used according to the invention and which are not visible to the human eye are in the infrared range with regard to the apparatus and method, preferably in the near infrared range. However, wavelengths in the UV range can also be used to obtain corresponding reflection behavior and thus signatures of eg skin structure, lower tissue, fat layer and thus face.

In practice, light in a wavelength range from 800 nm to 1,400 nm, particularly preferably in a wavelength range from 800 nm to 1,000 nm, has proven to be easy and inexpensive to use for the device and the method, especially since corresponding bulbs are readily available for this range are.

• - Ausleuchten des Beobachtungsbereichs mit mindestens zwei verschiedenen für das menschliche Auge nicht sichtbaren Wellenlängen des Lichts zeitlich nacheinander,
• - Erfassen der von den Objekten und/oder Personen im Beobachtungsbereich infolge der Ausleuchtung entstehenden Reflexionen nach den verschiedenen Wellenlängen zeitlich hintereinander und pixelweise,
• - Erfassen der von den Objekten und/oder Personen im Beobachtungsbereich infolge einer Ausleuchtung mit für das menschliche Auge sichtbarem Licht entstehenden Reflexionen mit wenigstens einer weiteren Kamera für Tageslicht, die die entstehenden Reflexionen zeitlich hintereinander zusätzlich pixelweise erfasst,
• - pixelweises Abspeichern der Reflexionswerte je eines mit einer der verschiedenen Wellenlängen und mit für das menschliche Auge sichtbarem Licht beleuchteten Bildes in mehreren Bildspeichern,
• - Überlagern aus den in den Bildspeichern abgespeicherten Reflexionswerten der Pixel der für das menschliche Auge nicht sichtbaren Wellenlängen und den Reflexionswerten der Pixel der für das menschliche Auge sichtbaren Wellenlängen und Erzeugung eines Überlagerungsbildes zur Darstellung des erfassten Beobachtungsbereichs auf einem Monitor, insbesondere zur Gesichtserkennung einer oder mehrerer Personen im Beobachtungsbereich.

The object is also achieved by a method for observing objects and / or persons in an observation area, with the steps

• Illuminating the observation area with at least two different wavelengths of the light which are not visible to the human eye, one after the other,
• Detecting the reflections arising from the objects and / or persons in the observation area as a result of the illumination, according to the different wavelengths in temporal succession and pixel by pixel,
• Detecting the reflections arising from the objects and / or persons in the observation area as a result of illumination with light visible to the human eye with at least one further camera for daylight, which additionally captures the resultant reflections pixel by pixel in temporal succession,
• pixel-by-pixel storage of the reflection values of a respective image illuminated with one of the different wavelengths and with light visible to the human eye in a plurality of image memories,
• Superimposing on the reflection values, stored in the image memories, the pixels of the wavelengths not visible to the human eye and the reflection values of the pixels of the wavelengths visible to the human eye and generating an overlay image for displaying the detected observation area on a monitor, in particular for facial recognition of one or more People in the observation area.

With particular advantage, the method can be used for facial recognition. Hereby, it is exploited that the reflection values of faces at different wavelengths in the non-visible region, e.g. in the infrared or UV range, depict a signature of the face that can not be detected using only daylight or only a single wavelength of light not visible to the human eye. These reflection values can be advantageously used for face recognition. Each skin behaves differently and also the layers of fat and the tissue under the skin form different infrared images, which can be used to create a signature of a face. As a result, increased safety in face recognition, e.g. be achieved at ATMs, especially if there is a pixel-by-pixel offsetting with the other values for the face recognition. Preferably, a pixel of a daylight camera with correspondingly allocated pixels can be superimposed by cameras for wavelength ranges in the region of the light which is not visible to the human eye, thereby achieving face recognition with greater reliability. This is expediently carried out each pixelwise.

The different reflective properties result in a specific signature of the recorded object, for example the recorded face of a person, so to speak under the skin, that is to say a signature that carries each person with him.

Preferably, the sub-picture is stored in a database. If the person moves further so that after a few seconds he is in the area of another pixel, the current reflection values can be compared in a mathematical evaluation process with the stored superimposed reflection values. Are they essentially, z. B. to 95%, the same, the software recognizes this. The thus recognized pixel on the monitor, in which superimpose the information obtained from the daylight and from the non-visible region of the light, can be lifted out or marked accordingly, for example in strong color, with a brightness value and or by another optical highlighting. This facilitates tracking the movement of a particular person whose face has been recognized in a crowd.

In order to obtain a fast and good result even with an actual image sequence, at least one illumination means is operated alternately in the cycle with the different wavelengths.

For the specific purpose of a surveillance, in particular a face recognition, it is advantageous if an object can be marked on the monitor and the reflection values of this object are taken over into the database. Above all, this makes tracking easier across different monitoring areas and networks.

Likewise, it is advantageous if reflection values from a superordinate database can be adopted for the correct color representation at night.

A follow-up is particularly advantageously facilitated by the fact that several cameras are provided for detecting the objects and / or people in the observation area. As a result, an object detected with respect to the reflection values can be tracked by way of its reflection values over several camera locations.

For the invention, it is advantageous if a person and above all a face is successively illuminated serially with different light, on the one hand in the wavelength range of the invisible light and on the other hand in the visible to the human eye range of light. If this is not possible, because e.g. At a cash machine, the daylight is present anyway, so mutually or additionally a lighting can also be done with a light in a different wavelength and from a matched camera, the corresponding reflection behavior can be detected. The images taken under these conditions are then superimposed, with significant information e.g. to win over a person standing in front of the ATM. If this information is then compared with information already existing about the authorized user, additional authentication can take place. This can also ensure that a person with a lifelike mask is recognized as "spurious".

Further advantages will become apparent from the subclaims and the following description of preferred embodiments.

list of figures

• 1 eine schematische Darstellung eines Beobachtungsbereichs,
• 2 die pixelweise Bilderfassung.
• 3 eine schematische Darstellung des Verfahrensablaufs,
• 4a, 4b eine schematische Abbildung eines Gesichts, einmal aufgenommen im Tageslicht und einmal aufgenommen mit einer für das menschliche Auge nicht sichtbaren Wellenlänge des Lichts,
• 5 eine Darstellung der Vorgehensweise der Objekterkennung im Bildspeicher,
• 6 ein markiertes Monitorbild.

The invention will be explained in more detail below with reference to an embodiment shown in the figures. Show it:

• 1 a schematic representation of an observation area,
• 2 the pixel-wise image capture.
• 3 a schematic representation of the procedure,
• 4a . 4b a schematic illustration of a face taken once in daylight and once taken with a wavelength of light invisible to the human eye;
• 5 a representation of the procedure of the object recognition in the image memory,
• 6 a marked monitor picture.

Detailed description of preferred embodiments

The invention will now be described by way of example with reference to the accompanying drawings. However, the embodiments are only examples that are not intended to limit the inventive concept to a particular arrangement. Although the invention has been described in detail, it should be understood that it is not limited to the particular components of the apparatus and the particular process steps, as these components and methods may vary. The terms used herein are intended only to describe particular embodiments and are not intended to be limiting. In addition, if singular or indefinite articles are used in the specification or claims, this also applies to the majority of these elements unless the overall context clearly makes otherwise clear.

In general, illuminating means that emit light that is invisible to the human eye are used to illuminate a surface. In principle, however, light which is not visible to the human eye in another spectral range can also be used for the present invention since the inventive procedure is not influenced thereby. An illumination means generates a light. This may be e.g. also act to a single LED as lighting means.

• • nahes Infrarot (englisch: near infrared, NIR) ist kurzwellige IR-Strahlung, die sich direkt an den sichtbaren (roten) Bereich anschließt von 780 nm bis 1400 nm.
• • kurzwelliges Infrarot (englisch: short wavelength, SWIR) 1,4 bis 3,0 µm
• • mittleres Infrarot (englisch: mid wavelength, MWIR) mit Wellenlängen von 3,0 µm bis 8 µm.
• • langwelliges Infrarot (englisch: long wavelength, LWIR) 8 bis 15 µm
• • fernes Infrarot (englisch: far infrared, FIR) ist langwellige IR-Strahlung von 15 µm bis 1 mm und reicht in den Bereich der Terahertzstrahlung.

Preferably, non-visible light is used in the infrared range, although light in the UV range may be used as well. A subdivision for infrared light commonly used in Anglo-American space is:

• • Near Infrared (NIR) is short-wave IR radiation that directly adjoins the visible (red) area from 780 nm to 1400 nm.
• • short-wavelength infrared (English: short wavelength, SWIR) 1.4 to 3.0 μm
• • mid-infrared (MWIR) with wavelengths from 3.0 μm to 8 μm.
• • long-wave infrared (English: long wavelength, LWIR) 8 to 15 μm
• • Far Infrared (Far infrared, FIR) is long-wave IR radiation of 15 μm to 1 mm and extends into the range of terahertz radiation.

For the invention, the range of the near infrared has proven to be advantageous, since there are readily available commercially available LEDs for this area in the trade. The wavelength of this light is then, for example, in a range of 800 to 1,400 nm, preferably in the range of 800 to 1000 nm. However, it is also possible to use other regions of the light, for example in the UV range. The light of modern lighting means, whether in the visible to the human eye or in the infrared or UV range, is usually generated by light emitting diodes, which are often installed in larger quantities in such a lighting means. Now different objects, but also people, do not have the same reflection properties for the different wavelengths. In the white light spectrum we see an object colored. For example, from the spectrum of white light, the wavelength of the red light (-700 nm) is reflected, while all other wavelengths are swallowed. If this wavelength hits our eye, we see this as a red color.

However, each object has similar properties when viewed with a light spectrum in the non-visible light region, such as light. illuminated in the infrared light area this object or person. The wavelengths are reflected differently.

For facial recognition, it is exploited that the reflection values of faces at different infrared wavelengths in the invisible area represent a signature of the face virtually under the skin, as shown schematically in FIG 4a . 4b demonstrate. These reflection values can be advantageously used for face recognition. Each skin and / or skin structure behave differently and also form the fat layers and tissue under the skin different infrared images. As a result, an increased security in the face recognition can be achieved, in particular if there is also a pixel-by-pixel offset with the other values for the face recognition. Preferably, a pixel of a daylight camera with correspondingly assigned pixels can be superimposed by infrared light cameras, thereby achieving face detection with greater reliability. This is expediently carried out each pixelwise.

Due to the different reflection characteristics, a specific signature of the recorded object, for example the recorded face of a person, i. a signature under the skin, so to speak.

The different reflection properties also contain, in indirect form, information about the color originally visible to the human eye in the white light spectrum. Indirectly, not one wavelength is particularly strongly reflected, as was the case with the red color in the above-described example, but at e.g. three different wavelengths in the infrared range will set a certain reflection ratio to each other. In this reflection ratio is also the information about the actual reflective material such. also the skin of a human. Thus, at different infrared wavelengths, cotton will reflect in a different ratio, e.g. Synthetic or semi-synthetic clothing and this also applies to the individual whose skin, skin texture, lower tissue and fat layers are different from other people. The distinction not only by color, but also by the material or material mix is quite advantageous in the identification of a person.

Wavelengths used can be e.g. 820 nm, 870 nm and 950 nm. For these or similar wavelengths light emitting diodes are produced in large quantities.

If one now sequentially illuminates the object or the face with at least two, but better with three or more different wavelengths, e.g. of the infrared spectrum, depending on the wavelength, different reflections will be set on one and the same object. Since light-emitting diodes have a very fast turn-on and turn-off behavior, almost any number of light changes can take place in one second. This makes use of the method described here.

In the embodiment are in a lighting means 1.1 a number of LEDs 1.2 . 1.3 . 1.4 with a first wavelength of 820 nm ( 1.2 ), a number of light-emitting diodes with a second wavelength 950 nm ( 1.3 ) and a number of light emitting diodes with a third wavelength 870 nm ( 1.4 ), which can be switched on per wavelength together. For better identification of an object 1.7 can be used even more wavelengths, here in the embodiment, it is limited to three sequentially emitted wavelengths 1.6 that the object 1.7 illuminate. It is also possible to combine two or more wavelengths in each case in order to obtain more detailed information of the reflection properties of the object.

With a usual infrared camera as the first camera 1.5 becomes the light reflected back sequentially with different amplitude 1.8 receive. The light reflected at one wavelength, eg at 950 nm, becomes a reflection value 2.6 of the object 1.7 Pixel by pixel ( A 1.1 ) to (A ... n) in a first image memory A stored. This takes, for example, a period of 3 ms. Thereafter, the lighting device, ie the lighting means 1.1 , on another wavelength of eg 870 nm switched and it is again a frame in another image memory B stored. This is followed by a further switching, for example to the wavelength 820 nm, the individual image thus obtained is stored in a further image memory C stored. Of course, the number of image memories is not limited to three, when using other wavelengths or the parallel combination of the wavelengths to each other as well as when using information of a daylight camera as another camera 11.5 the number of image memories can be increased. The latter increases the evaluable information about the properties of the object 1.7 , When working with other wavelengths, only the first camera is used 1.5 adapting to it, ie when using UV light, must also be the assigned camera (this is then the camera corresponding to the infrared camera 1.5 .) to be able to take pictures in this spectral range.

Likewise, with at least one other camera 11.5 for daylight those of the objects and / or persons in the observation area 1.9 as a result of illumination with visible to the human eye light at the object and / or persons resulting reflections 11.8 recorded consecutively pixel by pixel. The first camera 1.5 and the other camera 11.5 can be integrated into a device. They are preferably time-delayed, ie serial.

3 shows the schematic sequence. A control unit 3.1 specifies the timing. When using a commercially available first camera 1.5 This can also synchronize the control unit for each frame, so that in this case no special camera must be used and the image analysis is limited only to the lighting unit and the evaluation software.

The reflection values 2.6 in the infrared region and in the visible to the human eye area as described in a number of image memory A . B . C . D stored. The reflection values per pixel of the wavelengths not visible to the human eye are superimposed on the reflection values of the pixels of the wavelengths visible to the human eye, thus producing a sub-picture image by means of the monitor control 3.4 on the monitor 3.5 can be represented.

The reflection values per pixel can be stored in a database 3.2 with one or many predefined stored values in an evaluation unit 3.3 compared and the result using the monitor control 3.4 eg as a marked, for example color-coded object to the monitor 3.5 output. When the calculation of the individual reflection values has been completed, the image memories become A . B . C . D refilled in another cycle. The reading of the data, that is certain reflection values, into the database 3.2 This can happen, for example, by having a person on the monitor 3.5 click on a suspicious person with the mouse pointer. Simplified in 3 by means of the controller 3.7 the database 3.2 by the observer at the monitor.

The reflection values of the various wavelengths emitted sequentially with the illumination unit whose reflection is specific to the fat layers in the face and optionally also to the clothing material and the color (visible in "normal" light) and the reflection values in the wavelength range visible to the human eye become thereby in the database 3.2 stored. By pixel-wise superimposition of the reflection values, an accurate facial recognition of the person can take place.

This is exemplified for the here described in particular face recognition on hand of 4a and 4b explained in more detail. 4a schematically shows a face, as in daylight by a daylight camera (another camera 11.5 ) is recorded. Clearly visible are the facial features we know with eyes, nose, mouth and ears. 4b now shows the same face, but taken with a first camera 1.5 which works with another wavelength of light invisible to the human eye. For this purpose, the face was illuminated with light of this wavelength so that a camera tuned to the reception of light of this wavelength could take this picture. It can be clearly seen that the facial features recede into the background, but instead fatty layers and cushions as well as structures of the skin and the lower tissue come to light. Performing the same recording for further wavelengths of light, other representations arise again, so that with a corresponding overlay a significant signature for each person is identifiable.

By data comparison and mathematical evaluation of the similarity of the reflection pattern, the suspect person 5.2 on the monitor picture 5.3 regardless of prevailing lighting conditions, eg in a bright color compared to the usual representation of all other persons 5.1 be marked or marked. 6 shows the resulting monitor image, due to the not possible in a patent application color representation unfortunately only symbolically.

The total reflection, which indeed decreases with the distance of the object, is preferably calculated out by a mathematical relation, that means only the differences of the Reflection values are included in the calculation. As a result, the distance between the person to be monitored and the camera is irrelevant. The symbolic numerical values shown in 5 represent the already adjusted reflection values.

5 clarifies the procedure of object recognition in image memory 4.1 (shown here by way of example only for the reflection values in the infrared range). In the database 3.2 are already the reflection values (vertical rows 4.5 4.6 and 4.7 ) for different materials (vertical row 4.2 ). These values can be determined individually or transferred from a higher-level database. The numerical values correspond to the reflection values. Thus, for example, in the image memory A (vertical row 4.2 ) For each stored object at the given wavelength a specific reflection value 4.9 stored. These reflection values are repeated in the rows of the further pixels A (1.2), B (1.2), etc., following to the right, so that all the reflection values for a specific material for the image memory A are the same.

The italicized numerical values 4.8 correspond to those through the first camera 1.5 determined current reflection values of the illuminated object. In the illustration in 5 Only the determined reflection values for a pixel A (1.1) B (1.1) C (1.1) are shown. Of course, the reflection values may be different for another pixel A (1.2) B (1.2) C (1.2), since the camera has detected another object, for example. The pixel-by-pixel information from a daylight camera is preferably treated analogously. Suppose an observer on the monitor has identified a suspicious person by superimposing the pixels from the infrared range and the visible range and, for example, clicks this person with the mouse pointer, then the current iteration-related reflection values determined by the camera are displayed 4.8 in the image memory 4.1 stored.

Suppose the person has moved, and after a few seconds, it is in the area of the pixel A (1.1), B (1.1), C (1.1). The current reflection values 4.8 be in a mathematical evaluation process with the stored reflection values 4.9 compared. Are they essentially, z. B. to 95%, the same, the software recognizes this (represented by the oval 4.3 in 5 ) and correspondingly highlights or highlights the corresponding pixel on the monitor, for example, in strong color, with a brightness value and / or by other visual highlighting or marking.

Depending on the application of the software, e.g. only the representation of the suspicious person or persons as optically lifted or marked object on the monitor e.g. are shown in color, while all other objects are shown only in shades of gray.

Of course, the possibility of object identification is not limited to just one object, it is also possible to display several objects in different shades or other optical highlights simultaneously on the monitor. Thus, motions of several persons, e.g. be better rated in groups of people. By superimposing the stored reflection values of the pixels of the wavelengths not visible to the human eye on the reflection values of the pixels of the wavelengths visible to the human eye, different persons can be easily distinguished. Also, the detection of suspicious persons does not necessarily have to be initiated "by hand"; intelligent algorithms can also do this.

If people are to be observed over several camera outlook points, can over a Kameranetzwerk 3.6 the data of the databases are exchanged or passed on.

LIST OF REFERENCE NUMBERS

1.1

lighting means

1.2

LED with wavelength 820 nm

1.3

LED with wavelength 950 nm

1.4

LED with wavelength 870 nm

1.5

first camera

1.6

Sequentially emitted different wavelengths

1.7

Reflecting object

1.8

Sequentially light reflected back with different amplitude

2.6

reflectance

3.1

control unit

3.2

Database

3.3

rating

3.4

monitor control

3.5

monitor

3.6

Camera Network

3.7

Control database by the observer

4.1

Image memory symbolic

4.2

Reflection value for materials in the database (vertical row)

4.3

essentially equal reflection values

4.4

by observers reflected values of a person to be tracked

4.5, 4.6, 4.7

Reflection values for different materials

4.8

current reflection values determined by the camera

4.9

specific stored reflection value

5.1

Persons on the monitor picture

5.2

marked (suspect) person on the monitor picture

5.3

monitor image

11.5

another camera for daylight

11.8

reflected back visible light

A

first image memory

B

image memory

C

third image memory

A (1.1)

first pixel of a number of n pixels of the image memory A

A (1.2)

second pixel of a number of n pixels of the image memory A

B (1.1)

first pixel of a number of n pixels of the image memory B

B (1.2)

second pixel of a number of n pixels of the image memory B

Claims (12)

Device for observing objects (1.7) and / or persons in an observation area (1.9), with at least one, preferably a plurality of illumination means (1.1), which illuminate the observation area (1.9) in chronological succession with at least two different wavelengths not visible to the human eye, at least one first camera (1.5) which captures, pixel by pixel, the reflections arising from the objects (1.7) and / or persons in the observation area (1.9) as a result of the illumination as a result of the illumination, - At least one further camera (11.5) for daylight, the time of the objects (1.7) and / or people in the observation area (1.9) as a result of illumination with visible to the human eye light on the object (1.7) and / or persons reflections recorded one after the other pixelwise and in addition to the first camera (1.5), a plurality of image memories (A, B, C,..., n) which store the reflection values of a respective image illuminated with one of the different wavelengths and with light visible to the human eye, a monitor (3.5) for displaying the observation area (1.9) detected by the at least one first (1.5) and the at least one further camera (11.5), an evaluation unit (3.3) which superimposes the reflection values, stored in the image memories (A, B, C, ..., n), of the pixels of the wavelengths not visible to the human eye with the reflection values of the pixels of the wavelengths visible to the human eye and generates an overlay image, in particular for the face recognition of one or more persons in the observation area (1.9). Device after Claim 1 , characterized in that a database (3.2) is provided in which certain, the different wavelengths associated reflection values are stored, and that the evaluation unit (3.3) stored in the image memories (A, B, C, ..., n) Reflection values of the pixels of the wavelengths not visible to the human eye are compared with the reflection values of the wavelengths invisible to the human eye in the database and the color and / or a brightness value and / or another pixel of the respective pixel for display on the monitor based on the comparison assigns optical emphasis. Device after Claim 1 or 2 , characterized in that a driver (1.10) for the illumination means (1.1) is provided, which operates the illumination means in the cycle alternately with the different wavelengths. Device according to one of the preceding claims, characterized in that at least one illumination means emits the light of the different wavelengths. Device after Claim 4 , characterized in that the illumination means (1.1) comprises a plurality of light-emitting diodes (1.2, 1.3, 1.4), of which at least one light emits at one of the different wavelengths. Device according to one of the preceding claims, characterized in that the wavelengths not visible to the human eye lie in the infrared range, preferably in the near infrared range. Method for observing objects (1.7) and / or persons in an observation area (1.9), comprising the steps Illuminating the observation area (1.9) with at least two different wavelengths, which are not visible to the human eye, one behind the other, Detecting the reflections arising from the objects (1.7) and / or persons in the observation area (1.9) as a result of the illumination according to the different wavelengths in temporal succession and pixel by pixel, Detecting the reflections arising from objects (1.7) and / or persons in the observation area (1.9) as a result of illumination with light visible to the human eye with at least one camera for daylight (11.5), which additionally captures the resultant reflections pixel by pixel in temporal succession, pixel-by-pixel storage of the reflection values of a respective image illuminated with one of the different wavelengths and with light visible to the human eye in a plurality of image memories (A, B, C,..., n), Superimposing on the reflection values, stored in the image memories (A, B, C, ..., n), the pixels of the wavelengths not visible to the human eye and the reflection values of the pixels of the wavelengths visible to the human eye and generating a sub-picture for display of the recorded observation area (1.9) on a monitor (3.5), in particular for the face recognition of one or more persons in the observation area (1.9). Method according to Claim 7 , characterized in that a database (3.2) is created in which certain, the different wavelengths associated reflection values are stored, and that the evaluation unit (3.3) stored in the image memories (A, B, C, ..., n) Reflection values of the pixels of the wavelengths not visible to the human eye are compared with the reflection values of the wavelengths invisible to the human eye in the database and the color and / or a brightness value and / or another pixel of the respective pixel for display on the monitor based on the comparison assigns optical emphasis. Method according to Claim 7 or 8th , characterized in that at least one illumination means (1.1) is operated in the cycle alternately with the different wavelengths. Method according to one of Claims 7 to 9 , characterized in that the non-visible to the human eye wavelengths are in the infrared range, preferably in the near infrared range. Method according to one of Claims 7 to 10 , characterized in that an object (1.7) can be marked on the monitor (3.5) and that the reflection values of this object (1.7) are taken over into the database (3.2). Method according to one of Claims 7 to 11 , characterized in that for detecting the objects (1.7) and / or persons in the observation area (1.9) a plurality of first (1.5) and / or further cameras (11.5) are provided and that an object detected with respect to the reflection values on the basis of the reflection values over several Camera locations is traceable away.

Images