DE102014216366A1 - Method for monitoring an area - Google Patents

Abstract

The method (S1-S7) serves to monitor a region and comprises that image data of at least part of the region are detected (S1), the image data are compared with at least one reference value (S2) and at least one upon reaching and / or exceeding the reference value first lighting device is activated (S4). A monitoring device has at least one camera for recording image data, at least one first illumination device for illuminating a region to be monitored with a first, visible light and an evaluation device coupled to the at least one camera and to the at least one first illumination device for carrying out the method (S1- S7). The invention is particularly applicable to monitoring and presence-dependent lighting of outdoor areas and confined spaces.

Description

The invention relates to a method for monitoring a region in which image data of at least part of the region are detected. The invention is particularly applicable to monitoring and presence-dependent lighting of outdoor areas and confined spaces.

At present, mainly passive IR-sensitive detectors (so-called PIR detectors) are used for presence detection in dark environments. Their simple signal detection usually only reacts differentially to object movements. In particular, the commonly used, based on pyroelectric effects PIR detectors ("pyrodetectors") can only react to changing IR radiation due to their operating principle.

In particular, differential charge signals of two separate pyroelements of a pyrodetector are tapped for object recognition by means of pyrodetectors and supplied to an integrated processing logic in order to derive there a switching signal for switching on a lighting device. In conjunction with a typical Fresnel zone optics, pyro-detectors can only be used as motion detectors and can not be used for the detection of static objects. However, such a restriction to a motion message is not sufficient for many lighting concepts.

Another disadvantage of pyrodetectors is their relatively large form factor due to the IR-compatible Fresnel optics, which usually leads to a low customer acceptance and due to their low angular resolution and range to a relatively high misdetection rate. In addition, PIR detectors often require active user gestures to enable the lighting device to be activated or remain activated.

Another type of well-known motion detection for presence detection in dark environments includes "active" motion detectors that transmit microwaves in the sub-GHz range or ultrasonic waves to scan their echoes for Doppler shifts. The active motion detectors have the disadvantage that they are often unsuitable for object recognition due to their coarse resolution.

It is the object of the present invention to overcome the disadvantages of the prior art at least partially and in particular to provide a simple and inexpensive feasible way to object detection even at first dark areas.

This object is achieved according to the features of the independent claims. Preferred embodiments are in particular the dependent claims.

• – Bilddaten des Überwachungsbereichs erfasst werden,
• – die Bilddaten mit mindestens einem Referenzwert verglichen werden und
• – bei Erreichen und/oder Überschreiten des Referenzwerts zumindest eine erste Beleuchtungseinrichtung aktiviert wird.

The object is achieved by a method for monitoring a range (also referred to below as "monitoring range" without restriction of generality), in which:

• - image data of the surveillance area are captured,
• - The image data are compared with at least one reference value and
• - When reaching and / or exceeding the reference value at least a first illumination device is activated.

The method allows even with a weak illumination of the surveillance area, a determination of minute movements. Entering the dark surveillance area e.g. by a person or similar can be safely recognized because of its movement. This procedure is particularly sensitive to differential changes between two temporally spaced or temporally spaced image data, in particular images.

The image data may e.g. be recorded by a camera.

The monitoring area may be an open or closed three-dimensional volume, e.g. an outdoor area or an interior.

It is an embodiment that the first illumination device is activated and the surveillance area is illuminated by the first illumination device with a first, visible light.

Alternatively or additionally, another area than the monitoring area may also be illuminated when at least one illumination device is reached and / or exceeded the reference value.

The exceeding may, in particular, include an exceeding in relation to absolute values. If signed values are considered, the exceedance can also be below.

The first light is emitted at a first wavelength or at a first wavelength range. The first light may have an infrared light component. The first light can be white light.

It is still an embodiment that the area is illuminated by the first illumination device and / or by a second illumination device with a second light.

In a further development, therefore, the area of the first illumination device can be illuminated with the first light and with the second light.

In another development, the area can be illuminated by the first illumination device with the first light and by the second illumination device with a second light.

By means of the illumination with the second light, a sensitivity of the movement detection can be carried out safely even in very dark surveillance areas, in particular also by means of a camera. In this case, lighting with the second light with high luminous flux may be advantageous in order to achieve a high resolution even in the dark.

For this purpose, the reference value for detecting the object can in particular be determined or set adaptively to the basic illumination of the surveillance area by the second light and / or a prevailing background residual light (for example, a twilight light).

The first illumination device and / or the second illumination device may be separate from the camera or form a unit with the camera.

It is a further development that the image data are acquired upon illumination of at least part of the area, in particular with respect to the illumination by the second light.

Generally, comparing the image data with the at least one reference value may include determining a deviation of the image data from the reference value.

It is also an embodiment that a mean brightness value of the image data is compared with a reference value. This allows a particularly simple implementation of the method with high sensitivity. The average brightness value can be determined by averaging the brightness over one or more, in particular over all, pixels of a recorded image of the monitored area.

It is a further embodiment that the second light has or is infrared light, in particular in a wavelength range between 780 nm and 900 nm. The second light may in particular have a wavelength of about 820 nm. This results in the advantage that a human eye does not perceive the illumination by the second light or only very weakly, and thus monitoring of movements proceeds inconspicuously.

However, the second light may also have a visible portion to the human eye, in particular consist only of visible light. The second light may e.g. corresponding to the first light and e.g. be generated by the same lighting device. This development is particularly easy to implement and, for example, be used in the event that a repeated at short intervals visible illumination of the surveillance area does not bother.

It is yet another embodiment that the second light is emitted with temporal interruptions, in particular pulsating. As a result, energy consumption for generating the second light can be kept low. An image data acquisition rate may e.g. in the range of seconds, e.g. be about one second. A time of recording may be e.g. 10 milliseconds. The illumination with the second light may then be carried out in particular every second for at least 10 milliseconds.

It is also an embodiment that the image data are repeatedly recorded intermittently, in particular in the event that the reference value has not been reached and / or exceeded (so-called "dark mode"), the interruptions being in the range of seconds. So a movement can be detected quickly enough.

It is also an embodiment that an object recognition is performed with these image data and the first illumination device is deactivated if at least one predetermined object has not been detected by means of the object recognition. So can for certain Use cases relevant objects (eg persons) of non-relevant objects (eg animals or moving non-living objects) are distinguished.

It is also an embodiment that after reaching or exceeding the reference value and after the illumination of the surveillance area with the first light, image data is received from the surveillance area. With this image data, an object recognition can be carried out and, depending on a result of the object recognition, namely if at least one predetermined object has not been detected by means of the object recognition, the first light can be switched off again. Thus, for certain applications, relevant objects (e.g., persons) may be distinguished from non-relevant objects (e.g., animals or moving non-living objects).

It is also an embodiment that a background adjustment is performed by means of the image data of the area illuminated by the first light. Thus, an object recognition can be carried out with particularly high accuracy.

The background adjustment (in particular for adaptation, for example, to slowly changing illumination conditions, for example, by sidelights or variable natural daylight) may be e.g. by a remeasurement of the area illumination, in particular by a remeasurement of characteristic image histogram data, in particular and by a comparison with previously stored background image data, are determined.

For example, the brightness of the measured image may be adjusted or adjusted so that a certain number of the brightest pixels, e.g. of 50 brightest pixels, at an upper gain limit, e.g. at an upper saturation limit.

Alternatively or additionally, the image background can be adjusted or kept constant by calculation and / or by adjusting the camera sensitivity (for example via so-called "gain" and / or "exposure time" settings or a setting of a light measurement integration time).

It is also an embodiment that an object recognition is performed with these image data and the first illumination device is deactivated if at least one predetermined object has not been detected by means of the object recognition. Thus, for certain applications, relevant objects (e.g., persons) may be distinguished from non-relevant objects (e.g., animals or moving non-living objects).

It is also an embodiment that the background matching is repeated with a time constant in the range of, in particular, minutes, e.g. every two minutes, three minutes, etc. Thus, it can be prevented that a very little moving person or the like. is considered to belong to the background and is thereby included in the static background.

It is a further development that the time constant for the background adaptation is in the range of the expected exposure fluctuation (eg due to clouds) or lies only in the region of the latency for recognizing an object (eg depending on a used object recognition algorithm) or in the range of the provided time until Switch off the first light with negative object detection is.

The object recognition can be performed with a same time constant (e.g., an image acquisition rate) as the background determination, that is, in the range of one or more minutes. The time constant is then considerably longer than in the dark mode. However, the time constant of the object recognition may also differ from the time constant of the background determination and may be shorter (e.g., in the seconds range) or longer.

It is a development that the object recognition is carried out continuously or with only (short) interruptions, but their result is considered only when the background adjustment causes only a small change or the background barely changes or no longer changes. This waiting for a background which is only slight to practically non-changing may, for example, be in the range of seconds or one or more minutes.

The object is also achieved by a monitoring device, comprising at least one camera for recording image data, at least one first illumination device for illuminating a region to be monitored with a first, visible light and an evaluation device coupled to the at least one camera and to the at least one first illumination device for carrying out the method according to one of the preceding claims.

The camera can take color or black-and-white or gray-scale images. It can have a CMOS sensor. The camera may be equipped with optics, especially miniature optics.

It is an embodiment that the monitoring device has a second illumination device coupled to the evaluation device for illuminating the area to be monitored with a second light. This second illumination device may in particular illuminate infrared light ("IR light"), in particular in the near infrared ("NIR") region.

The above-described characteristics, features, and advantages of this invention, as well as the manner in which they are achieved, will become clearer and more clearly understood in connection with the following schematic description of an embodiment which will be described in connection with the drawings. In this case, the same or equivalent elements may be provided with the same reference numerals for clarity.

The figure shows a possible flowchart of the method.

In a first method step S1, image data of at least a part of an area to be monitored is detected by means of a camera, for example from a limited space or an outside area. The image data may be associated with pixels of a captured image.

The camera may have a CMOS sensor, for example with an image resolution of 640 × 480 or more pixels.

An effective image resolution or resolution from the CMOS sensor may be e.g. be set by binning settings in the CMOS sensor. In particular, the resolution for achieving fast image data processing, taking into account the requirement of a present application, may be kept low.

For example, in dark mode, a so-called "binning number" can be increased, so that the resolution is reduced so much that the mean value calculations (eg with respect to the mean value Amean described below) for a given latency (eg one second) as fast as possible and often can perform in a row.

For the image data evaluation for object recognition in the illuminated case or bright mode, the effective image data resolution can be increased so far that the algorithm obtains sufficient image information about the monitoring area.

High resolution is also advantageous for the dark mode with fast microprocessors, for example when a structure recognition algorithm is used to evaluate the noisy dark mode data with which the appearance of an object in the image data can be detected. For example, an entropy value calculation of one or more images or frames can be carried out for this purpose. A strong noise without structure then corresponds to a high entropy value. When a structure (especially an object) appears in the noise, the entropy value decreases.

The camera may be equipped with optics, especially a miniature optics.

The camera is e.g. sensitive in a spectral range from about 400 nm to about 900 nm. In particular, the camera is also sensitive to a portion of the near infrared spectrum, e.g. from about 780 nm to about 900 nm, ie, for example, for IR light with a wavelength of about 820 nm. The area to be monitored may therefore be monitored by means of the camera in a spectral range which is not for the human eye or barely visible. Thus, monitoring can be performed even if the area to be monitored appears dark to a human observer.

For illuminating at least a portion of the area to be monitored, without this attracting a human observer, this can optionally be illuminated at least temporarily with a second light in the form of an IR light from the near infrared spectrum ("NIR light") the camera is sensitive, eg by means of an NIR illumination device with one or more NIR light sources, which thus serves as the second illumination device. In the dark mode, the NIR illumination device can in particular illuminate the area to be monitored in such a way that detection of a movement of an object by the camera is reliably possible. In particular, monitoring of an otherwise non-illuminated area, e.g. a cellar, possible.

The camera is operated in a so-called "dark mode" in a dimly lit room in which it is comparatively fast (for example approximately every second) image data, e.g. a picture, takes. A typical image data acquisition duration may e.g. 10 milliseconds.

If an NIR illumination device is used, it is advantageous for energy-saving operation that it illuminates the monitoring area at least until the image data acquisition can be performed, e.g. every second for about 10 milliseconds. In particular, the illumination may be pulsating, i. always for a given period of time. For a NIR illumination device with a pulse power of one watt, a mean continuous power of only ten milliwatts can be achieved accordingly. For this purpose, the NIR illumination device can be controlled by the camera. Controlled by the camera, the NIR illumination device can be activated only briefly and in good time for and for image data acquisition.

In a second step S2, the image data are compared with at least one reference value.

In this case, in particular a mean brightness value Amean of the image recorded by the camera can be compared with a predetermined reference value Aref. The mean brightness value Amean corresponds, for example, to an averaging over a recording or over a predetermined amount of recordings, which are within a relatively short time Time interval, eg 1 second, to be made. The predefined reference value Aref is, for example, a value adapted over time (eg moving average value), which is adapted over a significantly longer time interval than is the case for the determination of the brightness value Amean. For example, the reference value Aref can be adapted over a period of minutes or hours, so that, for example, brightness changes due to daylight can be compensated accordingly by the comparison with the adaptive reference value Aref.

For example, if the area to be monitored is dark, then the average brightness value Amean is low and does not differ significantly or not significantly from a background noise. The dark mode can therefore correspond to a state in which the image captured by the camera has an average brightness value Amean that differs only slightly from the reference value Aref. If, for example, the mean brightness value Amean is below the reference value Aref, it is assumed that no object moves into the monitored area.

If an object moves into the monitored area, even at low brightness, the average brightness value Amean of the recorded or measured image changes (increases or decreases). In the present case, it can be assumed, for example, that an object moving into the region changes the brightness value Amean so strongly that a deviation from the reference value Aref is large enough to exceed a certain threshold value "thresh", ie abs (Amean - aref)> thresh applies, where "abs" is an absolute value. This possibility of determining the reaching and / or exceeding of the reference value Aref has the advantage that the noise effects can be suppressed by the threshold value thresh, which is advantageous in particular in the case where the values of Amean and Aref are close to one another. Threshold thresh may also take into account hysteresis, etc.

However, other possibilities are included for determining the achievement and / or exceeding of the reference value Aref, for example by comparisons Amean> = Aref or abs (Amean)> = abs (Aref).

The approaches described above are particularly sensitive to differential changes between two or more temporally successive recorded images, and allow detection of even small movements in low-lighted surveillance areas.

Entering dark surveillance areas e.g. by a person or similar can be recognized so because of their movement.

As already mentioned above, the reference value Aref can be smoothed in the presence-free dark or dark mode and thus in particular automatically adapted to the low-illumination situation prevailing in the dark mode (for example in the case of twilight).

The image analysis, e.g. the determination of the average brightness value Amean and its comparison with the predetermined reference value Aref can be carried out, for example, by means of an image evaluation device integrated into or coupled to the camera.

To compare the image data with the at least one reference value, it is possible to query whether the brightness value Amean deviates sufficiently from the reference value Aref, e.g. reaches and / or exceeds the reference value Aref.

If this is not the case ("N"), steps S1 and S2 are executed again.

If, on the other hand, the brightness value Amean has reached and / or exceeded a sufficiently large deviation from the reference value Aref ("J"), a known background image of the illuminated monitoring area is loaded and installed in a step S3.

The underlying background image has been recorded shortly before the last light shutdown. It may be especially the last picture before the last light shutdown. In one variant, the monitoring device then no longer reacts to differential changes between two or more temporally successively recorded image data sets, in particular images.

After step S3, in a step S4, a first, visible light, for example white light, is switched on. For this purpose, a (first) ambient lighting device with at least one visible light emitting light source, eg one or more LEDs, are activated. The ambient lighting device can be controlled by the camera or by the image evaluation device. When the visible light is on, the procedure is in a "bright mode".

In a step S5, image data is again taken in the light mode by the camera, and e.g. evaluated by the image evaluation device to detect at least one object in the monitored area.

In one variant, the camera can capture a newly appearing image background of an image now captured in bright mode. The static image background may be determined separately from previously detected moving objects, so that newly added persons or objects can be detected with high accuracy against this background, in particular spatially defined or "segmented".

Possibly slowly changing lighting conditions by side light or variable natural daylight can generally be determined by measuring the area lighting, in particular by retiming the image histogram data.

In general, the background image can be adjusted or kept constant (background adjustment) either by calculation or by adjusting the camera sensitivity (for example via so-called "gain" and / or "exposure time" settings).

In addition, in the bright mode, the image processing system has sufficient time to classify a segmented structure or shape of a recognized person with an image recognition algorithm at a time constant in the minute range.

If the object detection in step S5 yields a positive result (e.g., that an object has been detected), the background determination time constant may be increased or kept high to make further undisturbed observations.

Alternatively or additionally, the increase of the time constant of the background adaptation for stabilizing the image can also be initiated or carried out before object recognition. Thus, for example, the background image may already be performed for the first time after step S3 after loading the last background image or after a first acquisition and image evaluation in the light mode in step S5.

In this case ("J"), step S5 is repeated as long as the object recognition provides a positive result, that is, in particular, a predetermined object has been detected. The visible light or the ambient lighting device remains switched on in particular permanently.

In a negative result ("N"), in which the detected object is not a given object (for example, no object or other than a given object, eg a dog or a cat), in a following step S6, before the light is turned off, captured an image of the currently unoccupied monitoring area and stored as a background reference image.

In a further step S7, the visible light can be switched off again, the method then returning to step S1. For this purpose, the camera or the image evaluation device can issue a switching signal for switching off the ambient lighting device.

The first, visible light can also be switched off in step S7 if a person executes a gesture recognizable to the camera or the image evaluation device, which leads to switching off. In this case, step S1 may be carried out again after a predetermined delay time.

While the invention has been further illustrated and described in detail by the illustrated embodiment, the invention is not so limited and other variations can be derived therefrom by those skilled in the art without departing from the scope of the invention.

Thus, the second light may also be light that has a proportion in a visible to the human eye spectrum or even lies only in the visible spectrum. The second light may then correspond to the first light.

Generally, "on", "an", etc. may be taken to mean a singular or a plurality, in particular in the sense of "at least one" or "one or more" etc., unless this is explicitly excluded, e.g. by the expression "exactly one", etc.

Also, a number may include exactly the specified number as well as a usual tolerance range, as long as this is not explicitly excluded.

reference numeral

• S1

  Step of image data acquisition in dark mode

  S2

  Step of image data comparison in dark mode

  S3

  Step of loading a well-known background image

  S4

  Step of turning on a first light

  S5

  Step of image data acquisition and image data processing in bright mode

  S6

  Step of a storage of a background picture

  S7

  Step of turning off the first light

Claims (13)

Method (S1-S7) for monitoring an area in which Image data of the area are acquired (S1), - The image data are compared with at least one reference value (S2) and - When reaching and / or exceeding the reference value at least a first illumination device is activated (S4). Method (S1-S7) according to claim 1, wherein the first illumination device is activated and the area of the first illumination device is illuminated with a first, visible light. Method (S1-S7) according to one of the preceding claims, in which an average brightness value of the image data is determined and in which the mean brightness value is compared with a brightness reference value (S2). Method (S1-S7) according to one of the preceding claims, in which an entropy value of the image data is determined and in which the entropy value is compared with an entropy reference value. Method (S1-S7) according to one of the preceding claims, in which the area is illuminated by the first illumination device and / or by a second illumination device by a second light. The method (S1-S7) according to claim 5, wherein the second light has infrared light in a wavelength range between 780 nm and 900 nm. Method (S1-S7) according to one of claims 5 to 6, in which the second light is emitted with temporal interruptions, in particular pulsating. Method (S1-S7) according to one of the preceding claims, in which the image data are recorded repeatedly with interrupts, the interruptions being in the range of seconds (S1). Method (S1-S7) according to one of the preceding claims, in which - With this image data object recognition is performed (S5) and - The first illumination device is deactivated if at least one predetermined object has not been detected by the object recognition (S7). Method (S1-S7) according to one of the preceding claims, in which a background adaptation (S3) is carried out by means of the image data of the area illuminated by the first light. Method (S1-S7) according to Claim 10, in which the background adaptation is carried out by calculation and / or by adjusting a camera sensitivity. Monitoring device comprising At least one camera for taking image data, At least one first illumination device for illuminating a region to be monitored with a first, visible light and - An evaluation device coupled to the at least one camera and to the at least one first illumination device for carrying out the method (S1-S7) according to one of the preceding claims. Monitoring device according to claim 12, further comprising a second illumination device coupled to the evaluation device for illuminating the area to be monitored with a second light.

Images