DE102012223121A1 - Method for tracking at least one object and monitoring device - Google Patents

Abstract

The invention relates to a method for tracking at least one object (5) by means of at least one trackable imaging sensor, with which the at least one object (5) is detected, one or more comparison patterns (8.6), each of which has at least one object feature (9.6) of the have at least one object (5), are stored in chronological order in order to recognize the at least one object (5) in one or more images (4.6) subsequently acquired by the at least one imaging sensor on the basis of a match value to be determined of the at least one stored object feature (9.6) of the at least one object (5) to carry out at least one of the comparison patterns (8.6) with newly acquired object features (9.6), the at least one imaging sensor tracking the at least one object (5) and a course of movement of the at least one object (5 ) is saved, with the recognition of the weni at least one object (5), if several comparison patterns (8.6) are stored, for determining the match value at least one older comparison pattern (8.6) is taken into account at least to the extent that newer comparison patterns (8.6) of the at least one object (5) are taken into account and / or wherein when the at least one imaging sensor (2) is tracked, if the at least one object (5) is no longer recognized in an image (4.6) of the at least one imaging sensor (2) recorded subsequently, an estimate of the further movement of the at least one object (5) takes place, in which older stored movements of the at least one object (5) are taken into account at least to the extent as newer stored movements of the at least one object (5).

Description

The invention relates to a method for tracking at least one object by means of at least one trackable imaging sensor, with which the at least one object is detected. The invention also relates to a monitoring device with at least one imaging sensor.

"Tracking", in particular video tracking, means the tracking of objects by means of image processing by an automatic tracking of an imaging sensor or a camera system. The tracked object can, for. A user or a parent system. Here, a manually or automatically defined object is detected, determines its position and the detecting camera moves so that the object is shown as close to the center of the image and continue to remain there. The image sensor or the image acquisition system can be one or more cameras in the visual and / or thermal imaging field, for example long-wave infrared (LWIR), medium-wave infrared (MWIR), very long-wave infrared (VLWIR), far-infrared (FIR), short-wave Infrared (SWIR) and near infrared (NIR) as well as in the UV range. In order to be able to recognize the object, object representations in the image and / or object features or object properties are initially determined, stored as comparison patterns or so-called templates, and attempts in the subsequent images to retrieve objects having this appearance or these features. This can be accomplished, for example, by storing a picture section which is compared in the following pictures with image sections there, in particular by means of cross-correlation or the like, whereby the position with the maximum match is determined. This can also be done using other known object features, such. B. focus, border, phase, edge analysis or histograms (eg HoG / Histogram of Oriented Gradients).

However, all procedures must take into account the fact that objects can change their appearance and thus their characteristics or properties over time. This happens, for example, already by a rotation of the object (eg a vehicle), whereby its appearance changes significantly with respect to the camera system. For this purpose, it is known to make an adjustment of the stored object properties to the changed appearance by storing its newly detected features as changes in a still recognized object as startup characteristics for further searches in subsequent images. There is thus a kind of relearning of the tracker. However, this is problematic as soon as the properties of the object change because other objects are in the vicinity of it or at least partially obscure it. For example, if two different objects move towards each other or if the object to be tracked moves behind a stationary object, such as a tree, a building or a vehicle, then the object features the tracked object with the characteristics of the object encountered when the objects meet mixed. After the encounter, it can easily happen that the tracker pursues the wrong object. In existing tracker systems, this is a frequently observed negative effect.

Once an object to be tracked is no longer found in the image, for example due to rapid temporary obscuration, it must also be decided how the automatic camera tracking should continue. Known systems either stop or move on with the last set speed and direction. However, this can lead to a wrong direction of movement, in particular in the case of a previous diffuse movement of the tracked object, as a result of which the object area can be left completely.

The DE 103 02 306 A1 relates to an object tracking with a swivel camera.

The DE 10 2006 001 033 A1 discloses a method and apparatus for detecting objects in the driving environment for a motor vehicle.

The DE 11 2009 000 485 T5 generally shows a video surveillance and in particular the comparison of objects that are shown in multiple images.

The present invention has for its object to improve a method of the type mentioned above and to avoid the disadvantages of the prior art, in particular to be prevented on the one hand, that after encountering different objects, especially with a partial concealment, the wrong object is pursued On the other hand, given a loss of the object, for example by rapid temporary occlusion, an improved behavior in the tracking of the imaging sensor can be specified.

• – eines oder mehrere Vergleichsmuster, welche jeweils wenigstens ein Objektmerkmal des wenigstens einen Objekts aufweisen, in zeitlicher Reihenfolge gespeichert werden, um in einem oder mehreren nachfolgend von dem wenigstens einen bildgebenden Sensor erfassten Bildern eine Wiedererkennung des wenigstens einen Objekts anhand wenigstens eines zu bestimmenden Übereinstimmungswerts des wenigstens einen gespeicherten Objektmerkmals des wenigstens einen Objekts wenigstens eines der Vergleichsmuster mit neu erfassten Objektmerkmalen durchzuführen, wobei
• – der wenigstens eine bildgebende Sensor dem wenigstens einen Objekt nachgeführt wird und einen Bewegungsverlauf bzw. eine Trajektorie des wenigstens einen Objekts gespeichert wird, wobei
• – bei der Wiedererkennung des wenigstens einen Objekts, wenn mehrere Vergleichsmuster gespeichert sind, zur Bestimmung des wenigstens einen Übereinstimmungswerts wenigstens ein älteres Vergleichsmuster wenigstens in dem Umfang wie neuere Vergleichsmuster des wenigstens einen Objekts berücksichtigt wird und/oder wobei
• – bei der Nachführung des wenigstens einen bildgebenden Sensors, wenn das wenigstens eine Objekt in einem nachfolgend erfassten Bild des wenigstens einen bildgebenden Sensors nicht mehr wiedererkannt wird, ältere gespeicherte Bewegungen des wenigstens einen Objekts wenigstens in dem Umfang wie neuere gespeicherte Bewegungen des wenigstens einen Objekts berücksichtigt werden.

This object is achieved according to the invention by a method for tracking at least one object by means of at least one trackable imaging sensor with which the at least one object is detected, wherein:

• One or more comparison patterns, which each have at least one object feature of the at least one object, are stored in chronological order in order in one or more subsequent of the at least a recognition of the at least one object based on at least one to be determined matching value of the at least one stored object feature of the at least one object at least one of the comparison pattern with newly detected object features perform
• - The at least one imaging sensor is tracked the at least one object and a course of motion or a trajectory of the at least one object is stored, wherein
• Upon the recognition of the at least one object, if a plurality of comparison patterns are stored, at least one older comparison pattern is taken into account and / or wherein at least one older comparison pattern is determined at least to the extent that newer comparison patterns of the at least one object are used to determine the at least one match value
• In the tracking of the at least one imaging sensor, if the at least one object in a subsequently acquired image of the at least one imaging sensor is no longer recognized, older stored movements of the at least one object are taken into account at least to the extent of newer stored movements of the at least one object become.

The tracking method according to the invention achieves a considerable improvement in the object tracking in the case of at least partially obscured or lost object. In this case, the history is viewed in the case of the object features or the object representation and / or in the course of the movement. Older stored object features are considered at least as much as newer stored object features. This advantageously minimizes the risk that the tracker will erroneously shift to a new object. Here, the object features of a changing object are re-learned as before, but keep the previous object features, for example, in a list. Alternatively or additionally, the history of object movements in the case of temporary loss of object recognition can be taken into account in the control of the tracking of the imaging sensor. The object movements can be detected and / or stored, for example, as motion vectors. Thus, for example, a determined long-term direction of movement or, in the case of a wobbly movement, a stoppage for the tracking of the imaging sensor can also be selected. An older comparison pattern is stored earlier than a newer comparison pattern. Older and newer comparison samples are thus related to each other in terms of the chronological sequence of their storage.

Comparison patterns may include so-called object templates.

• – falls der zu bestimmende wenigstens eine Übereinstimmungswert größer oder gleich einem vorgebbaren ersten Grenzwert ist, eine sichere Wiedererkennung des wenigstens einen Objekts erfolgt ist, wobei kein neues Vergleichsmuster des wenigstens einen Objekts gespeichert wird, und/oder
• – falls der zu bestimmende wenigstens eine Übereinstimmungswert größer oder gleich einem vorgebbaren zweiten Grenzwert und kleiner als ein vorgebbarer erster Grenzwert ist, ein neues Vergleichsmuster unter Verwendung von neu erfassten Objektmerkmalen des wenigstens einen Objekts gespeichert wird, und/oder
• – falls der zu bestimmende wenigstens eine Übereinstimmungswert kleiner als ein inevorgebbarer zweiter Grenzwert ist, das wenigstens e Objekt nicht mehr wiedererkannt wurde.

According to the invention, it can further be provided that:

• If the at least one match value to be determined is greater than or equal to a predefinable first limit value, a secure recognition of the at least one object has taken place, no new comparison pattern of the at least one object being stored, and / or
• If the at least one match value to be determined is greater than or equal to a predefinable second limit value and smaller than a predefinable first limit value, a new comparison pattern is stored using newly acquired object features of the at least one object, and / or
• If the at least one matching value to be determined is smaller than a second limitable to be impeachable, the at least one object was no longer recognized.

The degree of agreement can z. B. be determined by a cross-correlation on the pixel level. The predefinable first limit may be at a match value of about 10% to about 100%, more preferably about 60% to about 90%, preferably about 88%. The predefinable second limit may be at a match value of about 10% to about 100%, more preferably about 60% to about 90%, preferably about 85%.

It is advantageous if, in the recognition of the at least one object for determining the at least one matching value, at least one older comparison pattern is used in preference to at least one newer comparison pattern of the at least one object.

For example, a minimum matchmaking unit may be specified using the oldest stored comparison pattern in which the at least one stored object feature has at least the predetermined minimum match score with the newly acquired object features. Starting with the oldest comparison pattern, the respective match value of the next stored comparison pattern can be determined one after the other in chronological order, after which it is checked whether it is greater than or equal to the predetermined minimum match value. By doing so, computational effort can be reduced. The minimum match value may correspond, for example, to the second threshold.

Alternatively, first the match values of the at least one stored object feature of all stored comparison pattern with the newly detected object features are determined. Subsequently, this match can be weighted by a factor, with older comparison patterns receiving a higher weighting. Thereafter, the stored comparison pattern is used, in which the at least one stored object feature has the highest weighted match value with the newly acquired object features.

As a result, older object features can be viewed further. When searching for the object position in the subsequent images, older object features are now preferred over newer ones. In this case, a corresponding weighting can also be used. Thus, if two objects were subsequently both separately visible again, even if the tracker were incorrectly relearned, the old object appearance would be prioritized prior to the encounter, and thus the correct object would be followed up.

It is advantageous if the time distribution of the storage of comparison patterns of the at least one object takes place uniformly or logarithmically.

Especially with a longer tracking of an object with the associated frequent changes, the number of to be processed comparison pattern or object features are getting higher, so the computational effort increases sharply. Therefore, additionally a selection of the object features to be stored can be made. In addition, the maximum number of stored object features or comparison patterns can be limited. The temporal distribution of the stored object features can be temporally uniform or logarithmic.

In a defined period unrecognized comparison pattern or object features can be removed or deleted. Comparison patterns or object features that were recognized and / or tracked at a predefined frequency over a defined, longer period of time can preferably be stored.

It is advantageous if the first comparison pattern or the first defined object features are stored permanently in order to prevent accidental frequent occlusions or encounters from causing undisturbed object features.

Object features of the at least one object may have appearance properties and / or representations of the at least one object.

The tracking of the at least one imaging sensor can take place in such a way that, taking into account the movement of the at least one object, it is displayed in each case in the region of the center of the image subsequently acquired. The capturing camera is moved in such a way that the recognized or recognized object is displayed as close as possible to the center of the image and remains there during further observation.

When tracking the at least one imaging sensor, an estimate or prediction of the movement of the at least one object can take place if the at least one object is no longer recognized in a subsequent image. Once an object to be tracked is no longer found in the image, for example by rapid temporary obscuration, it should be decided how the camera tracking should behave. In an advantageous manner, the history of the object or of the object movements can also be considered here, since each detected movement can have a certain inaccuracy, which is caused, for example, by a fluctuating movement of the object or its own position.

It can be determined whether the course of movement of the at least one object is diffuse or directed. By estimating the motion in a previously diffuse movement (eg, compensating for a wobbly movement), without considering the history or differentiation of the previous movement in "diffuse" or "directed", the previous object area would be left quickly.

During the tracking of the at least one imaging sensor, if the at least one object is no longer recognized in a subsequently acquired image, the current position of the at least one imaging sensor may remain unchanged upon detection of a diffuse motion pattern of the at least one object. If a diffuse motion is detected, the camera position can remain unchanged, as this is the best estimate for the further movement of the object.

It is advantageous if, during the tracking of the at least one imaging sensor, if the at least one object is no longer recognized in a subsequently acquired image, the further movement of the at least one object is detected by one, in particular, detection of a directed course of movement of the at least one object modified or normalized vector addition of a number of previous detected and / or estimated movements of the at least one object is estimated.

As a result, the history is also taken into account in the case of a directed course of motion, which leads to an increase in the estimation accuracy.

The length of the estimated further movement of the at least one object can be determined by averaging or normalizing the individual lengths of the preceding detected and / or estimated movements of the at least one object used in the vector addition.

It is advantageous if a search region in the acquired images of the at least one imaging sensor is extended or enlarged in step by step, if the at least one object is no longer recognized in the subsequent acquired images, for example after a defined period of time.

In addition, since all estimates are provided with some blurring, the search range can be increased or extended over time, since the deviation of the motion estimation from the actual behavior of the object with increasing time leads to a greater deviation of the expected position from the real position. In addition, as time progresses, so does the likelihood that the object will change its motion. After re-detecting an object and correcting the wrong or inaccurate camera position, the search area can then be reduced to its original size. If the lost object is not retrieved after an adjustable period of time (eg 15 seconds), automatic camera tracking can stop and the original area sizes are reset. The user or automatic object recognition can transfer a new object to the tracker.

Claim 16 specifies a monitoring device having at least one imaging sensor and at least one control device electrically connected to the at least one imaging sensor.

Advantageous embodiments and further developments of the invention are specified in the subclaims. The following is an exemplary embodiment of the invention described in principle with reference to the drawing.

Show it:

1 a schematic representation of a monitoring device according to the invention, which is adapted to carry out a method according to the invention;

2a to 2f simplified representations of an object tracking to explain the method according to the invention;

3 a simplified representation of a temporally logarithmically distributed storage of object features or comparison patterns;

4 a simplified representation to illustrate the method according to the invention in a diffuse movement pattern;

5 a simplified representation to illustrate the method according to the invention in a directed course of motion;

6 an illustration of a search area during object tracking; and

7 a representation of an extended search area for an object loss.

1 shows a monitoring device 1 , which is an imaging sensor 2 , For example, a camera or a thermal imaging device, and a control device 3 which, with the imaging sensor 2 electrically connected. The control device 3 can be configured as a software-parameterized recognition and control unit and also for digitization. The digitization can alternatively already in the imaging sensor 2 respectively. In addition, the monitoring device 1 an evaluation and / or display device 4 on. Especially with large surveillance areas up to the area of several kilometers, several imaging sensors can also be used 2 be provided, which record the image sequences. The imaging sensor 2 can be aligned, for example by means of a pan and tilt head on to be detected areas. In addition, the imaging sensor can 2 also be set up to carry out further, in particular linear movements.

In the 2a to 2f the operation of the method according to the invention is illustrated in more detail. 2a shows an object tracking of a detected object 5 that of the trackable imaging sensor 2 in a picture 4.1 is detected at a first time t ₀ . How out 2a visible, approaching the object 5 another object 6 , The object area is through a frame 7 indicated. In the 2a to 2f become comparison samples 8.1 - 8.6 with object features 9.1 - 9.6 stored at times t ₀ -t ₄ . As object feature 9.1 - 9.6 can each have a pixel-level object representation in the object area 7 be used. In addition, z. B. HoG, focus, border, phase position, edge viewing or the like as object features 9.1 - 9.6 suitable.

Like from the 2 B to 2d which show further times t ₁ , t ₂ and t ₃ conceals the other object 6 only partially, then completely the first object 5 , In the 2d becomes the object 5 almost completely from the other object 6 obscured, so that the object tracking a mixture between the two objects 5 . 6 detected. At the meeting of the objects 5 . 6 become the object features 9.1 of the tracked object 5 with the characteristics or characteristics of the encountering other object 6 mixed. How out 2e As can be seen, the wrong object, namely the other object, can accordingly come about 6 , is followed up.

The inventive method for tracking the at least one object 5 by means of the at least one trackable imaging sensor 2 with which the at least one object 5 This problem also avoids older features 9.1 - 9.6 be considered further. Here, as before, the object features 9.2 - 9.6 the changing object 5 newly learned, the previous object properties or object features 9.1 - 9.6 however, for example, kept in a list. Looking at the object area 7 in 2a , so this shows the taught-in object 5 at time t ₀ . The object areas 7 in the 2 B and 2c indicate at the times t ₁ and t ₂ a relearning with the object features 9.2 . 9.3 at. At time t ₃ takes place in 2d a relearning of the object features 9.4 to the oncoming other object 6 , Without viewing the object history according to the invention, at time t ₄ in FIG 2e the wrong object 6 followed, with a further relearning takes place. When searching for the object position in the subsequent images, older object features are used according to the invention 9.1 - 9.6 at least to the extent considered, such as newer features 9.1 - 9.6 or even older compared to newer features preferred. Would therefore at the meeting according to the 2a to 2f both objects 5 . 6 be again separately visible, the tracker would even when falsely relearning the old look of the object 5 prioritize the encounter and, as in 2f shown at time t ₄ , the correct object 5 pursue.

• – eines oder mehrere Vergleichsmuster 8.1–8.6, welche jeweils wenigstens ein Objektmerkmal 9.1–9.6 des wenigstens einen Objekts 5 aufweisen, in zeitlicher Reihenfolge gespeichert werden, um in einem oder mehreren nachfolgend von dem wenigstens einen bildgebenden Sensor 2 erfassten Bildern 4.1–4.6 eine Wiedererkennung des wenigstens einen Objekts 5 anhand wenigstens eines zu bestimmenden Übereinstimmungswerts des wenigstens einen tengespeicher Objektmerkmals 9.1–9.6 des wenigstens einen Objekts 5 wenigstens eines der Vergleichsmuster 8.1–8.6 mit neu erfassten Objektmerkmalen 9.1–9.6 durchzuführen, wobei
• – der wenigstens eine bildgebende Sensor 2 dem wenigstens einen Objekt 5 nachgeführt wird und ein Bewegungsverlauf 10.1, 10.2 (siehe 4 und 5) des wenigstens einen Objekts 5 gespeichert wird, wobei
• – bei der Wiedererkennung des wenigstens einen Objekts 5, wenn mehrere Vergleichsmuster 8.1–8.6 gespeichert sind, zur Bestimmung des wenigstens einen Übereinstimmungswerts wenigstens ein älteres Vergleichsmuster 8.1–8.6 wenigstens in dem Umfang wie neuere Vergleichsmuster 8.1–8.6 des wenigstens einen Objekts 5 berücksichtigt werden und/oder wobei
• – bei der Nachführung des wenigstens einen bildgebenden Sensors 2, wenn das wenigstens eine Objekt 5 in einem nachfolgend erfassten Bild 4.1–4.6 des wenigstens einen bildgebenden Sensors 2 nicht mehr wiedererkannt wird, eine Schätzung der weiteren Bewegung 11.1, 11.2inen des wenigstens e Objekts 5 erfolgt, bei welcher ältere gespeicherte Bewegungen 12.1, 12.2 des nenwenigstens ei Objekts 5 wenigstens in dem Umfang wie neuere gespeicherte Bewegungen 13.1, 13.2 des wenigstens einen Objekts 5 berücksichtigt werden.

The invention now provides a method for tracking the object 5 by means of the trackable imaging sensor 2 proposed with which the object 5 is detected, wherein:

• - one or more comparison samples 8.1 - 8.6 , which in each case at least one object feature 9.1 - 9.6 of the at least one object 5 stored in chronological order to be in one or more subsequent of the at least one imaging sensor 2 captured images 4.1 - 4.6 a recognition of the at least one object 5 on the basis of at least one match value of the at least one ten-stored object feature to be determined 9.1 - 9.6 of the at least one object 5 at least one of the comparison samples 8.1 - 8.6 with newly acquired object features 9.1 - 9.6 perform
• - The at least one imaging sensor 2 the at least one object 5 is tracked and a course of movement 10.1 . 10.2 (please refer 4 and 5 ) of the at least one object 5 is stored, where
• - At the recognition of the at least one object 5 if multiple comparison patterns 8.1 - 8.6 for determining the at least one matching value, at least one older comparison pattern is stored 8.1 - 8.6 at least to the extent of newer comparison samples 8.1 - 8.6 of the at least one object 5 be considered and / or where
• - When tracking the at least one imaging sensor 2 if that is at least one object 5 in a subsequently captured image 4.1 - 4.6 the at least one imaging sensor 2 is no longer recognized, an estimate of the further movement 11.1 . 11.2 inen of the at least e object 5 takes place at which older stored movements 12.1 . 12.2 of at least one object 5 at least to the extent of newer stored movements 13.1 . 13.2 of the at least one object 5 be taken into account.

For example, at least eight comparison patterns 8.1 - 8.6 with at least one object feature 9.1 - 9.6 of the at least one object 5 stored in chronological order.

• – falls der zu bestimmende Übereinstimmungswert größer oder gleich einem vorgebbaren zweiten Grenzwert, vorzugsweise etwa 85% und kleiner als der vorgebbare erste Grenzwert ist, wird ein neues Vergleichsmuster 8.1–8.6 unter Verwendung von neu erfassten Objektmerkmalen 9.1–9.6 des wenigstens einen Objekts 5, insbesondere zum Umlernen gespeichert, und/oder
• – falls der zu bestimmende Übereinstimmungswert kleiner als der vorgebbare zweite Grenzwert ist, wurde das wenigstens eine Objekt 5 nicht mehr wiedererkannt.

If the matching value to be determined is greater than or equal to a predefinable first limit value, preferably about 88%, then a secure recognition of the at least one object is ensured 5 done, with no new comparison pattern 8.1 - 8.6 of the at least one object 5 is stored, and / or

• - If the match value to be determined is greater than or equal to a predetermined second limit, preferably about 85% and less than the predetermined first limit, a new comparison pattern 8.1 - 8.6 using newly acquired object features 9.1 - 9.6 of the at least one object 5 , stored in particular for relearning, and / or
• If the matching value to be determined is smaller than the predefinable second limit value, the at least one object became 5 no longer recognized.

The degree of agreement can z. B. be determined by a cross-correlation on the pixel level.

When tracking the imaging sensor 2 it can be a variety of movements, such as pivoting movements or linear movements act.

When recognizing the at least one object 5 at least one older comparison pattern may be used to determine the match value 8.1 compared to at least one newer comparison sample 8.2 - 8.6 of the at least one object 5 preferably used.

For example, the oldest stored comparison pattern 8.1 - 8.6 be used to determine the match value at which the at least one stored object feature 9.1 - 9.6 at least one predetermined minimum match value with the newly acquired object features 9.1 - 9.6 of the at least one object 5 having. The minimum match value may correspond, for example, to the second threshold.

Alternatively, the match values of the at least one stored object feature may also be first 9.1 - 9.6 all stored comparison samples 8.1 - 8.6 with the newly acquired object features 9.1 - 9.6 be determined. Subsequently, this match can be weighted by a factor corresponding to older comparison patterns 8.1 - 8.6 receive a higher weighting. Thereafter, the stored comparison pattern 8.1 - 8.6 used in which the at least one stored object feature 9.1 - 9.6 the highest weighted match score with the newly captured object features 9.1 - 9.6 having.

The temporal distribution of the storage of comparison patterns 8.1 - 8.6 of the at least one object 5 can be uniform or logarithmic. In 3 is a time-logarithmic distribution of comparison patterns 8.1 - 8.6 Simplified by diamonds. The time t is plotted on the horizontal axis.

In a defined period of time no longer recognized comparison samples 8.1 - 8.6 can be deleted.

reference sample 8.1 - 8.6 which have been recognized and / or tracked at a predetermined frequency in a defined longer period of time may preferably be stored.

The first comparison sample 8.1 at time t ₀ can be stored permanently.

The object features 9.1 - 9.6 of the at least one object 5 may be appearance properties and / or representations of the at least one object 5 exhibit.

The tracking of the at least one imaging sensor 2 can be done such that, taking into account the movement or the course of movement 10.1 - 10.2 of the at least one object 5 , this in each case in the center of the image of the subsequently captured images 4.1 - 4.6 is pictured.

It is determined whether the trajectory or the course of motion 10.1 . 10.2 of the at least one object 5 diffused or directed. The in 4 illustrated movement history 10.1 is diffuse, whereas the in 5 illustrated movement history 10.2 is directed.

In tracking the at least one imaging sensor 2 can if that is at least one object 5 in a subsequently captured image 4.1 - 4.6 is no longer recognized when detecting a diffuse movement pattern 10.1 of the at least one object 5 with older stored movements 12.1 and newer stored movements 13.1 the current position of the at least one imaging sensor 2 remain unchanged. Sonach will be the current position 11.1 of the at least one object 5 as another movement of the object 5 suspected and not mistakenly a straightforward further movement 14 of the object 5 accepted.

The movements can be executed, for example, as motion vectors.

In tracking the at least one imaging sensor 2 can if that is at least one object 5 in a subsequently captured image 4.1 - 4.6 is no longer recognized, upon detection of a directed movement history 10.2 of the at least one object 5 the further movement 11.2 of the at least one object 5 by means of a, in particular modified or normalized, vector addition of a number of preceding detected and / or estimated motion vectors or movements 12.2 . 13.2 of the at least one object 5 to be appreciated. This can be older stored movements 12.2 and newer stored movements 13.2 of the at least one object 5 be evaluated the same. The length of the estimated further movement 11.2 of the at least one object 5 can be determined by averaging the individual lengths of the preceding detected and / or estimated movements used in the vector addition 12.2 . 13.2 of the at least one object 5 be determined. This averaging is necessary because the length otherwise results from the addition of all older movements 12.2 . 13.2 would result in what would be too big. The total length of the resulting vector could be approximately determined by the number of divided vectors are divided. For example, if the last ten previous movements 12.2 . 13.2 can be added, the total length of the resulting vector can be divided by 10. Alternatively, one tenth of the length of the individual vectors or movements could also be used 12.2 . 13.2 be added. However, the longer an estimate is required, the less accurate the estimated further movement will be 11.2 ,

Since all estimates are provided with a certain blurring, in addition, the search area 15 (please refer 6 and 7 ) are increased as time progresses, since the deviation of the estimate from the actual behavior of the object 5 with increasing time to a greater deviation of the expected from the real position of the at least one object 5 leads. Also, as time increases, so does the likelihood that the object will 5 his movement changes.

In 6 are an object area 7 and a search area 15 shown. An automatic adjustment of the search area 15 shows 7 , While the object area 7 maintains its unaltered size, becomes the search area 15 increased. The search area 15 in the captured images of the at least one imaging sensor 2 is progressively expanded as time progresses when the at least one object 5 in the pictures below 4.1 - 4.6 is no longer recognized. The camera tracking becomes a probable location of the object 5 customized. After rediscovered object 5 and correction of the eventual wrong camera position becomes the search area 15 then reduced back to its original size. Becomes the lost object 5 After an adjustable time (eg 15 seconds) not found again, the automatic camera tracking stops and the original area sizes are restored. The user or the automatic object recognition can give the tracker a new object 5 to hand over.

LIST OF REFERENCE NUMBERS

1

monitoring device

2

Imaging sensor

3

control device

4

Display / evaluation

4.1-4.6

Pictures of the recorded scene

5

tracked object

6

another object

7

Property area

8.1-8.6

reference sample

9.1-9.6

Property features

10.1, 10.2

movement patterns

11.1, 11.2

further appreciated movement

12.1, 12.2

older stored movements

13.1, 13.2

newer stored movements

14

straight-line movement

15

search area

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• DE 10302306 A1 [0005]
• DE 102006001033 A1 [0006]
• DE 112009000485 T5 [0007]

Claims (16)

Method for tracking at least one object ( 5 ) by means of at least one trackable imaging sensor ( 2 ), with which the at least one object ( 5 ), wherein: - one or more comparison samples ( 8.1 - 8.6 ), each of which has at least one object feature ( 9.1 - 9.6 ) of the at least one object ( 5 ) are stored in chronological order in order in one or more subsequent of the at least one imaging sensor ( 2 ) captured images ( 4.1 - 4.6 ) a recognition of the at least one object ( 5 ) based on at least one matching value of the at least one stored object feature ( 9.1 - 9.6 ) of the at least one object ( 5 ) at least one of the comparative samples ( 8.1 - 8.6 ) with newly acquired object features ( 9.1 - 9.6 ), and wherein - the at least one imaging sensor ( 2 ) the at least one object ( 5 ) and a movement course ( 10.1 . 10.2 ) of the at least one object ( 5 ), characterized in that - upon recognition of the at least one object ( 5 ), if several comparison samples ( 8.1 - 8.6 ) for determining the at least one match value at least one older comparison sample ( 8.1 - 8.6 ) at least to the extent that newer reference samples ( 8.1 - 8.6 ) of the at least one object ( 5 ) and / or that - during the tracking of the at least one imaging sensor ( 2 ), if the at least one object ( 5 ) in a subsequently captured image ( 4.1 - 4.6 ) of the at least one imaging sensor ( 2 ) is no longer recognized, an estimate of the further movement ( 11.1 . 11.2 ) of the at least one object ( 5 ), in which older stored movements ( 12.1 . 12.2 ) of the at least one object ( 5 ) at least to the extent that newer stored movements ( 13.1 . 13.2 ) of the at least one object ( 5 ). Method according to Claim 1, characterized in that: if the at least one match value to be determined is greater than or equal to a predefinable first limit value, reliable recognition of the at least one object ( 5 ), with no new comparative sample ( 8.1 - 8.6 ) of the at least one object ( 5 ), and / or - if the at least one match value to be determined is greater than or equal to a predefinable second limit value and smaller than a predefinable first limit value, a new comparison pattern ( 8.1 - 8.6 ) using newly acquired object features ( 9.1 - 9.6 ) of the at least one object ( 5 ) and / or - if the at least one match value to be determined is smaller than a predefinable second limit value, the at least one object ( 5 ) was not recognized again. Method according to claim 1 or 2, characterized in that upon the recognition of the at least one object ( 5 ) for determining the at least one match value at least one older comparison sample ( 8.1 . 8.2 ) compared to at least one recent comparison sample ( 8.3 - 8.6 ) of the at least one object ( 5 ) is preferably used. Method according to claim 1, 2 or 3, characterized in that the oldest stored comparison pattern ( 8.1 - 8.6 ) is used to determine the at least one match value in which the at least one stored object feature ( 9.1 - 9.6 ) at least one predetermined minimum match value with the newly acquired object features ( 9.1 - 9.6 ) of the at least one object ( 5 ) having. Method according to one of claims 1 to 4, characterized in that the temporal distribution of the storage of comparative samples ( 8.1 - 8.6 ) of the at least one object ( 5 ) takes place evenly or logarithmically. Method according to one of claims 1 to 5, characterized in that no longer recognized in a defined period of comparison samples ( 8.1 - 8.6 ) to be deleted. Method according to one of claims 1 to 6, characterized in that comparison samples ( 8.1 - 8.6 ), which were recognized and / or tracked at a predetermined frequency in a defined longer period of time, are preferably stored. Method according to one of claims 1 to 7, characterized in that the first comparison pattern ( 8.1 ) is stored permanently. Method according to one of claims 1 to 8, characterized in that the object features ( 9.1 - 9.6 ) of the at least one object ( 5 ) Appearance properties and / or representations of the at least one object ( 5 ) exhibit. Method according to one of claims 1 to 9, characterized in that the tracking of the at least one imaging sensor ( 2 ) such that taking into account the movement ( 10.1 . 10.2 ) of the at least one object ( 5 ) this each in the middle of the image center of the subsequently captured images ( 4.1 - 4.6 ) is pictured. Method according to one of claims 1 to 10, characterized in that it is determined whether the course of movement ( 10.1 . 10.2 ) of the at least one object ( 5 ) is diffused or directed. Method according to one of claims 1 to 11, characterized in that during the tracking of the at least one imaging sensor ( 2 ), if the at least one object ( 5 ) in a subsequently captured image ( 4.1 - 4.6 ) is no longer recognized when a diffuse movement pattern is detected ( 10.1 ) of the at least one object ( 5 ) the current position of the at least one imaging sensor ( 2 ) remains unchanged. Method according to one of claims 1 to 12, characterized in that during the tracking of the at least one imaging sensor ( 2 ), if the at least one object ( 5 ) in a subsequently captured image ( 4.1 - 4.6 ) is no longer recognized, upon detection of a directed course of motion ( 10.2 ) of the at least one object ( 5 ) the further movement ( 11.2 ) of the at least one object ( 5 ) by means of a vector addition of a number of preceding detected and / or estimated movements ( 12.2 . 13.2 ) of the at least one object ( 5 ) is estimated. Method according to claim 13, characterized in that the length of the estimated further movement ( 11.2 ) of the at least one object ( 5 ) by averaging the individual lengths of the preceding detected and / or estimated movements used in the vector addition ( 12.2 . 13.2 ) of the at least one object ( 5 ) is determined. Method according to one of claims 1 to 14, characterized in that a search area ( 15 ) in the captured images ( 4.1 - 4.6 ) of the at least one imaging sensor ( 2 ) is progressively expanded as time progresses when the at least one object ( 5 ) in the subsequently acquired images ( 4.1 - 4.6 ) is no longer recognized. Monitoring device ( 1 ) with at least one imaging sensor ( 2 ) and at least one with the at least one imaging sensor ( 2 ) electrically connected control device ( 3 ), which is adapted to carry out a method according to one of claims 1 to 15.

Images