FR3076039A1 - ADAPTIVE SYSTEM AND METHOD FOR AUTOMATICALLY TRACKING AT LEAST ONE TARGET IN AT LEAST ONE VIDEO STREAM - Google Patents

Abstract

This system comprises: at least one tracking device (141,...14n), receiving a video stream and dynamically configurable, designed for the detection and automatic tracking of at least one target by analyzing the video stream; a performance metric value calculator (20) based on a target tracking result (R1,...Rn) provided by the tracking device (141,...14n); a corrector (26) for the configuration parameter of the tracking device (141,...4n) as a function of the performance metric value; and a dynamic configurator (161,...16n) of the tracker (141,...14n) by applying the corrected configuration parameter. It further comprises an additional reference tracking device (14MSTR), receiving at least a portion of the video stream, and the calculator (20) calculates the performance metric value from a comparison, on the video stream portion, the target tracking result (R1,..., Rn) provided by the tracking device (141,..., 14n) and a reference tracking result (S1,..., Sn) provided by the additional reference tracking device (14MSTR).

Description

The present invention relates to an adaptive system for automatically monitoring at least one target in at least one video stream. It also relates to a corresponding method and computer program.

By “target” is meant a living object or being, human or animal, detectable and intended to be automatically tracked in the video stream or streams using known methods which can be implemented by micro programming or micro wiring in the device. which is itself a computer or electronic device. In other words, the target is the object or living being to be located in real time in space and time from a perception of a scene provided by one or more video cameras.

The fields of application include the detection and automatic tracking in real time of a target in a video stream, the detection and automatic tracking of a target in several video streams supplied in parallel by several cameras (functionality generally known as "real-time multi-target tracking"), but also the automatic detection and monitoring of several targets in several video streams provided in parallel by several cameras (functionality generally known as "multi-target multi-target time tracking" real ”).

Multi-camera multi-target tracking in particular is a complex operation requiring the implementation of sophisticated algorithms generating a large quantity of calculations. It requires the joint and simultaneous use of several electronic components such as central processing units (Central Processing Unit), electronic chips specialized in video decoding, graphics processors or graphics processing units GPU (from the English "Graphie Processing Unit"), etc. In addition, the algorithms implemented in the multi-target multi-camera tracking processing chain present a large number of configuration parameters influencing both the quality of the tracking and the solicitation of material resources.

In the field of video surveillance, it is also common to use a very large number of cameras: for example a few hundred cameras in closed public places up to several thousand or tens of thousands outdoors. Continuous optimization of the compromise between the quality of monitoring and the solicitation of material resources then becomes essential, failing which the calculation costs would quickly become prohibitive.

In addition, the dynamic sharing of material resources becomes a major issue in a context of their versatility. For example, at any time, priority can be given to monitoring a specific target, even if it means temporarily reducing the quality of other monitoring. To implement this flexibility, you must be able to continuously adapt the monitoring so that it operates according to a predetermined optimization aimed at reducing the demand on material resources while maintaining a certain level of quality. The invention thus applies more particularly to an adaptive automatic target tracking system in at least one video stream, comprising: - at least one tracking device, receiving said at least one dynamically configurable video stream, designed for detection and automatic monitoring of at least one target by analysis of said at least one video stream, a calculator of at least one performance metric value from a target tracking result provided by said at least one tracking device , - a corrector of at least one configuration parameter of said at least one monitoring device as a function of said at least one performance metric value, and - at least one dynamic configurator of said at least one monitoring device by application of said at least a corrected configuration parameter.

Such a system is for example envisaged in the article by Igual et al, entitled "Adaptive tracking algorithm to improve the use of computing resources", published in IET Computer Vison, volume 7, n ° 6, pages 415 to 424 (2013 ). It is designed to detect risk situations in terms of monitoring, that is to say situations in which the monitoring device (s) is or are likely to commit detection or monitoring errors. , to automatically adapt the calculation capacities (ie the solicitation of material resources) or change the tracking algorithm. For this purpose, it is capable of calculating at least one performance metric value from a target tracking result provided by the tracking device (s) and of using it to dynamically adapt its configuration parameters. .

However, the detection of risk situations is not a simple problem to solve and the way in which the article by Igual et al mentioned above is not fully satisfactory. It creates a lot of uncertainty about its reliability.

It may thus be desired to provide an adaptive automatic target tracking system in at least one video stream which makes it possible to overcome at least some of the above problems and constraints.

An adaptive automatic target tracking system is therefore proposed in at least one video stream, comprising: - at least one tracking device, receiving said at least one video stream and dynamically configurable, designed for automatic detection and tracking of at least one target by analysis of said at least one video stream, - a calculator of at least one performance metric value from a target tracking result provided by said at least one tracking device, - a corrector for at least one configuration parameter of said at least one monitoring device as a function of said at least one performance metric value, - at least one dynamic configurator of said at least one monitoring device by application of said at least one corrected configuration parameter, and - at least one additional reference tracking device, receiving at least a portion of said at least one video stream, the computer being more precisely designed to calculate said at least one performance metric value from a comparison, on said at least one portion of video stream, of the target tracking result provided by said at least one tracking device and of a reference tracking result provided by said at least one additional reference tracking device.

Thus, risk situations are detected in a much more reliable way and the optimization of the material resources which results largely compensates for the additional cost generated by the addition of the reference monitoring device (s).

Optionally, an adaptive automatic target tracking system according to the invention may also comprise at least one video capture device, connected in data transmission to said at least one tracking device and to said at least one additional tracking device of reference, for the respective supply of said at least one video stream and of said at least one portion of said at least one video stream.

An adaptive automatic target tracking method is also proposed in at least one video stream, comprising the following steps: - analysis of said at least one video stream, using at least one dynamically configurable tracking device designed for the automatic detection and monitoring of at least one target in said at least one video stream, - calculation of at least one performance metric value from a target tracking result provided by said at least one monitoring, - correction of at least one configuration parameter of said at least one monitoring device as a function of said at least one performance metric value, and - dynamic configuration of said at least one monitoring device by application of said at least one parameter of corrected configuration, the method further comprising providing at least a portion of said at least one video stream to at least one additional reference tracking device, and l e calculation of said at least one performance metric value being more precisely carried out from a comparison, on said at least one portion of video stream, of the target tracking result provided by said at least one tracking device and a reference tracking result provided by said at least one additional reference tracking device.

There is also a computer program downloadable from a communication network and / or recorded on a computer-readable medium and / or executable by a processor, comprising instructions for the execution of the steps of an adaptive automatic tracking method. target in at least one video stream according to the invention, when said program is executed on a computer. The invention will be better understood with the aid of the description which follows, given solely by way of example and made with reference to the appended drawings in which: - Figure 1 schematically represents the general structure of an adaptive system of automatic target tracking in at least one video stream, according to one embodiment of the invention, - Figure 2 illustrates the successive steps of an automatic target tracking process implemented by the system of Figure 1, - FIG. 3 schematically represents an example of a first simplified architecture for the system of FIG. 1, and FIG. 4 schematically represents an example of a second shared and decoupled architecture for the system of FIG. 1.

The adaptive automatic target tracking system of Figure 1 includes at least one video capture device for providing at least one video stream. In the nonlimiting example illustrated, it includes n, more precisely n video cameras 10i, ..., 10j, ..., 10n for the supply of n video streams. Each of them films a scene and transmits the corresponding video stream by computer connection.

It further comprises a collector 12 of the n video streams for their transmission to a primary stage of tracking devices designed for the detection and automatic tracking of at least one target by analysis of the video streams they receive respectively.

In the example illustrated, the primary stage comprises: - n monitoring devices 14u ..., 14 ,, 14n for the respective processing of the n video streams originating from the n video cameras 101; 10 ,, 10n, and transmitted by the collector 12, and - according to a first aspect of the present invention, an additional reference monitoring device 14MStr for the processing of respective portions of the n video streams from the n video cameras 101; ..., 10 ,, ..., 10n, these portions being selected and transmitted by the collector 12.

Each of the n monitoring devices 141; ..., 14i; ..., 14n is for example capable of simultaneously detecting and tracking several visible targets in real time in the video stream it receives. Such devices, hardware or software, are well known in the art and will not be detailed further. They are dynamically configurable using one or more parameters allowing them to influence their consumption of material resources, their performance and the quality of their results in terms of target tracking: for example a number of images to be processed or exploit per second in the video stream, an image resolution, a maximum number of targets to be detected and track in the video stream, etc. Dynamic configuration consists of being able to modify these configuration parameters during their execution. For this, they are respectively associated with n dynamic configurators 16i, ..., 16 ,, ..., 16n which are physically designed or software programmed to store, adjust and dynamically apply these configuration parameters.

According to a second aspect of the present invention, they are also respectively associated with n measurement devices 18 !, ..., 18 ,, ..., 18n which are materially designed or software programmed to measure one or more values representative of '' solicitation or exploitation of material resources by the n monitoring devices 14 !, ..., 14 ,, ..., 14n. These values may include an operating frequency of the CPU and GPU hardware units used, a rate of use of allocated memory space, a data input or output rate, a number of images processed or used per second in the video stream, a number of cores used in the CPU and GPU hardware units used, etc.

Note that the number of images processed or used per second in the video stream is a parameter which can be both configured (using the n dynamic configurators 16i, 16,, 16n) and measured (using n measuring devices 18i, 18 ,, 18n). In fact, it is not because a target consisting of a certain number of images to be processed or used per second is set using one of the configurators ^ !, ..., 16 ,,. .., 16n that it is necessarily reached. This objective can in particular be temporarily hampered by insufficient material resources and it is advantageous to be able to measure it.

Thus, by dynamically playing on the configuration parameters and also regularly measuring the demand on hardware resources, it really becomes possible to automatically optimize the processing performed by the n monitoring devices 141; ..., 14 ,, ... , 14n within the limits of available resources.

The video stream portions transmitted to the additional reference tracking device 14Mstr are successively selected by the collector 12 according to any predetermined algorithmic selection rule, for example according to a periodic scheduling of the n video cameras 10i, ..., 10 ,,. .., 10n. In the example of FIG. 1, n portions Pi, ..., P ,, ..., Pn are successively selected and extracted from the n video streams supplied by the n video cameras 10i, ..., 10 ,,. .., 10n. They are each advantageously of sufficiently short duration so as not to constrain the additional reference monitoring device 14Mstr to processing in real time. For example, each portion of video stream can have a duration of less than one tenth, or even one hundredth, of the time which is left to the additional reference monitoring device 14Mstr to carry out its analysis. Thus, unlike the n monitoring devices 14 !, ..., 14 ,, ..., 14n, the latter can have more efficient, remote hardware resources, ... both in memory and in computation capacities . It can therefore be considered that the detection and monitoring processing carried out by the additional reference monitoring device 14Mstr are optimal, or at least of higher quality and resource levels, in comparison with the n monitoring devices 14! .. ., 14 ,, ..., 14n. This is the reason why the results provided by this additional reference monitoring device 14Mstr can be considered as constituting a reference, without target detection or tracking error, to which the results provided by the n monitoring devices can be compared. followed 14i, ..., 14 ,, ..., 14n, according to the first aspect of the present invention.

The adaptive automatic target tracking system of FIG. 1 further comprises a metric calculator 20 materially designed or software programmed for the calculation of at least one performance metric value of each of the tracking devices 14u ..., 14, , 14n.

According to the first aspect of the present invention, the metric calculator 20 comprises a first hardware or software module 22 more precisely designed to calculate a first performance metric value from a comparison, on each of the n portions Pu ..., P ,, Pn of video stream, of tracking results of target (s) Ri, ..., R ,, ..., Rn respectively provided by the tracking devices 141 (..., 14 ,, ..., 14n on the n portions Pu ..., P ,, ..., Pn and of results of monitoring of target (s) S1; ..., Si, ..., Sn successively supplied by the additional monitoring device of reference 14Mstr on these same n portions Pu ..., P ,, ..., Pn. The first performance metric value can be expressed in the form of a ratio and be correlated to a number of interchanged targets or to a number of targets not identified by each of the n monitoring devices 14 ^ ..., 14 ,, ..., 14n when compared to the additional monitoring device of ref ence 14Mstr-

According to the second aspect of the present invention, the metric calculator 20 comprises a second hardware or software module 24 more precisely designed to calculate at least one second performance metric value from each of the representative values of solicitation of material resources provided by the n measuring devices 18i, ..., 18 ,, 18n. In accordance with the measures given in the example above, it can thus be a fraction of the maximum operating frequency of the CPU and GPU hardware units used, a rate of use of allocated memory space, a fraction of throughput maximum input or output of data, a result metric obtained in relation to a target set in terms of the number of images processed or used per second, a rate of use of the calculation cores used in the CPU and GPU hardware units requested in relation to a maximum allocated number, for each of the n tracking devices 14 ^ ..., 14 ,, ..., 14n.

The adaptive automatic target tracking system of FIG. 1 further comprises a corrector 26 materially designed or software programmed to recalculate the configuration parameters of the n tracking devices 141; ..., 14 ,, ..., 14n from the metric values provided by the metric calculator 20. In practice, this corrector 26 can have functions or systems of pre-established linear or non-linear equations, of a motor rules, an expert system, a neural network system, comparators, or any other artificial intelligence tool useful for updating the configuration parameters of each of the n monitoring devices 14 ,,. .., 14 ,, ..., 14n. It may in particular have a set of rules or laws which must be observed, such as a law according to which each tracking device 14 ,, ..., 14 ,, ..., 14n must process at least five images. per second, or that a law according to which each monitoring device 14 ,, ..., 14 ,, ..., 14n must not fail to detect more than 30% of the targets.

It can have an individual approach, focusing on the individual adjustment of each of the n monitoring devices 14 ,, ..., 14 ,, ..., 14n, but also a global approach, focusing on distributing equitably the material resources generally available according to the different needs of each. It provides its results to the n dynamic configurators 16 ,, ..., 16 ,, ..., 16n.

The additional reference tracking device 14MSTr, the metric calculator 20 and the corrector 26 form a hardware or software optimization platform 28 of the system of FIG. 1.

The n results of the n tracking devices 14i, ..., 14 ,, ..., 14n are supplied to a secondary stage 30 of the adaptive automatic target tracking system of FIG. 1. In a simple embodiment, this secondary stage 30 can be a system output stage, but it can also be a secondary stage for additional processing of the video streams.

The operation of the system detailed previously will now be described with reference to FIG. 2.

During a first step 100, the collector 12 transmits the n video streams supplied by the n video cameras 10i, ..., 10 ,, ..., 10n to the n monitoring devices 14 ,, ..., 14j , ..., 14n of the primary stage of the system for analysis. During this same step, the collector 12 transmits the n portions P ,, ..., Pi, ..., Pn of successive video streams to the additional reference tracking device 14MSTr for analysis also.

During a following step 102, each of the n monitoring devices 14 ,, ..., 14 ,, ..., 14n becomes aware of its current configuration using its dynamic configurator 16i, ..., 16 ,, ..., 16n, analyzes the video stream transmitted to it and provides its results on the secondary stage 30. It also provides its partial results Ri, ..., R ,, ..., Rn obtained on the portions Pi, ..., P ,, ..., Pn of video stream to the first module 22 of the metric calculator 20.

During a step 104 executed in parallel with step 102, the additional reference tracking device 14Mstr analyzes the n portions Pi, P ,, Pn of successive video streams which are transmitted to it and provides its results Si, S ,, Sn at the first module 22 of the metric calculator 20.

Following steps 102 and 104, the first module 22 of the metric calculator 20 calculates its aforementioned metric values and supplies them to the corrector 26 during a step 106.

During a step 108 executed in parallel with steps 102 and 104, each of the n measurement devices 18 ^ ..., 18 ,, ..., 18n supplies its measured values to the second module 24 of the metric calculator 20.

Following step 108, the second module 24 of the metric calculator 20 calculates its aforementioned metric values and supplies them to the corrector 26 during a step 110.

Following steps 106 and 110, the corrector 26 updates the configuration parameters of the n monitoring devices 14 ^ ..., 14 ,, ..., 14n as indicated previously during a step 112. It transmits the updated values in dynamic configurators 16i, ..., 16 ,, ..., 16n.

During a last step 114, each dynamic configurator 16i, ..., 16i, ..., 16n acts respectively on each tracking device 14i, ..., 14 ,, ..., 14n as indicated above.

Steps 100 to 114 are advantageously carried out continuously, during the entire operating time of the system.

A first example of a simple implementation of the system in FIG. 1 is illustrated in FIG. 3. This architecture has components whose control and cost make implementation easy.

A tracking server 40 forming, for example a conventional computer with an operating system such as Linux (registered trademark) and a network access card, comprises a set 42 of shared monitoring devices. This assembly 42 receives the video streams from the n video cameras 10i, ..., 10 ,, ..., 10n which can take the form of IP cameras (from the English "Internet Protocol") or conform to the Wi-Fi protocol standard transmitting their video streams over the network. Tracking devices are quite simply software components capable of executing several algorithms for detecting and tracking target (s), for example implemented in C ++ or Python (registered trademark) programming language with the help of the OpenCV software library ( trademark).

The optimization platform 28 is also implemented in a conventional computer with network access. The additional reference monitoring device 14Mstr, the metric calculator 20 and the corrector 26 which constitute it are also software components for example implemented in C ++ or Python (registered trademark) programming language with the help of the OpenCV software library. (trademark). In particular, the metric calculator 20 and the corrector 26 can be programmed according to state-of-the-art algorithms for calculating metrics and corrections based on simple equations tested on test data for regulation. In this example also, the additional reference monitoring device 14MSTr receives the video stream portions of the set 42 of shared monitoring devices.

A second example of a more complex but more flexible implementation of the system of FIG. 1 is illustrated in FIG. 4. This architecture allows pooling and decoupling of all of its components, so as to envisage an industrial implementation having the capacity to process a very large number of video cameras. In particular, the scale effect in this type of architecture makes it possible to produce a considerable improvement in system performance by benefiting from the multiplication of gains by video cameras and the use of inexpensive and shared calculation units for optimization by metric calculations and corrections.

A first component 50, fulfilling a function of distributor and communication bus, receives the video streams from the n video cameras 10i, ..., 10 ,, ..., 10n. It can be a message agent component (from the English "message broker") such as Apache ActiveMQ (registered trademark) or WebSphere (registered trademark), or an ESB company bus (from the 'Business Service Bus') such as Blueway (registered trademark) or Talend (registered trademark).

A second component 52, connected to the first component 50, has a scalable architecture of the laaS demand infrastructure type (from the English “Infrastructure as a Service”), making it possible to adapt to the demand for resources by dynamic allocation of computing units such as virtual machines. This second component 52 can be implemented in private laaS based on VMware components (registered trademark) or in public laaS based on services such as Amazon EC2 (registered trademark). It includes a monitoring module 54 to M monitoring devices 54u ..., 54m, a reference monitoring module 56 to N additional reference monitoring devices 56i, ..., 56N and a regulation module 58 to P regulators 58ι, 58P each comprising a metric calculator and a corrector. These three modules 54, 56 and 58 are deployed on a virtual machine.

Data exchanges between the first and second components 50 and 52 are carried out as follows: - transmission of video streams and corrected configuration parameters, from the first component 50 to the tracking module 54 of the second component 52, - transmission of video tracking results, from the tracking module 54 of the second component 52 to the first component 50, - transmission of the selected video stream portions, from the first component 50 to the reference tracking module 56 of the second component 52, - transmission of the results of reference tracking on the selected video stream portions, of the reference tracking module 56 of the second component 52 to the first component 50, - transmission of the video tracking results and of the reference tracking results, from the first component 50 to the regulation module 58 of the second component 52, - exchange of metric values between regulators of regulation module 58 of the two ith component 52 passing through the first component 50, - transmission of the corrected configuration parameters, from the regulation module 58 of the second component 52 to the first component 50, - transmission of the video monitoring results, from the first component 50 to a secondary stage or out.

The distribution of tasks within each of modules 54, 56 and 58 is decoupled according to the availability and resources of their components 54 !, ..., 54m, 56 !, ..., 56n and 58i , ..., 58P.

As already mentioned, the systems illustrated in FIGS. 1, 3 and 4 can be implemented in computer devices such as conventional computers comprising processors associated with memories for storing data files and computer programs. The monitoring and regulation devices (by metric calculation, correction and configuration) can be implemented in these computers in the form of computer programs or of different functions of the same computer program. These functions could also be at least partly micro programmed or micro wired in dedicated integrated circuits. Thus, as a variant, the computer devices implementing Figures 1, 3 and 4 could be replaced by electronic devices composed solely of digital circuits (without a computer program) for carrying out the same actions.

It clearly appears that an adaptive automatic target tracking system in at least one video stream such as one of those described above makes it possible to really optimize its processing automatically within the limits of its resources and to detect risk situations in a way significantly more reliable. The resulting optimization of material resources largely offsets the additional cost generated by the addition of the reference monitoring device (s) and measurement devices for the solicitation of material resources.

In particular, it becomes possible to minimize the occupancy rate of the equipment for the overall system including all of the monitoring devices and to maximize the quality of monitoring for each of them: in fact the chain of regulation by metric calculations and Corrections is advantageously implemented regularly, for example periodically, for each of the monitoring devices of the primary stage and makes it possible to correct their operation regularly according to objectives set in the system.

The reference reference tracking device (s) is or is not subject to real-time processing constraints because the frequency and latency of the control loops can vary without sudden deterioration in system performance. Thus, each additional reference monitoring device can be deployed on physical components shared with other system functionalities, in an architecture generally less expensive, reduced to the total cost per unit operation, than real-time oriented architectures.

The overall cost of the system depends on the complexity of the scenes, and the quality demanded of the monitoring devices. For a fixed monitoring quality, the following reasoning shows that the present invention reduces the expenditure of the system. A similar reasoning can be carried out according to which, for a given expense, the present invention improves the quality of monitoring.

Let us note: - ctreguiation-pour-i the quantity of material resources used for the regulation (calculation of metrics and correction) of the monitoring device 14 ,,

Ctsuivi-tracker-master-i> C ^ caclul-des-métriques-i> C ^ œrrecteur-i la C | U3ntity of material resources used for the evaluation by the additional reference monitoring device 14Mstr of the video stream portions of the camera 10 ,, of the resource consumption assessment of the monitoring device 14 ,, of the correction calculation to be applied to the monitoring device 14 „- ctunitaire-tracker-moy the average quantity of material resources requested by a monitoring device (average calculated over all the monitoring devices 14! to 14n), ctunitaire-tracker-maitre-moy the quantity of material resources requested for the evaluation by the additional reference monitoring device 14Mstr of the video stream portions of an average camera over a period and across all tracking devices, unit — tracker — avg — optimized the average amount of material resources required by a tracking device when the regulation described above is used (average calculated on all the monitoring devices 14! at 14n).

When the system benefits from the first and second aspects of the present invention:

and

For a simple implementation of the regulation:

When the system does not benefit from the first and second aspects of the present invention:

Hence a gain G:

The present invention therefore provides a performance gain when the average optimization for a monitoring device provided by the regulation is greater than the average expenditure induced by the additional reference monitoring device for a monitoring device:

Note also that the invention is not limited to the embodiments described above.

In particular, the embodiments described above combine the advantages of the first and second aspects of the present invention. But it should be noted that these two aspects are independent of each other. A system according to the first aspect of the present invention may not include the measurement devices and the calculations of second metric values. A system according to the second aspect of the present invention may not include the additional reference tracking device (s) and the calculations of first metric values.

It will more generally appear to those skilled in the art that various modifications can be made to the embodiments described above, in the light of the teaching which has just been disclosed to him. In the detailed presentation of the invention which has been made previously, the terms used should not be interpreted as limiting the invention to the embodiments set out in this description, but must be interpreted to include all the equivalents the prediction of which is within the reach of the skilled person by applying his general knowledge to the implementation of the teaching which has just been disclosed to him.

Claims (4)

1. Adaptive automatic target tracking system in at least one video stream, comprising: at least one tracking device (14i, 14n), receiving said at least one dynamically configurable video stream, designed for automatic detection and tracking of '' at least one target by analysis of said at least one video stream, a calculator (20) of at least one performance metric value from a target tracking result (Ri, ..., Rn) provided by said at least one monitoring device (14i, ..., 14n), a corrector (26) of at least one configuration parameter of said at least one monitoring device (14i, ..., 14n) as a function of said at at least one performance metric value, and at least one dynamic configurator (16i, ..., 16n) of said at least one monitoring device (14i, ..., 14n) by application of said at least one corrected configuration parameter, characterized in that it further comprises at least one additional reference monitoring device (14Mstr ), receiving at least a portion (Pi, ..., Pn) of said at least one video stream, and in that the computer (20) is more precisely designed to calculate said at least one performance metric value from a comparison, on said at least one portion of video stream, of the target tracking result (Ri, ..., Rn) provided by said at least one tracking device (14 !, ..., 14n) and of a reference monitoring result (Si, ..., Sn) supplied by said at least one additional reference monitoring device (14MStr) · 2. Adaptive automatic target tracking system according to claim 1, further comprising at least one video capture device (10i, ..., 10n), connected in data transmission to said at least one tracking device (14! ..., 14n) and to said at least one additional reference tracking device (14MSTr), for the respective supply of said at least one video stream and of said at least one portion (Pi, ..., Pn) of said at least a video stream. 3. Adaptive automatic target tracking method in at least one video stream, comprising the following steps: analysis (102) of said at least one video stream, using at least one tracking device (14i, ... , 14n), dynamically configurable, designed for the automatic detection and monitoring of at least one target in said at least one video stream, calculation (106) of at least one performance metric value from a monitoring result target (Ri, Rn) provided by said at least one tracking device (14 ^ 14n), correction (112) of at least one configuration parameter of said at least one tracking device (14 ^ 14n) as a function of said at least one performance metric value, and dynamic configuration (114) of said at least one monitoring device (14 !, ..., 14n) by application of said at least one corrected configuration parameter, characterized in that it comprises in addition to the supply (100) of at least a portion (Pt, ..., Pn) of said at least one video stream to at least one additional reference tracking device (14Mstr), and in that the calculation (106) of said at least one performance metric value is more precisely carried out from a comparison, on said at least one portion (Ρ1 (..., Pn) of video stream, of the target tracking result (Ri, ..., Rn) provided by said at least one tracking device (14i, ..., 14n) and of a reference monitoring result (Si, ..., Sn) supplied by said at least one additional reference monitoring device (14Mstr) · 4. Computer program downloadable from a communication network and / or recorded on a medium readable by computer and / or executable by a processor, characterized in that it comprises instructions for the execution of the steps of an adaptive method automatic target tracking in at least one video stream according to claim 3, when said program is executed on a computer.