FR2942333A1 - Monitoring system integration and modularization simplifying method for e.g. petrochemical complex installation, involves sending heterogeneous physical data to corresponding processing modules of monitoring system - Google Patents

Abstract

The method involves defining a common and single repository for storing heterogeneous physical data e.g. pressure, transiting in a monitoring system and for storing associated metadata. A translation is provided to a metadata dictionary for constructing a bijective correspondence in a representation understandable by different information processing elements for video samples having digitized physical parameters or for context information element captured by sensors. The physical data are sent to corresponding processing modules (3) of the monitoring system. An independent claim is also included for a monitoring system comprising a processing module.

Description

The present invention relates to a method for simplifying the integration and modularization of surveillance systems involving heterogeneous physical data, these data being used to simplify the integration and modularization of surveillance systems using heterogeneous physical data. being, for example, video data or provided by sensors. This type of monitoring, involving complex industrial processes for example, which is one of the priority areas of the present invention, is often called supervision. The supervision of large and complex systems (supervision of a city, an industrial site, a transport network, ...) calls for more and more advanced processing and visualization devices, which are 'support software and hardware solutions, themselves in constant progress. Their interfaces differ from one provider to another, and even from one model to another. As for the supervised systems, they are regularly supplemented by extensions (new district, new workshop, additional station, ...). As a result, each system differs from similar installations, which leads to complex configuration management, high maintenance operating costs and difficulties in implementing any change. This situation can become particularly inextricable if several applications have to deal with the same physical information, themselves relating to mobiles or devices with rapidly varying characteristics, and if it is necessary to present on the same display device these data and the results. treatments respecting temporal correlation and location on a map. The typical example is that of an industrial installation for which one will want to follow in real time an incident by displaying on the same map map geolocated physical measurements (electrical voltage, engine speed, pressure, ...), videos of fixed cameras close to the incident site, the video transmitted by the intervention vehicle and the residual volume and pressure calculated in real time for each tank, the calculated safety zones, ....

It is also obvious for the system suppliers that it is possible to reuse subassemblies already developed for another system for a new system, to be able to test these subsystems in isolation and to ensure that they are in good condition. integration as long as the interface requirements have been met. In order to be able to apply modularity to architectures that have become fully computerized and networked, we have been led to imagine new levels of abstraction and to develop design and interface standards that allow us to simply connect each feature to a single network. architecture provided for this purpose. A recent example of these techniques is Service Oriented Architecture (commonly referred to as SOA). Such an architecture takes into account the management of the applications, the sequencing of the actions and the exchanges between applications, but does not concern with the way the physical data (pressure, flow, video, power, sound, ...) that feed the system are received and processed by the system. If the sources are multiple, dispersed and varied, it is in particular extremely complex to precisely locate the source of information, to date it with certainty after it has passed through extensive networks and has undergone processing. To avoid creating errors or inconsistencies when presenting to users (for example on a basemap) or in the correlation processes, which is likely to induce completely erroneous interpretations of situations, costly and specific interfaces and adaptations are then indispensable. The only known beginning of the answer to this difficulty concerns a very different field, that of the audiovisual production which was very early faced with the difficulties induced by the correlation of multiple video and audio sources dealing with a single scene and processed by processing systems. complex information. The audiovisual community has developed standards for the production of content, constantly associating with the audio or video signal the data characterizing its creation or capture (instant, location, cameraman, intellectual property, etc.). These data, known as metadata (or metadata in English), are kept throughout the process of processing these audiovisual contents with the appropriate repetition or refresh rates. Typically, thanks to these metadata, and despite different paths or different treatments, the shots can be precisely localized for their editing, the sound and movement of the lips can be reproduced synchronously, the intellectual rights respected, ... The SMPTE (Society of Motion Pictures and Television Engineers) has developed several levels of standards providing: A dictionary of metadata taken into account, which provides in particular the model of these magnitudes (typically the address on the street, the number of the camera, its GPS coordinates, its angle of sight, ...), the unit of measurement to use, the number of significant digits, ... This dictionary, which already has a wide variety of terms, It is easy to extend to new applications classified in sections that can be public or private.

The way to dynamically encode this metadata, to link them and, where appropriate, to compress them to limit the volume of data to be taken into account The means of transmitting on various media, material or not, these so-called metadata-rich contents.

The subject of the present invention is a method for simplifying the integration and modularization of surveillance systems using physical data, a process making it possible to reduce the costs of maintaining these systems under operational conditions while allowing them to be easily changed. , as well as accurately locate on a map the source of information and date it with certainty. This method must also make it possible to reuse for a new system subsets already developed for another system, to be able to test these subsets in isolation and to ensure proper integration. The invention also relates to a monitoring system implementing such a method. The method according to the invention is a method of simplifying the integration and modularization of surveillance systems involving heterogeneous physical data, and is characterized in that it comprises the following steps: metadata dictionary compatible with the nature of the data passing through the system, we attach to each data a label including at least the time of creation or modification of this data, we define a single and common repository for each type of data transiting in the system and for the associated metadata, which is provided to the dictionary, for each type of metadata a translation allowing, for each digitized physical quantity sample or for each piece of context information collected by the sensors to construct a one-to-one correspondence in a representation understandable by the different elements of information processing this data. The monitoring system according to the invention is a system involving heterogeneous physical data, and it is characterized in that it comprises at least one processing module connected to sensors of physical quantities and comprising an indexing engine of analysis means, clock generation means, the sensors being associated with geolocation means, the processing module being connected to display and control means.

The present invention will be better understood on reading the detailed description of an embodiment, taken by way of nonlimiting example and illustrated by the appended drawing, in which: FIG. 1 is a simplified block diagram of a Figure 2 is an example of a metadata tag representation according to the present invention, and Figure 3 is an example of a metadata tag block representation according to the present invention. the present invention.

Using the state of the art in terms of abstractions and architectural design standards allowing interoperability of the functional modules (typically, but not exclusively, of the SOA type), the invention, in its general principle, which can be according to different applications, as described below, consists in the implementation of a standardization of the input interfaces for each type of physical quantity, a mechanism associating them with all the parameters available during their acquisition (dating, zoom value, viewing angle, valve position, ...) and means for keeping this context information unaltered and attached in the course of the different processes, so that each application can transmit it without distortion to the rest of the system , associated with the results of his own treatments. In this way, it becomes possible to objectively and unambiguously define interfaces for each application and to obtain both a true consistency of operation, interoperability between functional blocks and modularity over the long term.

The system of the invention is described with reference to the simplified example of FIG. 1, which relates, in the present case and in a nonlimiting manner, to a monitoring site of a petrochemical complex. This system comprises, in addition to the usual installations of the petrochemical complex, heterogeneous physical data sensors distributed, in this case, in two modules 1.2 placed closest to the sources of the physical data to be captured or scenes to be monitored by video link. . It is understood that the number of these sensor modules and associated devices can be much higher, and that it is a function, in particular, the nature of the installation and the number and location of different sensors connected in module 3. Each of these modules comprises, in addition to the sensors, their possible manipulation devices (camera steering mechanisms, ...) and their pre-processing interfaces and circuits (shaping the signals from the sensors, modulation of a carrier frequency, ...). These physical data can be a pressure, a chemical composition, a density, a pressure, ... or video. The data from these sensors are sent to a processing module 3. This module 3 comprises means 3a for acquiring data from the sensors, an indexing engine 5 and means 4 for storing these data. The role of the indexing engine is to perform the necessary calculations or analyzes on the data received from modules 1 or 2 and to add the results of these calculations or analyzes to the metadata previously associated with these data, thus constituting a metadata tag. enriched, a process that can be iterative. To do this, it is advantageously equipped with analysis tools 6 taking into account the different types of quantities or situations and implementing appropriate algorithms, algorithms specific to each application and obvious to establish by the man of the art on reading the present description. The storage means 4 are either physically included in the module 3 which, in practice, is most often an industrial PC, or external to the module 3 and connected by network to this module 3. In the case where the system of the invention comprises several sites and / or buildings and / or vehicles, each of them may, without inconvenience, include one or more modules 1 or 2 connected by any wired or wireless communication device to a single processing module 3 or each of them can have a set of elements 1 to 6. In any case, the parameterization of some sensors can change dynamically, as is the case, for example, with adjustable zoom cameras, PTZ cameras that are automatically controlled or an operator, or according to the geographical position of moving vehicles. Each set of elements 1 to 6 is itself connected (wired or otherwise) to a secure data path 7, which is for example the backbone of a SOA type architecture. Various functional modules are connected to this artery 7, such as a process management module 8, of the SCADA or other type, and the system interface modules 9, advantageously arranged in a central control room 10. These modules 9 comprise for example viewing screens and control levers. The storage means 4, if they are external to the module 3, as mentioned above and if the artery 7 ensures a sufficient data rate, can be connected to it by this artery 7.

From a dictionary of quantities relevant to the field of applications concerned, and in particular the metadata to be implemented, known to all the users of the system, a first step of implementing the method of the invention consists of adoption of a single, common repository for each data type and associated metadata. Typically, for each of the quantities captured by the sensors of groups 1 and / or 2, it will be the definition, a reference system (for example: angles measured with respect to the geographic north), the unit of measure to use to quantify, the number of significant figures to take into account, ... This dictionary to be shared by the user community, there is interest in it has already been subject to international standardization . This may be that of the SMPTE described above and many chapters of which are directly applicable to the fields targeted by the invention, but also other recognized dictionaries, specific to specific application areas.

The selected dictionary must also propose for each type of metadata a translation allowing, for each digitized physical quantity sample or for each item of context information collected by the sensors of the modules 1 and / or 2, to construct a one-to-one correspondence in a representation understandable by the various information processing elements of module 3. In order to ensure all the necessary flexibility, each of these elements of information is constructed to be self-sufficient, typically including a description of its nature from the dictionary, a measure or an identification and a complement allowing to reconstitute the limits between elements (length, synchronization, ...). An example of this approach is the representation in KLV format (Key Length Value, with K = key, L = total length of the sample, and V = value of the data for the sample considered) proposed by the SMPTE 377M standard. FIG. 2 shows an exemplary information element according to the invention, in KLV form of the SMPTE 377M standard. In this example of representation, illustrated by a latitude, the key (K), which here has a normalized length of 16 bytes, indicates that the following describes the latitude of the object (geographical position of the object expressed in fractions decimals of degrees, positive values corresponding to the northern hemisphere). The length (L) of the sample that follows is here expressed by a number of 14 bytes, and the sample itself is finally expressed by a representative number of the measurement or parameter concerned.

In modules 1 and 2, blocks are formed to group together any number of these information elements, as long as they constitute a coherent set, ie representing a single spatio-temporal set, like all the samples. video of the guardhouse at a given time, or all temperatures in a tank. These blocks are constituted with at least an absolute dating information and gather information that has all been acquired simultaneously, whatever the nature. The resolution of this dating is naturally determined to allow the identification of the most rapidly variable phenomena likely to be encountered. To give an example, if the fastest parameter of a system is conventional video, a resolution better than 10 milliseconds is superfluous, but if the video camera is mounted aboard a plane in supersonic flight and looks at the ground this resolution must be increased by several orders of magnitude. This implies that the system is totally equipped with a unique time distribution device with the required resolution, for example of the NTP type.

These blocks are elementary tags that can be dynamically attached to all physical parameters at modules 1 and 2, which can themselves be mobile and remote. The way of attaching these labels can take different forms as soon as they point with the dynamic precision required the physical size they characterize. This combination of labels and physical parameters can be done in a format where they are nested with the samples of physical quantities, for example: in a conventional telemetry format, in a standard format at a free location such as a proprietary data in a stream transport MPEG-2 or even as an associated file, the dynamic correspondence being provided by the dating information.

When stored in 4, data and tags are always associated. The low volume extracts are transferred as such to the artery 7. Often, however, because of the flow limitations, only the labels are transferred to the artery 7. As part of the results and dressings (or base card) that receives the display and control station 9, these labels are sufficient for this cockpit 9 to obtain precisely from the storer 4 the required data. One of the possible implementations of these blocks in the very simple example of a video camera moving on board a vehicle could consist in transmitting the video (and the possible sound) using an MPEG-2 transport format, as it is done today in France to broadcast digital terrestrial television and use the space left free in the format after each image (location of the proprietary data) to insert a block of tracking of the vehicle trajectory consisting of four elements KLV such as that of Figure 2, arranged according to the representation of Figure 3. A first element 11 describes the nature of the block (trajectory) and its length; it is followed by the element 12 providing the time at which element 13 (latitude) and element 14 (longitude) were recorded. As can be seen with the simple example above on each image of a video stream, these tags allow, whatever the fluctuations due to the transmissions or the processing that can undergo the corresponding data, to guarantee that the presented information on the display screen of the modules 9 in a mosaic of small images or on a base map or combined together to undergo processing in the module 8 are synchronous and correctly located. These are exchanges of blocks of this type that are observed at the interface between the modules 1, 2 and 3, possibly supplemented with parameter data of the sensors.

All of the above applies not only to data passing through modules 1 and 2, but also to all the processes in which they participate later. In concrete terms, if a processing is performed on one or more variable physical data identified as indicated above, processing such as intelligent search of particular information in the video by an analysis algorithm (6) embedded itself in a data engine. indexing (5), flammability of an atmosphere, volume of liquid remaining, ..., the result of this treatment, of whatever nature, is dynamically associated in module 3 with a label constructed from labels information sources used. This approach applies in the same way to physical data that could be synthesized, to a new indexation of previously acquired data and to the commands and controls that the different functional modules exchange through, for example, a type interface. SOA. This is symbolized by the arrows representing the exchanges represented between the module 3 and the artery 7 (for example, as indicated in FIG. 1, context, control and clock data to the module 3, and data extracts and index from the engine 5 and sent to the artery 7. By cons data, such as video, which usually require high bandwidths, advantageously pass directly between the module 3 and the module 9 with their labels, so as to can be compatible with other data transmitted via the data route 7. In this way, the coherence of information (temporal, geographical, nominal, ...) can be guaranteed in all circumstances and in various situations described, in a non A first aspect of the method of the invention is to generalize it, for real-time uses, to all the physical data covered by a surveillance system. including presence detections or alarms. This implies the implementation of a unique system of time distribution, geolocation and, if necessary, identification for all data sources, which can most often rely on GPS receivers located in the immediate vicinity. each data source. This makes it possible to have for each type of data a standardized block (or format) associating the signal or the event with the conditions of its capture, directly compatible with a cartographic representation (for example as shown in FIG. 3) and guaranteeing everything new service that would be fully compatible with the existing system. According to another aspect of the method of the invention, one keeps these metadata dynamically attached to each of the parameters (and can therefore evolve as the successive processing of these parameters) over the entire process of data processing. to ensure traceability, especially when it comes to storing them, for later analysis or legal requirements. This allows, without complex reprocessing or use of specific equipment, to reconstruct, from a simple extract, the conditions of an incident, to provide intrinsically to a police force all the data necessary for its investigations or to regenerate in simulation and faithfully a scenario. Another aspect of the method of the invention consists in extending the above to the results of the processing generated by each of the applications (or abstract services) described above, such as those taken into account by the SOA architectures, ie each once it makes sense, it will be a question of attaching the detections, the calculated results, the alarms deduced from physical parameters of the metadata associated with the original parameters or parameters allowing to go up one way one way with this metadata, this even if the exchange formats are different. To resume the simple example of the vehicle video described above, if an analysis algorithm 6 of the indexing engine 5 detects in a picture of the video a vehicle whose registration number has been reported, it will provide its detection index associated with the location block of Figure 3, to locate and date unambiguously the detection. In this way, the results of the processing, although using a different logical path, are directly homogeneous with the input data and can thus easily be associated with them synchronously and represented on the same cartographic background without any particular treatment. According to another aspect of the method of the invention, complementary to the previous ones, it ensures, in addition to a service control by an architecture of the SOA type or equivalent, the normalization of the input and output interfaces (physical data) of each of the services and therefore the exhaustive definition of the service interfaces, thus allowing a true modularity (multiple providers, performance ranges, ...) of specification and interfaces for each of the services.

Still taking the example of vehicle video, it becomes possible to specify, for constructing a new system, that the input data that will be received by the indexing engine 5 and its analysis algorithms 6, will always be of the type described higher (MPEG-2 transport stream with trajectory tracking blocks). If necessary, this format can even be the subject of a formal standardization process.

According to yet another aspect of the method of the invention, which is complementary to the preceding ones, an extensible library of interoperable services is created, whose inputs and outputs relating to the physical realities are themselves standardized and interoperable, so that systems Various monitoring systems may be formed from these basic building blocks without having to perform functional adaptation and so that pre-packaged basic offers can easily be made for each type of application. In a manner parallel to that presented in the example above, which shows that it becomes possible to specify the indexing engines 5 and the analysis algorithms 6, the developers of these video analysis functions can define accurately the types of video format they are able to take into account and thus constitute catalogs of standard offers. Another aspect of the method of the invention, complementary to the above, results from the practical finding that some unique or indispensable analysis or processing software is only available as frozen products whose interfaces can not be modified. According to this other aspect; the creation of a generic software or even hardware hosting structure, allowing the implementation of the first three aspects mentioned above without resorting to the resources of the rest of the system, and thus to be able to have services with completely standardized interfaces. and apply all of the above. Thus, if in the example already mentioned, there existed on the market an extremely powerful video analysis algorithm, but whose interfaces would only take into account very particular video streams different from those described above and using geolocation formats. owners, this algorithm would be implanted in a specific subsystem that can be a dedicated PC (eg connected to the rest of the system by standard IP links) with the development of translators both input and output data of such so that, seen from the processing module 3, it behaves like a totally standard functionality.

Claims (10)

REVENDICATIONS1. A method for simplifying the integration and modularization of surveillance systems involving heterogeneous physical data, characterized in that it comprises the following steps: a metadata dictionary is chosen which is compatible with the nature of the data transiting through the system; attaches to each datum a label including at least the creation or modification time of this datum, defining a single and common repository for each type of data to be transmitted in the system and for the associated metadata, which is provided to dictionary, for each type of metadata, a translation allowing, for each digitized physical quantity sample or for each piece of context information collected by the sensors, to construct a one-to-one correspondence in a representation that can be understood by the various elements of information processing processing these data and part of the system, and that this information is sent to the corresponding processing devices of the system. 2. Method according to claim 1, characterized in that the metadata dictionary is that of the SMPTE and that it provides for each digitized physical quantity sample a representation in KLV format. Method according to claim 1 or 2, for real-time uses, characterized in that it processes all the physical data covered by a monitoring system, that the time distribution is done by a single system, and that the data are geo-localized. 4. Method according to claim 3, characterized in that the single system identifies all the data sources. 5. Method according to one of the preceding claims, characterized in that the metadata are dynamically attached to each of the parameters over the entire process of data processing. 6. Method according to one of the preceding claims, characterized in that one attaches the detections, the calculated results, the alarms derived from physical parameters metadata associated with the original parameters or parameters to go back one-to-one to these metadata. 7. Monitoring system for implementing the method according to one of claims 1 to 6 and involving heterogeneous physical data, characterized in that it comprises at least one processing module (3) connected to sensors of physical quantities (1, 2) and comprising an indexing motor (5) of the analysis means (6), clock generation means, the sensors being associated with geolocation means, the processing module being connected display and control means (9, 10). 8. Monitoring system according to claim 7, characterized in that it comprises means for capturing and processing video signals. 9. Monitoring system according to claim 7 or 8, characterized in that it comprises several processing modules each connected to at least one group of corresponding sensors, the different processing modules being connected to a secure artery (7) to which the display and control means are also connected. 10. Monitoring system according to claim 7, 8 or 9, characterized in that it comprises an extensible library of interoperable services, whose inputs and outputs concerning the physical realities are themselves standardized and interoperable