WO2008038289A2

WO2008038289A2 - A system and a method for detecting and classifying damage in a pipeline

Info

Publication number: WO2008038289A2
Application number: PCT/IL2007/001202
Authority: WO
Inventors: Gil Pogozelits; Boris Greenstein; Alexander Pikus; Vladimir Yagnatinsky; Eugene Ostrovsky
Original assignee: Soniclynx Ltd.
Priority date: 2006-09-28
Filing date: 2007-10-07
Publication date: 2008-04-03
Also published as: WO2008038289A3

Abstract

The present invention discloses a system and a method for intrusion detection in a protected zone, the system is comprised of a multiplicity of sensors (100) strategically located in the region of interest (ROI) and a main processing unit (130) which analyzes the input signals of these sensors (100) for the purpose of distinguishing between actual intrusions and harmless events. It is the purpose of the present invention to provide a system and a method which allow the optimization of the scanned parameters. According to embodiments of the present invention the input from the sensors (100) is analyzed using a method which minimizes the time it takes to scan the ROI and maximizes the coverage of the ROI by the sensors.

Description

A SYSTEM AND A METHOD FOR DETECTING AND CLASSIFYING

DAMAGE IN A PIPELINE

[001] BACKGROUND OF THE INVENTION

[002] The present invention relates to systems and methods for detecting intrusions, more particularly it relates to systems and methods for territory protection from unauthorized intruders and from malicious sabotaging of objects in security-sensitive aieas. [003] Different methods and systems for the protection of territories, buildings and other constructions are known. Known in the art are methods that make use of geophones which are sensitive to the vibrations in the environment and in the ground in particular for gathering acoustic signals created by actions of intruders or caused by processes occurring in an environment. Geophones are used in passive intrusion detection systems. [004] Geophones, depending on the character of the protected zone, are placed under a superficial layer of the ground or on constructed elements of buildings such as the walls, the floor, or the ceiling. Alternatively, geophones may be attached to wire barriers and used to detect a variety of inputs such as a walking person, a running person, a walking bird, destruction of constructed elements, industrial noise or different kinds of vehicles. The main challenge in such systems is in the signal analysis process. The detected signals must be reliably classified according to their location and origin. Misidentification of signals may cause one of two problems: false positive and false negative errors. False positive errors are false alarms caused by a misidentification of signals produced by a benign source and erroneously setting off the alarm. False negative errors prevent the system from identifying intruders.

[005] Many publications exist addressing the above-mentioned challenge. All known solutions that are based on using these sensors may be separated into two groups as follows: direct analysis of signals based on predetermined threshold values; comparison of received signals with reference signals, which are stored in advance. Amrine et al., "Intruder detecting security system", Pat. No. US 4,001,771, which is incorporated herein as if fully set forth herein, suggests measuring differences of times of seismic wave spreading and/or their velocities (the waves being raised by the intruder). The time differences are then used to identify the intruder and his location. However, this approach requires a large number of geophones of different orientation, and the appropriate quantity of connecting cables, thereby rendering the system proposed by Amrine et al. rather cumbersome and expensive. [006] Chleboun, "Intruder detecting security system", Patent no. US 4,107,660 and Barowitz et al. "Vibration responsive intruder alarm system", Pat. US 4,223,304, both of which are incorporated herein as if fully set forth herein, use frequency separation of the energy of the detected signal to detect intrusion. Using frequency-based methods do not enable separating in all cases events adequately enough, because in real conditions there are no precisely separated frequency ranges for each event, and chosen ranges do not always meet the real requirements of intruder identification. For example, the frequency based methods may not enable distinguishing between different gaits of an intruder. Some frequencies may overlap and result in a loss of the intrusion information that took place or on the contrary may result in false alarms. Furthermore, by performing displacement of ranges the external conditions can impact the detection. [007] Kerr, "Intrusion detection apparatus having multiple channel signal processing", Patent number US 5,194,848, which is incorporated herein as if fully set forth herein, describes multi-channel signal processing with changing filter on each channel to allocate intruders. An alarm is triggered if the input amplitude for a given frequency range exceeds the threshold level during a predetermined period of time. The resulting signal (alarm signal) is then formed by a logical circuit that collects confirming signals from all channels based on a predetermined combination.

[008] Pakhomov "System for detecting intruders" US 6,529,130, which is incorporated herein as if fully set forth herein, describes a system that performs analysis of the bending of initial signals to allocate signals of intruders. The approach proposed by Pakhomov includes using the amplitude threshold, definition of the average distance between maximal values and moments of time appropriate to them, in order to move accurately definition root- mean-square values and values of root-mean-square deviations for these intervals. Pakhomov uses accumulation of data during a time interval of 4-6 seconds for recognition. For many systems this time interval does not correspond to system requirements. During this time interval an intruder may easily cross a protected zone (10-20 meters) and essential information for analysis may be lost. Furthermore, the approach proposed by Pakhomov requires a high signal/noise ratio. To reduce the time interval, Everett, Jr. et al. presented an "Intelligent security assessment system", US patent number 4,857,912, which is incorporated herein as if fully set forth herein. Reduction in time interval may be achieved by incorporating into the system other types of gauges such as, for example, vibration gauges, sound gauges, infra-red and high-frequency gauges. [009] Jensen, "Acoustic alert sensor", patent number US 5,007,032 and US 5,966,406, both of which are incorporated herein as if fully set forth herein, suggest comparing between the numbers of zero-crossings in a predetermined time-interval of a signal according to a given threshold for the identification of events.

[010] Cecic et al., "Method and apparatus for noise burst detection in signal processors", patent number US 5,504,473, which is incorporated herein as if fully set forth herein, suggests a method of detecting the movement of intruders having background noise by comparing the distribution of a minor noise function. The minor noise function may be approached by a known function of probability distribution of the casual source of a signal, and functions of the measured signal distribution. Both noise and signal distribution functions from previously known sources are defined during the predetermined period. [011] Krjukov I.N. et al ", A^aπTHBHoe ycτpoHCTBθ oβHapy^ceHHH H κjiaccHφHκaD,HH ceficMKrøecKHX CHraajioB", patent number 2,212,691, which is incorporated herein as if fully set forth herein uses an auto regression model for the classification of sources of seismic signals. Signals are adapted for concrete conditions with the help of correction factors that are determined according to test checks. The auto regression model allows classifying signals according to their sources. However, some distances between curves may be too small for detection, although the distances may refer to different signal sources. [012] GEOQUIP company (England) developed and manufactures the perimeter security and intrusion classification (PSICON) system, comprised of a chain of geophones and an analyzer. The geophones are used for vibration detection in the perimeter area of the protected object. The vibrations received by the geophones, which may be caused by the activity of intruders, are transformed to electric signals and sent to an analyzer which uses coded representation of the alarm images appropriate to known events, like the occurrence of an intruder, for recognition. These images are stored in system memory and compared with measured data by a method of 'neuron nets'. PSICON allows the finding out and identifying the vibrations caused by the attempts of intruders to overcome a protective barrier with or without the help of a ladder. However, the patent does not relate to protecting territories that have no surrounding walls.

[013] Taylor et al., "Intrusion detection process and device" Patent number US 6,337,625, which is incorporated herein as if fully set forth herein, describes protection of an active system comprising a source of signals and a receiver, which uses at least two acoustic or electromagnetic signals with different frequencies inside the specified zone by comparing the received signals for intrusion detection.

[014] Macalindin, "Intrusion detection system and signal processing circuitry therefore", EP 0 448 290 and Kerr, "Intrusion detection system incorporating deflection-sensitive coaxial cable mounted on deflectable barrier", Patent number US 5,268,672, both of which are incorporated as if fully set forth herein, suggest intrusion detection using microphone cables that are sensitive to the fluctuation of pressure changes. Such systems are widely used. The operational principle of a cable based intrusion detection system is similar to that of a system based on local sensors. Therefore, similar methods may be used to detect intrusion and similar problems may arise. However, since the design of cables allows to protect only a strip having a width of 2-3 meters, it is necessary to cover and network many cables in order to be able to provide protection for a larger territories having a perimeter of several hundred meter or even kilometers. Therefore, using cables and/or local sensors that are sensitive to vibrations and/or pressure leave many problems unsolved. [015] In order to prevent false negative errors the sensitivity of any intrusion detection system may be increased. However, increasing the sensitivity of an intrusion detection system may give rise to higher occurrences of false positive errors by causing false alarms. False alarms may be caused by animals such as cats, birds and the like. None of the above- referenced patents or patent applications that enable classifying a source signal regards the source origin of signals. Additionally, not enough attention is devoted to the problem of detection reaction time, which is extremely important especially in protecting small zones. [016] Implementations of the above-referenced patents and patent application do not provide a comprehensive solution for intrusion detection, i.e., they do not provide a system and method that enable classifying the source signal accurately. However, seismic gauges have many advantages such as low cost, and it is therefore desirable to try to improve systems and methods that are based on seismic gauges, which may be used for protecting strategically important locations or sites such as oil, gas and water pipelines. [017] Optical fibers may be used as a means for detecting intrusion in a protected zone. An optical fiber enables detecting an intrusion along the length of the fiber, whereby the fiber is the sensing element and is suitable for protecting zones that have a large perimeter of, e.g., several hundred meters or kilometers.

[018] Davidson, "Security system and strip or strand incorporating fiber-optic wave-guide means therefore", patent number US 4,275,294, which is incorporated as if fully set forth herein, describes a system comprising optical fibers that are strung out such that breakage or sever distortion anywhere along the length of the fiber will be detected. An obvious disadvantage of the system proposed by Davidson is the visibility of the fibers to a potential intruder, possible destruction of the fiber when an intrusion occurs, and exposure to environmental hazards such as weather conditions (e.g., rain, snow, sand, heat) and the like.

[019] Butter presents in patent number US 4,297,684, "Fiber optic intruder alarm system", which is incorporated herein as if fully set forth herein, is a system utilizing underground installation of a fiber optic cable in order to overcome the disadvantages of the system presented by Davidson. Such systems utilize measurement in changes of the phase of the light transmitted in the optical fiber. Measuring phase changes renders the system presented by Butter extremely costly, and is subject to have a high likelihood to trigger false alarms.

[020] No known prior art physical security system measures changes in the intensity of light transmitted threw fiber optic strands. The only technology known to applicants related to measurement of change in intensity of light in fiber optic is embodied in the teachings of Pat. U.S. 4,342,907 to Macedo et al." Optical sensing apparatus and method", which is incorporated herein as if fully set forth herein. In that patent it is noted that a cladded optical fiber subjected to pressure has inherent therein a change of the refractive index in the cladding material. The change of the refractive index results in a change of the light intensity exiting the fiber.

[021] SUMMARY

[022] Disclosed is a method for detecting intrusions in a protected zone in which at least one geophone sensor is installed by distinguishing between different event types. The detection relies on sensor output analysis. The method comprises the steps of collecting signals from the geophone sensors, performing analysis on different mathematical dimensions of the signals, and identifying at least one event type according to signal patterns appearing in the analysis.

[023] The event may be a vehicle riding and the identification is performed by analyzing the density of distribution of the transformed signals. The event may be a person walking and the identification is performed by using a correlation function on inputs signals from two geophones which are in close proximity, and which have received signals with maximal energy. The event may be caused by the presence of birds and the identification is performed by using a correlation function on inputs signals from two geophones which are in close proximity, and which have received signals with maximal energy. The event may be caused by an external event wherein the identification is performed by using a correlation function on inputs signals from two geophones which are in close proximity, and which have received signals with maximal energy.

[024] The method may also include the steps of sampling the signals and dividing the sampled signals into frames. The analysis may be performed on each frame. The sufficient duration for the recognition of the considered events should be greater than the frame duration. The analysis may include calculating the first norm of the absolute value of the signals and a second norm of the energy of the signals.

[025] The method may also include the step of selecting the output signal of a single sensor. The selected sensor is the sensor with the highest norms. The method may also include the step of filtering the samples of the sensor using a high-pass filter, wherein the filtering enables emphasizing features of the signals. The different mathematical dimensions may include the signal dimension and the spectrum dimension. The analysis may be performed on the envelope of the signal dimension. The analysis of the envelope may be performed according to the construction of the density of the transformed samples distribution. The analysis may include an estimation of the cross-spectral density of the sequences of the filtered samples of two sufficient informative seismograms of a given frame. The estimation may be performed by preliminary calculations of a cross-correlation function of the frames. [026] The sensors may be extended in the protected zone in different directions. The cross- correlation function may be performed based on power spectral density estimations. The method may also include the step of activating an alarm if the identified event is identified as an intrusion of the protected zone.

[027] The method may also include the step of calculating the differential output beams transmitted through at least two fibre optic strands of fibre optic strands which are installed in the protected zone. The method may also include the step of identifying an intrusion in the protected zone in accordance with the calculated deferential output of the beams. [028] Also disclosed is a system for detecting intrusions in a protected zone in which at least one geophone sensor is installed by distinguishing between different event types. The detection relies on sensor output analysis. The system includes a data collecting device for collecting signals from the geophone sensors, a calculating device for performing analysis on different mathematical dimensions of the collected signals, and a module for identifying at least one event type according to signal patterns appearing in the analysis. [029] The event may be a vehicle riding or a person walking. The event may be caused by the presence of birds or by an external event. The system may also include alarming means for alerting if the event is identified as an intrusion.

[030] The system may also include at least two fibre optic strands which installed in the protected zone and a transmitter device for transmitting a laser beam through the fibre optic strands. The system may also include a receiver device for measuring the differential output in the beams transmitted through the fibre optic strands, and a calculating device for identifying intrusions whenever the measured differential in the output beams exceeds a predetermined threshold.

[031] BRIEF DESCRIPTION OF THE DRAWINGS

[032] The subject matter regarded as the invention will become more clearly understood in light of the ensuing description of embodiments herein, given by way of example and for purposes of illustrative discussion of the present invention only, with reference to the accompanying drawings, wherein

[033] Figure 1 is a block diagram of the data collection and data processing system in accordance with some embodiments of the present invention;

[034] Figures 2a, 2b, 2c and 2d are examples for signals collected by the sensors of the system, whereas Figure 2a is an example for the signal produced by a person walking; Figure

2b is an example for the signal produced by a vehicle in motion; Figure 2c is an example for the signal produced by an easy impact such as by the presence of birds; Figure 2d is an example for the signal produced by a sound which was not produced by an event in the monitored zone;

[035] Figure 3 is a flowchart illustrating an intruder detection method in accordance with an embodiment of the present invention;

[036] Figure 4 is a flowchart illustrating the first and the second algorithms of the intruder detection method in accordance with an embodiment of the present invention; [037] Figure 5 is an illustration of the density of distribution functions of the transformed signals for four events: a person walking, a vehicle riding, the presence of a bird and external noise;

[038] Figure 6 includes illustrations of separate representation of density of distribution functions for four events: a person walking, a vehicle riding, the presence of a bird and external noise;

[039] Figure 7 includes illustrations of comparative diagrams of the power spectral density estimation for three events: a person walking, the presence of a bird and external noise;

[040] Figure 8 includes illustrations of comparative diagrams of the power spectral density estimation for three events: a person walking, the presence of a bird and external noise;

[041] Figure 9 includes illustrations of comparative diagrams of the power spectral density estimation for three events: a person walking, the presence of a bird and external noise;

[042] Figure 10 includes illustrations of comparative diagrams of the power spectral density estimation for three frames of event of a vehicle riding;

[043] Figure 11 is an illustration of a schematic diagram of conventional common view of the components of an embodiment of the present invention;

[044] Figure 12 is an illustration of a schematic diagram of conventional view in plane of the components of an embodiment of the present invention.

[045] The drawings together with the description make apparent to those skilled in the art how the invention may be embodied in practice.

[046] No attempt is made to show structural details of the invention in more detail than is necessary for a fundamental understanding of the invention. [047] It will be appreciated that for simplicity and clarity of illustration, elements shown in the figures have not necessarily been drawn to scale. For example, the dimensions of some of the elements may be exaggerated relative to other elements for clarity. Further, where considered appropriate, reference numerals may be repeated among the figures to indicate corresponding or analogous elements.

[048] DETAILED DESCRIPTION

[049] The present invention discloses a system and a method for intrusion detection in a protected zone. The system is comprised of a multiplicity of sensors strategically located in the region of interest (ROI) and a main processing unit which analyzes the input signals of these sensors for the purpose of distinguishing between actual intrusions and harmless events. It is the purpose of the present invention to provide a system and a method which allow the optimization of the scanned parameters. According to embodiments of the present invention the input from the sensors is analyzed using a method which minimizes the time it takes to scan the ROI and maximizes the coverage of the ROI by the sensors. [050] It is to be understood that an embodiment is an example or implementation of the inventions. The various appearances of "one embodiment," "an embodiment" or "some embodiments" do not necessarily all refer to the same embodiments.

[051] Although various features of the invention may be described in the context of a single embodiment, the features may also be provided separately or in any suitable combination. Conversely, although the invention may be described herein in the context of separate embodiments for clarity, the invention may also be implemented in a single embodiment. [052] Reference in the specification to "one embodiment", "an embodiment", "some embodiments" or "other embodiments" means that a particular feature, structure, or characteristic described in connection with the embodiments- is included in at least one embodiments, but not necessarily all embodiments, of the inventions.

[053] It is to be understood that the phraseology and terminology employed herein is not to be construed as limiting and are for descriptive purpose only.

[054] The principles and uses of the teachings of the present invention may be better understood with reference to the accompanying description, figures and examples.

[055] It is to be understood that the details set forth herein do not construe a limitation to an application of the invention.

[056] Furthermore, it is to be understood that the invention can be carried out or practiced in various ways and that the invention can be implemented in embodiments other than the ones outlined in the description below.

[057] It is to be understood that the terms "including", "comprising", "consisting" and grammatical variants thereof do not preclude the addition of one or more components, features, steps, or integers or groups thereof and that the terms are to be construed as specifying components, features, steps or integers.

[058] The phrase "consisting essentially of", and grammatical variants thereof, when used herein is not to be construed as excluding additional components, steps, features, integers or groups thereof but rather that the additional features, integers, steps, components or groups thereof do not materially alter the basic and novel characteristics of the claimed composition, device or method. [059] If the specification or claims refer to "an additional" element, that does not preclude there being more than one of the additional element.

[060] It is to be understood that where the claims or specification refer to "a" or "an" element, such reference is not be construed that there is only one of that element.

[061] It is to be understood that where the specification states that a component, feature, structure, or characteristic "may", "might", "can" or "could" be included, that particular component, feature, structure, or characteristic is not required to be included.

[062] Where applicable, although state diagrams, flow diagrams or both may be used to describe embodiments, the invention is not limited to those diagrams or to the corresponding descriptions. For example, flow need not move through each illustrated box or state, or in exactly the same order as illustrated and described.

[063] Methods of the present invention may be implemented by performing or completing manually, automatically, or a combination thereof, selected steps or tasks.

[064] The term "method" refers to manners, means, techniques and procedures for accomplishing a given task including, but not limited to, those manners, means, techniques and procedures either known to, or readily developed from known manners, means, techniques and procedures by practitioners of the art to which the invention belongs.

[065] The descriptions, examples, methods and materials presented in the claims and the specification are not to be construed as limiting but rather as illustrative only.

[066] Meanings of technical and scientific terms used herein are to be commonly understood as by one of ordinary skill in the art to which the invention belongs, unless otherwise defined. [067] The present invention can be implemented in the testing or practice with methods and materials equivalent or similar to those described herein.

[068] Unless specifically stated otherwise, as apparent from the following discussions, it is to be understood that utilizing terms such as "computing", "processing", "determining", "calculating," or the like, refer to the action or processes or both of a computer or computing system, or similar electronic computing device, that manipulate and/or transform data represented as physical, such as electronic quantities within system computing registers and/or memories into other data similarly represented as physical quantities within system computing memories, registers or other such transmission, information storage or display devices.

[069] Embodiments of the present invention are comprised of several aspects. The first aspect consists of using signal analysis of the signals received from seismic gauges-sensors which are positioned in the ROI using several mathematical methods. In particular, this processing includes analysis of several mathematical dimensions such as the signal dimension and the spectral dimension. The method enables the making of distinctions between signals which are closely related in form.

[070] The second aspect of the invention includes processing which is made by arrays- frames allocated by sliding windowing. Each frame comprises a defined quantity of seismograms and every seismogram corresponds to its own geophone. Each seismogram contains a defined quantity of digitalized samples, which is equal to window duration multiplied by digitization frequency. Window duration is minimal, but is no less than the duration of the intervals containing a part of the energy of each analyzed signal which is sufficient for recognition. [071] An additional aspect consists of analyzing the density of distribution of the transformed signals, which enables separating the signal produced by a vehicle from other groups of signals including signals produced by a person walking and by external noises. This enables the exclusion of frames containing vehicle signals, and the reduction of the number of possible events to be analyzed at the following steps.

[072] Another aspect of embodiments of the present invention consists of using a correlation function as an additional stage in the recognition procedure. This stage allows distinguishing between different signals which are similar in their form and duration. Distinguishing between these signals is difficult due to the disorder of parameters of real signals. The arguments of this correlation function are inputs signals from two geophones which are in close proximity, and which have received signals with maximal energy. Cross- correlation of these signals, received from such geophones, especially if these geophones are located in environments with rather low wave distribution velocity, enable distinguishing between signals which have similar forms, like the steps of a person walking and the presence of a bird.

[073] The proposed system and method for intruder identification allows decreasing the analysis time to approximately one second and decreasing the probability of false alarms even when only two geophones are used. The method is programmed to provide optimal results for typical events which have the greatest probability of occurring, both within the limits of a protected zone and in its nearest vicinity. According to one embodiment of the present invention the method is programmed to identify four types of events: a person walking, the movement of a vehicle, the presence of a bird and external noises. The method also enables identifying other types of events such as the impacts of a hammer or a driii. [074] According to an embodiment of the present invention the proposed system combines a laser light source, optical fibre strands which are pressure sensitive transducers, a photo- optic receiver, a signal processing unit and an external response system operatively connected to produce a relatively simple, trouble free physical security system. The divided laser beam is directed into two optical fibres, built in the ground along the protected pipeline. [075] The radiation received from each of the two photodiode receivers is directed to a differential amplifier. The two fibre waveguides are optically and electrically arranged so that only changes in one fibre strand with respect to the other are detected. The differentia measurement capability results in an extremely high sensitivity concurrent with a high common mode rejection against effects or changes in both fibres. System noise is reduced by the provision of band pass filter that is connected to the output of the differential amplifier. If the differential optical signal exceeds a predetermined threshold, it triggers a response such as setting off the alarm.

[076] An automatic signal processing or servo loop operates to maintain the optical fibre loops balanced with respect to each other so as to compensate for long term drifts in conditions such as temperature and the like. Without such a balancing control environmental changes could effect the differential measurement referred to above. Transient signal changes are not sensed by the servo control circuit. Thus, when a change of condition of one fibre with respect to the other occurs, which may indicate an intrusion, it is detected. The optical fibre strands are positioned far enough apart from one another so that an intruder would cause a change in one with respect to the other. Either optical fibre strand loop can serve as the pressure sensitive transducer. [077] Figure 1 is a block diagram illustrating the components of the disclosed system in accordance with an embodiment of the present invention. Geophones sensors 100 are placed within a protected zone according to their sensitivity and in view of the environmental parameters. Cables 110 are connected to each geophone 100 and their other outputs are connected to data collecting device 120. The output of collecting device 120 is connected to a calculating device such as a computer 130, containing the processing program. Alarm system 140 is connected to an output of computer 130. The scheme of the connection of geophone 100 to a device of data collection 120 can be consecutive or parallel, depending on the organization of the interface.

[078] Upon moving on the surface of a protected zone or in its vicinity each living being or technical mechanical means causes the occurrence of seismic fluctuations extending in the ground and perceived with the geophones 100. The numbers of the needed geophone sensors 100 is defined by the size of a protected zone, their sensitivity and resolution. For example, using standard geophones in an environment of sandy ground, the distance between geophones should not exceed 10-20 meters.

[079] Collecting device 120 receives electrical signals forming by geophones sensors 100. Computer 130 performs the transformation of signals into digital form (sampling) with frequency ranging between several hundreds and several thousand Hertz. The digital data stream forming from received signals, inputted into computer 130, consists of separate seismograms each of which corresponds to its own geophone sensor 100. [080] First, the dataflow is divided by a sliding window into separate arrays (frames) according to the accepted scheme of processing, as well as the duration of an energetically significant part of a signal or ϊhe frequency characteristic of the intruder and extraneous preventing events. Each frame represents a matrix each line of which, consisting of a sequence of samples, corresponds to one sensor control and represents a seismogram. The quantity of lines is equal to the number of sensors. The quantity of columns (quantity of elements in a seismogram) is equal to the duration of the record of said frame increased by the frequency of digitization.

[081] The division into frames is possibly carried out with or without overlapping depending on the sampling requirements. Creating overlaps between consecutive frames ensure that signals are not divided between frames, and enables avoiding erroneous recognition of the truncated parts.

[082] The duration of a sliding time window is determined so as to be as small as possible, but sufficient to allow analysis of all considered events. This requirement corresponds to that of the duration of the significant part of energy. The sufficient duration for the recognition of the considered events should not be greater than the frame duration. Accordingly, all seismograms within the limits of said frame have the same duration. The quantity of samples in a seismogram is defined by its duration and frequency of digitization; it usually ranges from 1000 Hz up to 4000 Hz for the ground, and is within a considerably wider range for dense environments, such as concrete.

[083] In the given examples the frame duration is 1 second. The frequency of digitization is 1000 Hz. According to this each seismogram has 1000 samples (denote this number by N; N=IOOO). Examples for signals caused by four types of actions which are essential to the protection of a protected zone are shown in Figures 2a, 2b, 2c and 2d. These signals are shown as the frames consisting of two seismograms after their digitization. In these figures the horizontal axes, abscissae, represent values of time (in terms of digitization samples), and the vertical axes, ordinatae, the amplitudes of the samples.

[084] Examples for signals caused by a person walking are recorded by the sensors as shown in Figure 2a. Examples for signals caused by the presence of large birds are shown in Figure 2b. These examples show that the signals caused by walking and birds are very similar in duration and form, a fact which makes their recognition more difficult. [085] Examples for signals caused by a moving vehicle are shown in Figure 2c. An example for a signal caused by external noise is shown in Figure 2d. These diagrams show the affinity between expected signals.

[086] Comparing these signals shows small differences in their characteristic envelope and frequency; however, the complex and diverse forms of actual signals make it difficult to perform a direct analysis of these parameters and the identification of events using only one procedure. This is especially evident when comparing signals caused by a person walking and the signals caused by the presence of birds.

[087] Neither frequency analysis nor amplitude analysis would enable making definite distinctions between the events specified above. The disclosed method makes use of a two- stage procedure which includes two algorithms. If needed additional algorithms may be used. At each stage full or at least partial allocation of one event is made.

[088] For an increase in the efficiency of recognition, the analysis is made in a different mathematical dimension at each stage, in particular, in the signal dimension and the spectrum dimension. The analysis is done with the help of different mathematical algorithms determined by the features of the corresponding event types, whereas each type must include at least one event. [089] Figure 3 is a flowchart illustrating the method of signal analysis and the detection of the presence of an intruder in accordance with some embodiments of the present invention. The signals, received by the sensors, are gathered by collecting device 120 (step 210) and transmitted to computer 130 where they are transformed to the digital form. Then they are sampled (step 220) and divided into frames (step 230). These frames are sequentially analyzed.

[090] The analysis of the obtained data begins with the reception of the next frame. The norm is calculated (algorithm step 240) for each seismogram of a given frame. The sum of the absolute value of samples of each seismogram of said frame is used to calculate this norm. The other norm which may be calculated is energy (equal the sum of quadrate samples). The disclosed method may use any one of the two.

[091] Next, a seismogram is selected (step 250), such as the seismogram with the greatest sum. This seismogram is selected as a sufficiently informative seismogram; it consists of an array of samples. The samples of this seismogram are denote by xl(i), where i= 1, 2,...N. [092] The samples of the chosen seismogram are filtered (step 260) using a high-pass filter which emphasizes features of various signals. This high-pass frequency filtering is realized by a differentiation of said sequence with respect to time. For this differentiation the calculation of finite differences is used. Then the simple finite differences Δxl(i)=xl(i+1)~ xl(i) is the sequence of the filtered samples, and i=l, 2... N-I.

[093] Than the first algorithm (step 270) and the second algorithm (step 290) are activated and their consequent conditions - first condition (step 280) and second condition (step 300) - are checked as described below. Provided that both conditions are found to indicate that an intrusion has occurred, the alarm is activated (step 310). The first algorithm is designed to distinguish between the event of a vehicle moving in the protected zone and all other events; the second algorithm is designed to distinguish between the remaining events and the event of a person walking.

[094] The array of the filtered samples is used for the subsequent realization of the method. Figure 4 is a detailed block diagram illustrating the first and the second algorithm in accordance with the present invention. The first algorithm 270 begins with calculating the absolute values (step 272) of the sequences of filtered samples, and defining the maximal value of sequence elements (step 273). For improving the division of various events the distribution function of the elements of the sequence are defined (step 274). The resulting sequence of transformed samples is { abs( Δxl(i) ) }, [where i=l, 2,...N-I] and its distribution function is formed.

[095] The first algorithm performs processing in the signal dimension and carries out the analysis envelope of the sequence using features of signals envelope caused by vehicle movement. The analysis of these envelopes is carried out according to the construction of density of the transformed samples distribution (step 275). For this analysis (step 276) the estimation parameter designed accordingly to ns(k),(where Ic=I, 2,...kmax), and at least two categories of these parameter estimation are set. Two categories denoted by "< "and ">" are used in this method.

[096] The parameters ns(k) values are equal to the number of the transformed samples values which belong to the k-th interval of values: from g (k+l)-mg to g (k)-mg, where g(k+l)<g(k) and g(kmax)<l. Where g (k) is the predetermined boundaries of the k-th interval. Experimentations have shown that it is sufficient to use boundaries g(k) =1.2 -0.2^»k, where k<6. These boundaries are used by calculating the density distribution of the examined events, such as the events mentioned above, and calculating the function ns(k) for these events. This distribution function is constructed for the transformed sequence of samples and allows the estimation of the form of the sequence envelope in an indirect way. [097] For instance, the envelopes corresponding to events of a person walking and the presence of a bird, have the peak form of the limited duration (~0.15-0.2 sec). Thus, most of the transformed samples correspond to the interval with minimal "k" values, in particular, k=l. A small minority of these samples correspond to the values k>l. Consequently, a distribution function corresponding to these events has the abrupt maximum in the first interval (k=l) and small values in other.

[098] The signal created by the movement of a vehicle has a smooth envelope with limited range of values. The distribution function for similar events is, therefore, a flattened curve. The transformations of initial signals enable obtaining the substantially significant differences of said distribution functions for these two groups of events. Experimentation shows that in the analysis of the presented dependences only two intervals may be used for distinguishing between these events. One of the following two may suffice: ns(l) <tk(l) or ns(2) >tk(2) where tk(l) and tk(2) are predetermined parameters of the first and second intervals accordingly. The predetermined threshold values which may be used are tk(l)=650 and tk(2)=200.

[099] The first condition is ns(l) < tk(l). The satisfaction of the first condition defines the event as the movement of a vehicle. In this case the following steps are carried out: identifying intruder absence, ending given frame check, accepting next frame and repeating this procedure according to the preceding. Provided that the first condition is not satisfied then event of a moving vehicle is excluded from the group of possible events. The subsequent analysis defines a person-intruder among the remaining three events using the second algorithm 290.

[0100] The second algorithm 290 affords performance of an estimation of the cross-spectral density of the sequences of the filtered samples of two sufficient informative seismograms of a given frame. The second sufficient informative seismogram outputs, for example, the seismogram which has the second largest sum of absolute values of initial samples following the first. There are several methods of performing the cross-spectral density estimation. Various methods may be distinguished by their different ability to separate. [0101] In the given algorithm this estimation is afforded by preliminary calculation of a cross-correlation function (step 292) of two sequences of the filtered samples. The cross- correlation function allows allocating the common features of signals which are captured by two sensors extended in the environment in different directions. A spectral transformation of the cross-correlation function enables comparing the results for different events. The use of the power spectral density (PSD) estimation (step 293) as a function of cross-correlation, instead of the commonly used Fourier-transformation, allows receiving functions which are more convenient for comparison.

[0102] Figures 7, 8, 9 and 10 illustrate comparative diagrams of the power spectral density estimation for different events. The values of PSD estimation are given in relative units and are plotted as ordinates - amplitude. The frequency parameter "f" is plotted on the horizontal axis.

[0103] In Figures 1, 8 and 9 examples of the calculated PSD estimation for three events - a person walking, the presence of a bird and external noise - and their differences are shown. In Figure 10 the PSD distribution for three frames, containing the signals of the event of a moving vehicle are shown. The main peaks of the event of a moving vehicle have a spread in all frequency ranges and their location does not provide means for distinguishing between it and the event of a person walking. This explains the necessity of using the first algorithm to distinguish between the event of a moving vehicle and other events.

[0104] The analysis of peak distribution allows distinguishing between other analyzed events. For this purpose the frequency range is divided into jmax intervals by the predetermined frequency borders f(j) (where j=0, 1, 2,...jmax) and f(j)≤f(j+l). Accordingly, the interval from f(j) to f(j+l) is the j-th interval, and when j=2 the second interval corresponds to the intruder actions.

[0105] Experiments have shown that the values of these borders are f(0)=f(l)=40 Hz, f(2)=120 Hz, f(3)=240 Hz , f(4)=320, f(5)=400 Hz etc for sandy environments. Additionally, the weight parameter a(j) is set, where a(j)>0, the weight parameters factors were a(l)=0, a(2)=l, a(3)=l, a(4)=0.6, a(5)=0.5, for all other values a=l. For identifying the event of a person walking the second algorithm is used MaxPSD(2) > Max { a(j)^« MaxPSDQ }, where:

MaxPSD(j) is the maximal value of PSD on the j-th interval, a(j) is the predetermined weight factors on j-th interval, j= 1,3,4,5...jmax. Q≠ 2).

[0106] The necessity of these weight factors (step 294) is dictated by the complicated and unsteady forms of real signals. Then the analysis of said distribution function is carried out and the second condition is checked (step 295).

[0107] If this condition is not satisfied then the given frame corresponds to either the event of external noise or to the presence of a bird, and this frame does not carry information about intrusion. Since no intrusion was detected the method ends the analysis of the given frame, retrieves the next frame and starts the analysis procedure from the beginning. Provided that intrusion was detected, the method activates the alarm (step 310), and the method ends the analysis of the given frame, retrieves the next frame and starts the analysis procedure from the beginning. In the above examples the PSD calculations make use of a window size of 256 samples and the sampling frequency of 1000 Hz.

[0108] In environments in which there is high speed acoustic wave distribution, such as building walls made of dense materials, it is sufficient to use only one seismogram with high levels of energy. The function of cross-correlation in such cases is simpler since an autocorrelation function may be used. Experiments have shown that in such environments it is sufficient to use a small number of sensors. Using even only two sensors the above described method may be used to identify intrusions to the protected zone and the operation of tools in the building.

[0109] Figure 11 is a schematic diagram of conventional common view of the components of an embodiment of the present invention; Figure 12 is an illustration of a schematic diagram of conventional view in plane of the components of an embodiment of the present invention. Illustrated is a fibre optic security system for detecting unauthorized activity in the vicinity of pipelines. The pipelines optionally include oil, gas or water pipelines. The system consists of transmitter subsystem 6, pairs of fibre optic waveguides lχ and 7₂ receiver 8 and alarm device 9.

[0110] Transmitter subsystem 6 consists of laser source 10 and beam divider 11 operatively connected to fibre optic strands forming two waveguides 1\ and 1%. Receiver 8 consists of photodiodes 12 and 13, preamplifiers 14 and 15, differential amplifier 16, band pass filter 17, threshold detector 18 and output signal timer 19 which are operatively connected. [0111] Fibre optic waveguides I_x and I₂ may be, for instance, low-loss silicone polymer clad glass core fibre; however, any other type of fibre is also within the scope of the present invention. Fibres 7i and 7₂ are installed underground.

[0112] The ends of waveguides 7χ and 7₂ are connected to photodiodes 12 and 13 via the connectors in receiver 8. Photodiodes 12 and 13 may be any type of PIN diodes, such as Honeywell fibre optic detectors. Photodiodes 12 and 13 operate in the photovoltaic mode and are connected to the input lads of two operational amplifiers preamplifiers 14 and 15. Amplifiers 14 and 15 are contained in a simple package to minimize temperature drift. Amplifiers 14 and 15 are identical with regard to the arrangement and value of the components.

[0113] Preamplifiers 14 and 15 are connected to a differential amplifier 16. The purpose of this amplifier 16 is to amplify any voltage difference between the output from preamplifiers 14 and 15. The output of differential amplifier 16 is connected to bandpass filter 17. Any extraneous noise introduced by the photodiodes 12, 13 preamplifier 14, 15 and differential amplifier 16 is eliminated. The output from bandpass filter 17 is connected to threshold detector 18. If the input exceeds a predetermined level (for example 10V), a signal appears at the output of the threshold detector 18. Either a positive or negative signal exceeding the 50 milli-volt level will cause a 10V signal output.

[0114] The output of threshold detector 18 is connected to output signal timer 19. Output signal timer 19 provides an output of fixed duration every time the threshold detector 18 generates an output signal. The duration of the output signal of timer 19 is determined by predefined parameters. [0115] With no external intervention the laser beams in waveguides 7i and 7₂ are identical and no difference is detected by differential amplifier 16. In such Cases the output of differential amplifier 16 is zero and the subsequent circuits 17, 18, 19 and 9 are inoperative. Whenever a small pressure is exerted anywhere along the length of either fibre 7i or 7₂ an optical loss occurs at that point and less light is received at photodiodes 12 or 13. Such pressure may result from an intruder stepping on optical fibre 7i or 7₂ buried in the ground near the pipeline. The radiation received at detectors 12 and 13 is no longer equal as a result of an optical loss occurring in one of the fibres 7i or 1_%. This signal difference is amplified by the differential amplifier 16. After passing through the band filter 17 the differential voltage is incident on a threshold detector 18. If the signal exceeds a preset threshold, threshold device 18 issues a large voltage signal which triggers timer circuit 19. Timer circuit 19 activates alarm device 9 for a preset amount of time.

[0116] By comparing the voltage levels at diodes 12 and 13 and amplifying the differences in amplifier 16 rather than measuring absolute values, greatly increases the sensitivity the proposed system. In order to prevent false alarms any noise spikes which may be amplified by the operational amplifier 16 and may trigger the threshold detector 18 have to be suppressed. This is achieved with a narrow bandpass filter 17 which transmits only signals received from the laser and thus eliminates noise spikes generated by the photodiodes 12, 13, preamplifiers 14, 15 or differential amplifier 16.

[0117] In addition to noises which might trigger the threshold detector 18, long term drift of one fibre output with respect to the other fibre is controlled. For instance, temperature differences in the environments of fibres 7i and I₂ or temperature differences in the electronics may cause a slow drift of the outputs at 12 and 13. Small differences might be amplified by differential amplifier 16 and these signals might be transmitted by bandpass filter 17 and trigger the threshold detector 18. To overcome this problem slow voltage changes are sampled by the servo control loop. Small voltage differences which appear at the output of bandpass filter 17 charge a capacitor in the integrator circuit. The voltage at this capacitor, which is located at the output of the integrator, is amplified and used to control the current of the transistor.

[0118] It is to be understood that some embodiments of the invention may be implemented, for example, using a machine-readable medium or article which may store an instruction or a set of instructions that, if executed by a machine, cause the machine to perform a method or operations or both in accordance with embodiments of the invention.. Such a machine may include, for example, any suitable processing platform, computing platform, computing device, processing device, computing system, processing system, computer, processor, or the like, and may be implemented using any suitable combination of hardware or software or both. The machine-readable medium or article may includes but is not limited to any suitable type of memory unit, memory device, memory article, memory medium, storage article, storage device, storage medium or storage unit such as, for example, memory, removable or non-removable media, erasable or non-erasable media, writeable or re-writeable media, digital or analog media, optical disk, hard disk, floppy disk, Compact Disk Recordable (CD- R), Compact Disk Read Only Memory (CD-ROM), Compact Disk Rewriteable (CD-RW), magnetic media, various types of Digital Versatile Disks (DVDs), a tape, a cassette, or the like. The instructions may include any suitable type of code, for example, an executable code, a compiled code, a dynamic code, a static code, interpreted code, a source code or the like, and may be implemented using any suitable high-level, low-level, object-oriented, visual, compiled or interpreted programming language. Such a compiled or interpreted programming language may be, for example, C, C++, Java, Pascal, MATLAB, BASIC, Cobol, Fortran, assembly language, machine code and the like.

[0119] While the invention has been described with respect to a limited number of embodiments, these should not be construed as limitations on the scope of the invention, but rather as exemplifications of some of the embodiments. Those skilled in the art will envision other possible variations, modifications, and programs that are also within the scope of the invention. Accordingly, the scope of the invention should not be limited by what has thus far been described, but by the appended claims and their legal equivalents. Therefore, it is to be understood that alternatives, modifications, and variations of the present invention are to be construed as being within the scope and spirit of the appended claims.

Claims

What is claimed is

1. A method for detecting intrusions in a protected zone by distinguishing between different event types, wherein at least one geophone sensor is installed in said protected zone and said detection relies on sensor output analysis, said method comprising the steps of:

- collecting signals from said geophone sensors;

- performing analysis on different mathematical dimensions of said signals; and

- identifying at least one event type according to signal patterns appearing in said analysis.

2. The method of claim 1 wherein said event is a vehicle riding wherein said identification is performed by analyzing the density of distribution of the transformed signals.

3. The method of claim 1 wherein said event is a person walking wherein said identification is performed by using correlation function on inputs signals from two geophones which are in close proximity, and which have received signals with maximal energy.

4. The method of claim 1 wherein said event is caused by the presence of birds wherein said identification is performed by using correlation function on inputs signals from two geophones which are in close proximity, and which have received signals with maximal energy.

5. The method of claim 1 wherein said event is caused by an external event wherein said identification is performed by using correlation function on inputs signals from two geophones which are in close proximity, and which have received signals with maximal energy.

6. The method of claim 1 further including the steps of - sampling said signals;

- dividing said sampled signals into frames, wherein said analysis is performed on each frame and the sufficient duration for the recognition of the considered events is greater than the frame duration..

7. The method of claim 1 wherein said analysis includes calculating at least one of the following: a first norm of the absolute value of said signals, a second norm of the energy of said signals.

8. The method of claim 7 further including the step of selecting the output signal of a single sensor.

9. The method of claim 8 wherein said selected sensor is the sensor with the highest said norms.

10. The method of claim 8 further including the step of filtering said samples of said sensor using a high-pass filter, wherein said filtering enables emphasizing features of said signals.

11. The method of claim 1 wherein said different mathematical dimensions include the signal dimension and the spectrum dimension.

12. The method of claim 11 wherein said analysis is performed on the envelope of said signal dimension.

13. The method of claim 12 wherein said analysis of said envelope is performed according to the construction of density of the transformed samples distribution.

14. The method of claim 6 wherein said analysis includes an estimation of the cross-spectral density of the sequences of the filtered samples of two sufficient informative seismograms of a given frame.

15. The method of claim 14 wherein said estimation is performed by preliminary calculation of a cross-correlation function of said frames.

16. The method of claim 15 wherein said sensors are extended in said protected zone in different directions.

17. The method of claim 15 wherein said cross-correlation function is performed based on power spectral density estimation.

18. The method of claim 1 further including the step of activating an alarm if said identified event is identified as an intrusion of said protected zone.

19. The method of claim 1 further including the step of calculating the differential output beams transmitted through at least two fibre optic strands, wherein said fibre optic strands are installed in said protected zone.

20. The method of claim 19 further including the step of identifying an intrusion in said protected zone in accordance with said calculated deferential output of said beams.

21. A system for detecting intrusions in a protected zone by distinguishing between different event types, wherein at least one geophone sensor is installed in said protected zone and said detection relies on sensor output analysis, said system comprising:

- data collecting device for collecting signals from said geophone sensors;

- a calculating device for performing analysis on different mathematical dimensions of said collected signals; and

- a module for identifying at least one event type according to signal patterns appearing in said analysis.

22. The system of claim 21 wherein said event is a vehicle riding.

23. The system of claim 21 wherein said event is a person walking.

24. The system of claim 21 wherein said event is caused by the presence of birds.

25. The system of claim 21 wherein said event is caused by an external event.

26. The system of claim 21 further including alarming means for alerting if said event is identified as an intrusion.

27. The system of claim 21 further including:

- at least two fibre optic strands, wherein said fibre optic strands are installed in said protected zone;

- a transmitter device for transmitting a laser beam through said fibre optic strands;

- a receiver device for measuring the differential output in said beams transmitted through said fibre optic strands; and

- a calculating device for identifying intrusions whenever said measured differential in said output beams exceeds a predetermined threshold.