CN116980520A - Fraudulent telephone identification method, system and medium based on call echo delay - Google Patents

Abstract

The invention relates to a method, a system and a medium for identifying fraudulent telephone based on call echo delay, and the method for identifying fraudulent telephone based on call echo delay comprises the following steps: playing a preset far-end reference signal at the first end of the conversation through a loudspeaker and sending the preset far-end reference signal to a microphone at the first end of the conversation; acquiring a near-end signal acquired by a microphone at the second end of the conversation, wherein the near-end signal comprises an echo signal of the far-end reference signal, environmental noise and a voice signal of a near-end speaker acquired by the microphone; calculating the echo time delay based on the near-end signal and the far-end reference signal; and if the echo time delay is not consistent with the distance information from the attribution of the first end to the second end, identifying the suspected fraudulent call. Compared with the prior art, the method has the advantages of high efficiency, quick response, low cost and easy implementation for identifying suspected fraudulent telephones, and meets the application requirements.

Description

Fraudulent telephone identification method, system and medium based on call echo delay

Technical Field

The present invention relates to the field of communication technology and fraudulent telephone identification. And more particularly to a method, system, and medium for identifying fraudulent calls based on call echo delay.

Background

With the rapid advancement of communication technology, traditional voice services are gradually moving towards IMS (IP Multimedia Subsystem, IP multimedia system) networks and VoIP (Voice over internet Protocl, voice over IP) transfer supporting VoLTE voice. This development has prompted the widespread use of packet-switched based VOIP network telephones. However, because the IP telephony gateway does not verify the calling party number sufficiently, it is difficult to recognize the number-changed call and to track the overseas call, which provides an opportunity for lawbreakers to conduct telecommunications fraud. In recent years, telecommunication network fraud is increasing, and properties of people suffer from huge losses to threaten social stability.

Current technology for identifying fraudulent telephones has focused mainly on the following: technology one: the identification technology based on the call number blacklist specifically comprises the following steps:

(1) Frequency of calls by the calling number: for the case that the same calling number frequently calls the called number in a short time, the number can be regarded as a suspected fraudulent phone, and the number is added to a blacklist.

(2) Call completing rate of calling number: for the case that the same calling number has almost no call completing rate after multiple calls, the calling number can also be regarded as a suspected fraudulent telephone, and the number is added into a blacklist.

(3) Called dispersion: for the case that the same calling number calls a plurality of irrelevant called numbers or the called numbers are frequently replaced, the number can be regarded as a suspected fraudulent phone and added into a blacklist.

(4) User identification or complaints: and establishing a suspected number blacklist according to the user feedback or complaint, and adding the number fed back or complaint by the user into the blacklist.

When the suspected number is added into the blacklist, the network switch loads the suspected number into the blacklist and performs matching when calling, if the matching is successful, the incoming call can be automatically intercepted, and voice, short message or USSD (Unstructured Supplementary Service Data) reminding message can be sent to the called user.

However, for a number change call, this solution is almost ineffective because the fraudster can avoid interception of the blacklist by changing the call number, and therefore, it needs to be used in combination with other solutions to improve the accuracy and efficiency of identifying fraudulent telephones.

And (2) a technology II: recognition techniques based on speech recognition and semantic analysis.

The technology based on voice recognition or the technology based on semantic analysis needs to record calls, or compare a voice sample library, or transfer voice texts, and search keywords and key numbers for semantic matching, so that large-capacity voice recognition equipment is required to be deployed, normal operation of the existing network cannot be influenced in operation, the technology is complex to realize, the cost is high, the recognition rate depends on the number of training data, and because the number of fraudulent telephones is relatively small, collecting enough fraudulent telephone data for training a model is a challenging task.

In addition, in the case where the language difference such as polyphones or dialects is large, there are cases where erroneous judgment is possible, and a large amount of collection of telephone data involves a problem of personal privacy.

And (3) a technology III: recognition techniques based on machine learning. The common machine learning models are classified into supervised, semi-supervised and unsupervised, and compared with the Light GBM, the clustering model, the isolated forest and the like which are typically applied to the anti-fraud field, the model cluster has the characteristics of considerable recognition rate, and the effect is better than that of the first and second methods and the efficiency is higher.

However, the research is focused on analysis and induction of conversation behaviors and features of users, a large amount of expert experience is needed to accumulate, the calculation capability is higher, a large amount of data is needed to train a model, the calculation result has weak interpretability, the integrated tree model can obtain the feature important weight of the whole dimension, but the feature weight of a sample level cannot be obtained, the interpretable reason of each selected feature corresponding to a person cannot be provided, and the influence of which features have on judging whether a specific single user is fraudulent cannot be known.

In the existing telecommunication fraud detection means, a better identification effect can be achieved on the behavior of obvious fraud features, but a large number of telephone fraud with stronger concealment, insignificant features and newer techniques appear, for example, in the scenes of arbitrary number changing of fraudulent numbers, the effect of the method is difficult to ensure. Further research on the characteristics of the fraudsters themselves and fraudsters over the road is still under investigation. These all present significant challenges to the telecommunications fraud detection efforts, requiring comprehensive use of a variety of techniques and methods, and adjustments and optimizations based on the actual situation.

In recent years, as national impact on national telecommunication fraud has increased, criminal partners have moved place of fraudulent telephone dialing to overseas to avoid the investigation by national law enforcement. To make fraudulent activity more confusing, criminals use GOIP (Gsm Over Internet Protocol) and VOIP devices to make phone transfers in the country, camouflaging as a national number to conduct fraudulent activity. Such fraudulent telephones have long call paths and require multiple transfers, which can result in long end-to-end delay times during the call, poor call quality, and traffic and acoustic characteristics that are different from those of a real domestic telephone. However, the current communication protocol cannot accurately reflect the end-to-end delay of the multiple forwarding calls, and even the VoLTE and VoNR supporting the RTP/RTCP protocol can calculate the call delay, but the RTCP Round Trip Delay can not calculate the end-to-end delay because of the recalculation during the multiple forwarding calls (the characteristics of fraudulent calls), so that it is needed to realize the detection and the judgment of the abnormal incoming call by calculating the end-to-end delay.

Disclosure of Invention

The invention provides a method, a system and a medium for identifying fraudulent telephones based on call echo delay, which are used for solving the problems of low timeliness and low effective rate of identifying the existing telephone for identifying the overseas fraudulent telephone.

To achieve the above object, in a first aspect, the present invention provides a method for identifying a fraudulent call based on a call echo delay, including:

playing a preset far-end reference signal at the first end of the conversation through a loudspeaker and sending the preset far-end reference signal to a microphone at the first end of the conversation;

acquiring a near-end signal acquired by a microphone at the second end of the conversation, wherein the near-end signal comprises an echo signal of the far-end reference signal, environmental noise and a voice signal of a near-end speaker acquired by the microphone;

calculating the echo time delay based on the near-end signal and the far-end reference signal;

and if the echo time delay is not consistent with the distance information from the attribution of the first end to the second end, identifying the suspected fraudulent call.

Preferably, if the echo time delay is determined to be inconsistent with the distance information from the home location of the first end to the second end, the method identifies a suspected fraudulent call, specifically:

establishing an echo delay map based on a call delay test, wherein the echo delay map comprises an echo delay range corresponding to a call distance;

obtaining corresponding call distance information by comparing the echo delay map based on the echo delay;

and if the difference value between the distance information from the attribution of the first end to the second end of the caller ID display and the call distance information in the delay map corresponding to the echo delay is more than the preset percentage of the distance information corresponding to the caller ID display, identifying the caller ID display as a suspected fraudulent telephone, wherein the preset percentage is more than or equal to 30%.

Preferably, before calculating the echo delay based on the near-end signal and the far-end reference signal, the method further includes:

and the sound engine of the second end carries out echo cancellation on the near-end signal to obtain an echo cancellation signal, encodes the echo cancellation signal, and sends the encoded echo cancellation signal to the first end.

Preferably, the sound engine at the second end performs echo cancellation on the near-end signal to obtain an echo cancellation signal, which specifically includes:

separating the near-end signal from the echo signal by a filtering algorithm and Fourier transformation to obtain an echo cancellation signal;

the calculating the echo delay based on the near-end signal and the far-end reference signal specifically includes: and calculating echo time delay based on the echo signal and the far-end reference signal.

Preferably, the calculating the echo delay based on the near-end signal and the far-end reference signal specifically includes:

calculating the similarity between signals by using a cross correlation function, and determining the time delay by comparing the correlation between the near-end signal and the far-end reference signal and finding the maximum cross correlation peak value between the near-end signal and the far-end reference signal; or (b)

Calculating the time delay between signals by calculating the inverse Fourier transform of the product of the Fourier transform of the near-end signal and the far-end reference signal through a cross-correlation function method; or (b)

The time delay is calculated by comparing the phase difference in the frequency domain of the near-end signal and the far-end reference signal.

Preferably, when the corresponding call distance information is obtained based on the echo time delay and against the echo delay profile: if the echo signals are more than one, the longest echo time delay is determined to be matched with the distance information from the attribution of the first end to the second end of the caller identification.

Preferably, an echo delay map is established based on a call delay test, wherein the echo delay map comprises an echo delay range corresponding to a call distance, and the echo delay map specifically comprises:

based on the region, the call distance and the operator situation, a call delay experiment is carried out, and a call delay map is established, wherein the call delay map comprises call delay ranges corresponding to various call distances and call operator situations.

Preferably, the echo delay profile is established based on a call delay test, which specifically comprises:

obtaining an echo presumption distance based on the distance information between the attribution of the caller identification first end and the second end and the operator information, and estimating to obtain network exchange delay;

and setting the echo delay according to the echo presumption distance, the packet delay and the network switching delay to establish a call delay map, wherein the call delay map comprises call delay ranges corresponding to various call distances and call operator conditions.

Preferably, the obtaining the corresponding call distance information based on the echo delay against the echo delay profile specifically includes:

and comparing with the call delay map based on the echo time delay and the operator of the current call.

In a second aspect, the present invention also relates to a fraudulent telephone identification system based on a call echo delay, comprising: a sending unit, configured to play a preset far-end reference signal at a first end of a session through a speaker, and send the far-end reference signal to the first-end microphone;

the acquisition unit is used for acquiring a near-end signal acquired by a microphone at the second end of the conversation, wherein the near-end signal comprises an echo signal of the far-end reference signal, environmental noise and a voice signal of a near-end speaker acquired by the microphone;

a time delay calculation unit, configured to calculate the echo time delay based on the near-end signal and the far-end reference signal;

and the fraud identification unit is used for determining that the echo time delay is not consistent with the distance information from the attribution of the first end to the second end when the incoming call is displayed, and identifying the incoming call as a suspected fraudulent call.

In a third aspect, the present invention further relates to a computer readable storage medium having stored therein instructions that when executed perform a method of identifying a fraudulent call based on a call echo delay as described above.

The invention relates to a fraudulent telephone identification method, a system and a medium based on call echo delay, which have the following beneficial technical effects:

the echo is the only one end-to-end, has return, is not influenced by transfer and can be measured, and the invention utilizes the call echo delay to calculate and analyze, so that the actual distance between two parties of the call can be approximately restored. The invention adopts the conversation echo analysis technology, realizes the detection and judgment of abnormal calls by reversing the time delay of the end-to-end echo signals, and further identifies whether the incoming call comes from an overseas transfer call, thereby effectively detecting the overseas number-changing telecommunication fraud, not only having the advantage of real-time performance, but also being capable of overcoming the defect that the current mainstream mode based on the incoming call number blacklist early warning cannot identify the overseas number-changing telecommunication fraud.

Compared with the prior art, the method has the advantages of high efficiency, quick response, low cost and easy implementation for identifying suspected fraudulent telephones, and meets the application requirements.

Drawings

FIG. 1 is a flowchart of a method for identifying fraudulent calls based on a call echo delay according to a first embodiment of the present invention;

FIG. 2 is a flow chart of a method for identifying fraudulent telephone based on echo delay in a first embodiment of the present invention;

FIG. 3 is a schematic diagram of audio signal definition during communication according to a first embodiment of the present invention;

fig. 4 is a schematic diagram of the type of call echo according to the first embodiment of the present invention;

FIG. 5 is a system frame diagram of a second end of a conversation according to a first embodiment of the present invention;

FIG. 6 is a table of items for echo detection according to the G.168 proposal according to the first embodiment of the present invention;

fig. 7 is a schematic structural diagram of a fraudulent phone identification system based on call echo delay in a second embodiment of the present invention.

Detailed Description

The invention is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting thereof. It should be further noted that, for convenience of description, only some, but not all of the structures related to the present invention are shown in the drawings.

Example 1

Referring to fig. 1-6, a method for identifying fraudulent calls based on call echo delay is disclosed. A fraudulent telephone identification method based on a call echo delay according to this embodiment will be described below based on an example: the method is applied to the voice call process and comprises the steps S10-S50.

S10, playing a preset far-end reference signal at the first end of the conversation through a loudspeaker, and sending the preset far-end reference signal to a microphone at the first end of the conversation.

The first end of the conversation is the far end, and the second end of the conversation is the near end. The far-end reference signal is a predetermined known signal sequence x (n), which is known to both the first and second end, and is transmitted to the second end of the conversation via the microphone of the first end of the conversation via the voice communication network.

S20, acquiring a near-end signal acquired by a microphone at the second end of the conversation, wherein the near-end signal comprises an echo signal of a far-end reference signal, environmental noise and a voice signal of a near-end speaker acquired by the microphone.

As shown in fig. 2, after the echo signal y (n), i.e. the signal x (n) is played by the speaker at the second end of the conversation, the echo signal y (n) collected by the microphone at the second end cannot be equal to the signal x (n), and includes a linear reflection and a nonlinear reflection, where the linear superposition is denoted by y '(n), and the nonlinear superposition is denoted by y "(n), where y (n) =y' (n) +y" (n). s (n): the voice signals of the near-end speaker collected by the microphone, namely the signals which are really needed to be extracted and sent to the far-end; v (n): ambient noise, this portion of the signal is attenuated in the ANS; the near-end signal d (n), i.e. the original signal before echo cancellation after microphone acquisition, can be expressed as: d (n) =s (n) +y (n) +v (n).

S30, calculating echo time delay based on the near-end signal and the far-end reference signal.

The audio engine of the current telephony device is able to acquire a near-end signal d (n) and a far-end reference signal x (n). The talk echo delay estimation algorithm may be based on the time domain or the frequency domain.

In this embodiment, the time domain method is: the similarity between the signals is calculated using a cross correlation function, and the time delay is determined by comparing the correlation between the near-end signal and the far-end reference signal, finding the maximum cross correlation peak between them.

The frequency domain method comprises the following steps: calculating the time delay between signals by calculating the inverse Fourier transform of the product of the Fourier transforms of the near-end signal and the far-end reference signal through a cross-correlation function method; or (b)

The delay is calculated by comparing the phase difference in the frequency domain of the near-end signal and the far-end reference signal.

In some other embodiments, prior to S30, S31:

the sound engine of the second end carries out echo cancellation on the near-end signal to obtain an echo cancellation signal, the echo cancellation signal is coded, and the coded echo cancellation signal is sent to the first end so as to cancel the influence of echo on the call quality.

The sound engine at the second end performs echo cancellation on the near-end signal to obtain an echo cancellation signal, which specifically includes:

and separating the near-end signal from the echo signal by a filtering algorithm and Fourier transformation to obtain an echo cancellation signal. S30 specifically comprises: and calculating the echo time delay based on the echo signal and the far-end reference signal.

Of course, in step S30, test analysis may be performed with a headset and VOIP ATA that are commonly available in the market, and the test item is an existing algorithm for performing echo cancellation test according to the g.168 proposal, as shown in the table of fig. 6, and the echo detection algorithm is further optimized according to the test result.

And S40, if the echo time delay is not consistent with the distance information from the attribution of the first end to the second end of the caller identification, identifying the caller identification as a suspected fraudulent call.

If the echo is more than one, the longest echo time delay is determined to be matched with the distance information from the attribution of the first end to the second end of the caller identification. As shown in fig. 4, if the GOIP/VoIP is switched using a fixed telephone line, an electrical echo caused by the GOIP ATA and an end-to-end acoustic echo are simultaneously generated within the network, and the lengths of the two echoes are significantly different, so that the receiving end detects the echo, and if the electrical echo and the end-to-end echo occur simultaneously, the two echoes are detected. If an electrical echo is necessarily generated by the in-home GOIP device, the end-to-end echo is overseas to in-home, i.e. the interval between the two echoes is longer. For example, the delay of the conventional fixed telephone adopting the G.711/PCM A/U coding algorithm is only 125us, and the delay of the G723/1 algorithm commonly used by the GOIP switching fixed telephone is up to 67.5ms.

Specifically, in some embodiments, the echo time delay may be far beyond a time delay threshold corresponding to the distance information from the preset first end home location to the second end, and generally, the echo time delay should be greater than more than thirty percent of the distance information from the preset first end home location to the second end, if the echo time delay is not matched, the suspected fraudulent call is identified.

Typically, the fixed network telephone is one-way 50ms delay, the mobile telephone is one-way 300-600 ms, the network telephone is one-way 500-600 ms, and the end-to-end delay for the overseas call is around 1248 ms. Thus, even if the Burmese calls the transfer telephone in the border, the round trip is still about 600ms more than the normal telephone, if the incoming call displays the domestic number, if the echo delay time is obviously longer than the echo delay time of the domestic call, the suspected fraudulent telephone is identified. The echo delay is typically found to be an abnormal end-to-end delay, and is considered a rogue phone.

In other embodiments, steps S41-43 are employed as shown in FIG. 2.

S41, an echo delay map is established based on the call delay test, wherein the echo delay map comprises an echo delay range corresponding to the call distance.

In some embodiments, a call delay experiment is performed based on the region, the call distance and the operator situation, and a call delay map is established, where the call delay map includes call delay ranges corresponding to various call distances and call operator situations.

The method specifically comprises the following steps: s411, obtaining echo estimated distance based on the distance information of the attribution of the caller display first end and the second end and the operator information, and estimating network exchange delay.

S412, setting echo delay according to the echo estimated distance and the network exchange delay to establish a call delay map, wherein the call delay map comprises call delay ranges corresponding to various call distances and call operator conditions.

The echo delay can be calculated and estimated based on the following formula 1 according to the echo estimation distance and the network switching delay, so as to establish a call delay map.

At present, a fraudulent telephone adopts an environment external transfer access environment, a mobile phone incoming call generally displays an incoming call number, a system can infer a two-place call distance according to the attribution place of the incoming call number, and because optical cable routes of operators at different places are different, each voice coding algorithm can cause certain packet delay, and thus, a call experiment is carried out by considering the factors so as to determine a call delay map. Because the fraudulent party needs at least one transfer, i.e. re-decoding the code and re-accessing the China telecom operator network, the communication time is greatly prolonged. As can be seen from the figure, in packet-switched communication networks, the main sources of delay include:

(1) The encoder (processing) Delay of the communication gateway is a fixed Delay, typically around 20 milliseconds.

(2) The packet delay (Packetization Delay) is the time it takes to fill the packet payload with encoded/compressed speech, and is a fixed delay, typically around 30 milliseconds in length.

(3) The serialization delay (Serialization Delay) refers to the fixed delay required to convert voice or data frame bits into a network interface serial data stream. In China's communication backbone gigabit (Gbps) MPLS network, the serialized delay data is usually very small and negligible.

(4) Queuing/buffering delay (Output Queuing Delay) refers to the time that it takes to wait for a period of time for a packet to be processed and sent to an output port when it arrives at an input port of a network device, and is a variable delay, typically on the order of hundreds of microseconds to milliseconds in duration. As the number of switching nodes through which packets pass increases, the queuing delay time increases exponentially.

(5) The network switching Delay (Switch Delay) refers to the time interval between when a Switch receives a packet and when it starts to replicate the packet to the destination port. Different forwarding techniques can affect the size of the delay, typically between 10 and 50 milliseconds (ms). As the number of switching nodes through which packets pass increases, the network switching delay time increases exponentially.

(6) The De-Jitter Buffer (De-Jitter Buffer) is used for eliminating adverse effects caused by factors such as network side packet loss, disorder, repeated packets and the like at the cost of increasing delay at the receiving end. A nominal delay setting of 40 milliseconds is typically used.

(7) Propagation delay (propagation delay) refers to the time required for a signal to travel from a source to a destination, propagation delay = transmission distance/propagation speed. In public frame relay or ATM networks, the propagation speed estimate typically used is 10 microseconds/mile (i.e., 6 microseconds/kilometer) (g.114). In satellite communication, the propagation delay is about 270 milliseconds (30 km/s) from the ground to the satellite and back to the ground of a radio wave, because the satellite is at a high altitude of 36000 km.

Network delay estimation: the total delay estimate of the end-to-end one-way call for the network communication is as in equation 1,

one-way call total delay = encoder delay + packet delay + serialization delay + n x queuing buffer delay + n x network switching delay + debounce delay + propagation delay (equation 1)

From the above formula, it can be seen that the total call delay is positively correlated with the call distance (the number of switching nodes passing through), and the fixed-line telephone is typically a unidirectional 50ms delay, the mobile telephone is unidirectional 300-600 ms, and the network telephone is unidirectional 500-600 ms.

Taking a foreign place to dial domestic telephone as an example, the optical cable distance is calculated according to 4800 km, the round trip transmission time is more than that of a nearby telephone of 800 km in China by about 48 ms, the total echo (round trip) delay is about 300×2+48=648 seconds, the above is the delay calculation of the end-to-end call, and as the fraudulent party needs at least one transfer, i.e. re-decoding the code and re-accessing the network of the China telecom operator, the communication time is greatly prolonged, as shown in fig. 7, the end-to-end delay of the foreign call is: 300 (first-segment VoIP) ×2+ is about 300 (domestic operator VoLTE or VoNR) ×2+48 (delay of transmission distance) =around 1248 ms. Thus, even if the overseas site makes a call to the border, the round trip is still about 600 milliseconds more than the normal call. A particularly significant problem of very long echo from an outbound call is the focus of the present invention. S50, obtaining corresponding call distance information based on echo time delay and against an echo delay map.

S42, obtaining corresponding call distance information based on the echo time delay and the echo delay map.

Specifically, based on the echo time delay and the operator of the current call, the corresponding call distance information is obtained by comparing with the call delay map.

S43, if the difference value between the distance information from the attribution of the first end to the second end of the caller ID and the call distance information in the delay map corresponding to the echo delay is more than the preset percentage of the distance information corresponding to the caller ID, the suspected fraudulent call is identified. Wherein the preset percentage is more than or equal to 30 percent. The caller identification corresponding distance information is the distance information from the attribution of the caller identification first end to the second end.

The actual distance between the two parties of the call can be approximately restored by utilizing the call echo delay to calculate and analyze; and further identifies whether the incoming call comes from an overseas transfer phone, thereby effectively detecting the overseas telecommunications fraud and reminding the user of answering cautiously. A distance that generally matches the atlas far beyond the distance information of the first end home to the second end is considered a fraudulent phone.

Example two

As shown in fig. 7, the present invention further relates to a fraud phone identification system based on call echo delay, which can be implemented by a computing device with a central processing unit, such as a personal computer, an upper computer, a server, etc., and implemented by a speech engine, and includes a sending unit 50, an obtaining unit 60, a delay calculating unit 70, and a fraud identification unit 80.

And a transmitting unit 50, configured to play a preset far-end reference signal at the first end of the session through a speaker, and transmit the far-end reference signal to the first-end microphone.

The acquisition unit 60 is configured to acquire a near-end signal acquired by a microphone at a second end of the conversation, where the near-end signal includes an echo signal of a far-end reference signal, an ambient noise, and a voice signal of a near-end speaker acquired by the microphone;

a delay calculation unit 70 for calculating an echo delay based on the near-end signal and the far-end reference signal;

and the fraud identification unit 80 is configured to identify as a suspected fraudulent call if the echo time delay is not consistent with the distance information from the home location of the first end to the second end of the caller identification.

Wherein the fraud recognition unit 80 specifically comprises:

a spectrum establishment subunit 81, configured to establish an echo delay spectrum based on a call delay test, where the echo delay spectrum includes an echo delay range corresponding to a call distance;

a call distance obtaining subunit 82, configured to obtain corresponding call distance information based on the echo delay versus the echo delay profile;

and the judging unit 83 is configured to judge that the distance information from the home location of the first end to the second end displayed by the incoming call does not accord with the call distance information in the delay profile corresponding to the echo delay, and identify the call as a suspected fraudulent call.

The implementation process and effects of the method for obtaining parameters in the fraudulent use of telephone identification system based on call echo delay in this embodiment are the same as those of the method for obtaining parameters in the fraudulent use of telephone identification system based on call echo delay described in the first embodiment, and are not described here again, and the modeling and calculation process is completed with reference to the above-described fraudulent use of telephone identification system based on call echo delay.

Example III

The invention relates to a computer readable storage medium, the storage medium stores instructions, the instructions execute a method for identifying fraudulent call based on call echo delay when running, the method and effect of the implementation process are the same as the method for identifying fraudulent call based on call echo delay described in the first embodiment, and the description is omitted here.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.

Claims (11)

1. A method for identifying fraudulent telephone based on call echo delay, comprising:

playing a preset far-end reference signal at the first end of the conversation through a loudspeaker and sending the preset far-end reference signal to a microphone at the first end of the conversation;

acquiring a near-end signal acquired by a microphone at the second end of the conversation, wherein the near-end signal comprises an echo signal of the far-end reference signal, environmental noise and a voice signal of a near-end speaker acquired by the microphone;

calculating the echo time delay based on the near-end signal and the far-end reference signal;

and if the echo time delay is not consistent with the distance information from the attribution of the first end to the second end, identifying the suspected fraudulent call.

2. The method for identifying fraudulent use of telephone according to claim 1, wherein said determining that said echo delay is not identical to the distance information indicating the location of said first end to said second end is identifying as a suspected fraudulent use of telephone, specifically:

establishing an echo delay map based on a call delay test, wherein the echo delay map comprises an echo delay range corresponding to a call distance;

obtaining corresponding call distance information by comparing the echo delay map based on the echo delay;

and if the difference value between the distance information from the attribution of the first end to the second end of the caller ID display and the call distance information in the delay map corresponding to the echo delay is more than the preset percentage of the distance information corresponding to the caller ID display, identifying the caller ID display as a suspected fraudulent telephone, wherein the preset percentage is more than or equal to 30%.

3. The method for identifying fraudulent telephone of claim 1, wherein before said calculating said echo delay based on said near-end signal and said far-end reference signal, further comprises:

and the sound engine of the second end carries out echo cancellation on the near-end signal to obtain an echo cancellation signal, encodes the echo cancellation signal, and sends the encoded echo cancellation signal to the first end.

4. A method for identifying a fraudulent call based on a call echo delay according to claim 3, wherein the sound engine of the second end performs echo cancellation on the near-end signal to obtain an echo cancellation signal, and specifically includes:

separating the near-end signal from the echo signal by a filtering algorithm and Fourier transformation to obtain an echo cancellation signal;

the calculating the echo delay based on the near-end signal and the far-end reference signal specifically includes: and calculating echo time delay based on the echo signal and the far-end reference signal.

5. A method for identifying a fraudulent call based on a call echo delay according to claim 1,

based on the near-end signal and the far-end reference signal, the echo time delay is calculated, which specifically comprises:

calculating the similarity between signals by using a cross correlation function, and determining the time delay by comparing the correlation between the near-end signal and the far-end reference signal and finding the maximum cross correlation peak value between the near-end signal and the far-end reference signal; or (b)

Calculating the time delay between signals by calculating the inverse Fourier transform of the product of the Fourier transform of the near-end signal and the far-end reference signal through a cross-correlation function method; or (b)

The time delay is calculated by comparing the phase difference in the frequency domain of the near-end signal and the far-end reference signal.

6. The method for identifying fraudulent telephone call based on echo delay of claim 1, wherein when obtaining corresponding call distance information based on said echo delay map against said echo delay map: if the echo signals are more than one, the longest echo time delay is determined to be matched with the distance information from the attribution of the first end to the second end of the caller identification.

7. The method for identifying fraudulent use of telephone according to claim 2, wherein an echo delay map is established based on a call delay test, and the echo delay map includes an echo delay range corresponding to a call distance, and the method specifically includes:

based on the region, the call distance and the operator situation, a call delay experiment is carried out, and a call delay map is established, wherein the call delay map comprises call delay ranges corresponding to various call distances and call operator situations.

8. The method for identifying fraudulent use of telephone according to claim 7, wherein the echo delay map is established based on a call delay test, and specifically comprises:

obtaining an echo presumption distance based on the distance information of the attribution of the caller identification first end and the second end and the operator information, and estimating to obtain network exchange delay;

and setting the echo delay according to the echo presumption distance and the network switching delay to establish a call delay map, wherein the call delay map comprises call delay ranges corresponding to various call distances and call operator conditions.

9. The method for identifying fraudulent use of telephone according to claim 7, wherein said obtaining corresponding call distance information based on said echo delay versus said echo delay profile is specifically:

and comparing with the call delay map based on the echo time delay and the operator of the current call.

10. A fraudulent telephone identification system based on call echo delay, characterized in that: comprising the following steps:

a sending unit, configured to play a preset far-end reference signal at a first end of a session through a speaker, and send the far-end reference signal to the first-end microphone;

the acquisition unit is used for acquiring a near-end signal acquired by a microphone at the second end of the conversation, wherein the near-end signal comprises an echo signal of the far-end reference signal, environmental noise and a voice signal of a near-end speaker acquired by the microphone;

a time delay calculation unit, configured to calculate the echo time delay based on the near-end signal and the far-end reference signal;

and the fraud identification unit is used for determining that the echo time delay is not consistent with the distance information from the attribution of the first end to the second end when the incoming call is displayed, and identifying the incoming call as a suspected fraudulent call.

11. A computer storage medium having stored therein instructions that when executed perform a method of identifying fraudulent telephones based on call echo delay according to any one of claims 1 to 9.