WO2009111957A1 - 一种线路状态检测方法、装置和预测式外呼系统 - Google Patents

一种线路状态检测方法、装置和预测式外呼系统 Download PDF

Info

Publication number
WO2009111957A1
WO2009111957A1 PCT/CN2009/070334 CN2009070334W WO2009111957A1 WO 2009111957 A1 WO2009111957 A1 WO 2009111957A1 CN 2009070334 W CN2009070334 W CN 2009070334W WO 2009111957 A1 WO2009111957 A1 WO 2009111957A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
tested
information
music
zero
Prior art date
Application number
PCT/CN2009/070334
Other languages
English (en)
French (fr)
Inventor
李瑞华
莫运能
冯蔚腾
万木杨
关少宁
覃景繁
唐繁荣
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2009111957A1 publication Critical patent/WO2009111957A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M1/00—Substation equipment, e.g. for use by subscribers
    • H04M1/82—Line monitoring circuits for call progress or status discrimination
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use

Definitions

  • the present invention relates to the field of telephone communication technologies, and in particular, to a line state detection method, a line state detection device, and a predictive outbound system.
  • the predictive outbound system will automatically filter out some invalid answering telephone numbers, such as filtering out the wrong number such as the empty number, answering machine, modem and fax number.
  • the number such as filtering out the phone number of the called user busy.
  • the predictive outbound call system filters out the invalid answering phone number by means of signaling detection.
  • the specific implementation process may be as follows: The called side detects the busy, empty number, invalid and other signaling of the called user, and The single-double-frequency pulse signal such as the busy tone and the ring-back tone is played to the calling side; the calling side performs the line state detection according to the characteristics of the single-double-frequency pulse signal such as the busy tone and the ringback tone, and reports the line state detection result, thereby reporting according to the report.
  • the line status detection result can determine the subsequent operation; if the line status detection result is the inactive line status of the called user busy, empty number, etc., the subsequent connection operation is not performed to filter out the invalid answering telephone number;
  • the status detection result is the normal line status such as the normal ring back tone (that is, the called telephone number is valid, the line is normal, and the called user has not answered the call), and then the subsequent connection operation is performed, such as connecting the agent, so that the agent conducts product promotion and the client Return visits and other services.
  • the inventors found that: the pleasing music ring back tone is gradually replacing the ring back tone of the traditional single dual frequency pulse signal, and the called side of some communication systems is no longer to the calling side.
  • the wrong phone number filtering operation was generated.
  • the existing predictive outbound system needs to be further improved.
  • Embodiments of the present invention provide a line state detection method, apparatus, and predictive outbound system.
  • the signal in the line includes a music signal ⁇ , which can accurately identify the state of the line, thereby enabling normal connection operation and improved filtering.
  • the accuracy of the invalid answering call improves the predictive outbound system.
  • the receiving module 310 is configured to receive a signal to be tested in the communication line
  • the acquiring module 320 is configured to acquire signal feature information of the signal to be tested received by the receiving module 310.
  • the first detecting module 330 is configured to: acquire signal feature information of the signal to be tested acquired by the acquiring module 320 and preset music signal characteristics The information is compared. If the result of the comparison is that the signal characteristic of the signal to be tested conforms to the characteristics of the music signal, the information indicating that the line state is valid is output.
  • the predictive outgoing call system provided by the embodiment of the present invention includes the above-described line state detecting device.
  • the music signal in the telephone communication line can be effectively recognized, for example, the music ring back tone can be effectively recognized, so that the signal in the telephone communication line is a music signal.
  • Figure 1 is a schematic diagram showing the comparison of the characteristics of the voice signal and the characteristics of the music field
  • 2 is a flow chart of detecting a music ringback tone according to an embodiment of the present invention
  • 3 is a schematic diagram of a line state detecting device according to an embodiment of the present invention
  • FIG. 4 is a schematic diagram of a predictive outgoing call system according to an embodiment of the present invention.
  • Music signals have different signal characteristics relative to other types of signals such as voice response signals, single-double-frequency pulse signals, for example, in the ⁇ domain, music signals have unique characteristics; for example, in the frequency domain On, the music signal also has unique characteristics. Therefore, by comparing the signal to be tested in the communication line with the signal characteristic of the music signal, it can be identified whether the signal to be tested is a music signal.
  • the embodiment of the present invention uses the signal characteristics of the music signal to realize line state detection.
  • the signal to be tested in the communication line is received, and the signal to be tested is a response signal after dialing the called telephone number.
  • the embodiment of the present invention can receive the signal to be tested in the communication line on the calling user side, or receive the signal to be tested in the communication line at a network device on the network side, such as receiving the signal to be tested at the gateway.
  • the embodiment of the present invention can obtain the ⁇ domain feature information of the received signal to be tested after acquiring the signal characteristic information of the received signal to be tested.
  • the frequency domain characteristic information is used to obtain signal characteristic information of the received signal to be tested.
  • only the received domain information of the signal to be tested may be obtained, or only the frequency domain information of the signal to be tested may be acquired, and the domain information of the received signal to be tested may be acquired and the frequency domain information may be obtained.
  • the domain characteristic information, such as energy change information of the adjacent signal to be tested, the frequency domain feature information may include any one of zero-crossing rate information and frequency component change information, and may also include zero-crossing rate information and frequency component change information.
  • the embodiment of the present invention can also obtain the energy change of the adjacent signal to be tested. Other information, such as obtaining the average value of the energy change of the received adjacent signal to be tested, and the like. Moreover, embodiments of the present invention do not limit the specific implementation process of acquiring energy change information of adjacent signals to be tested.
  • Obtaining the zero-crossing rate information of the received signal to be tested such as obtaining the maximum value of the zero-crossing rate change of the adjacent signal to be tested received in the predetermined inter-day interval, and acquiring the signal to be tested received in the predetermined inter-day interval, for example.
  • the zero crossing rate average may only use the maximum value of the zero-crossing rate change, or may only If you use the zero-crossing average, you can also use the zero-crossing rate change maximum and the zero-crossing rate average.
  • the embodiment of the present invention does not limit the specific implementation process of obtaining the zero-crossing rate information of the signal to be tested.
  • the embodiment of the present invention can directly compare the acquired signal feature information with the preset music signal feature information. Considering the existence of a single dual-frequency pulse signal, a music signal and a voice signal coexisting in the current communication network, the embodiment of the present invention can first identify the single-double-frequency pulse signal, and recognize that the received signal is not a single pair. After the frequency signal, the acquired signal feature information is compared with the preset music signal feature information. Of course, the process of identifying a single dual frequency signal and the process of comparing with the preset music signal characteristic information can also be performed simultaneously. The embodiments of the present invention can identify a single dual-frequency pulse signal by using existing technical means. The embodiment of the present invention does not limit the specific implementation process of identifying a single dual-frequency pulse signal.
  • the embodiment of the present invention compares the signal characteristic information of the signal to be tested with the preset music signal characteristic information to determine whether the signal characteristic of the signal to be tested conforms to the characteristics of the music signal.
  • the pre-set music signal characteristic information here may be for energy change, or for zero-crossing rate, and may be for both energy change and zero-crossing rate.
  • the preset music signal characteristic information may be any one or any two or three of an energy change maximum threshold, a zero-crossing rate change maximum threshold, and a zero-crossing rate average threshold.
  • any of the signal characteristics may be met in advance.
  • the characteristic of the music signal is set to determine that the signal characteristic of the signal to be tested conforms to the characteristics of the music signal, that is, all the signal characteristics of the signal to be tested do not conform to the characteristics of the music signal, and it is determined that the signal characteristics of the signal to be tested do not conform to the characteristics of the music signal. .
  • the signal characteristic of the signal to be tested conforms to the preset music signal feature, it can be determined that the signal in the line is a music signal, so that the music ring back tone (ie, the color ring tone) can be successfully detected.
  • the signal in the line is a music signal
  • the information of the line state can be outputted, so as to facilitate the subsequent normal connection operation.
  • the line After judging that the signal characteristic of the signal to be tested does not conform to the preset characteristics of the music signal, the line can be determined.
  • the signal in the signal is not a music signal, such as a voice signal. Therefore, it is impossible to determine whether the line to which the called user is dialed is in a valid state, such as whether it is impossible to determine whether the line cannot be connected due to the called user not in the service area, shutdown, or the like.
  • the judgment result may be output instead of the information that the line state is valid.
  • the output judgment result may include a cause value.
  • the cause value is that the signal is not a music signal or the like.
  • the predictive external call system uses the signal to be tested and the frequency domain feature to identify the single dual frequency pulse signal, the music ring back tone (ie, the color ring tone) and the voice prompt tone as an example, and the present invention is implemented with reference to the accompanying drawings.
  • the line state detection method provided by the method is described in detail.
  • the signals in the communication line mainly include single dual frequency pulse signal, music ring back tone (ie ring tones) and voice signal.
  • Single-duplex pulse signals such as traditional busy tone, ring back tone, etc.
  • Voice signals such as voice prompts.
  • the predictive outbound system needs to accurately identify single and double frequency pulse signals, music ring back tones and voice signals.
  • the predictive outbound system can easily identify single and dual frequency pulse signals using existing techniques.
  • the main technical problems that need to be solved in the predictive outbound system are: how to accurately identify the music ringback tone and the voice signal.
  • the predictive outbound system uses the domain characteristics and frequency domain characteristics of the signal to identify the music ringback tone and voice signal.
  • the upper part of Fig. 1 is the ⁇ domain feature of the speech signal
  • the lower part of Fig. 1 is the ⁇ domain characteristic of the music signal. Comparing the characteristics of the upper and lower parts of the region, we can see that there are differences in the energy changes between the temporal characteristics of the speech signal and the temporal characteristics of the music signal. That is to say, the speech signal has a characteristic of smoothness within a short period of time, for example, the energy variation of the signal is relatively small within 30 to 40 ms. However, in the long inter-segment, there is bound to be a phenomenon of low energy and high energy interaction, that is, the signal energy changes are more obvious. The music signal tends to be stable over long periods of time.
  • the difference between the frequency domain characteristics of the speech signal and the music signal includes:
  • the frequency component of the music signal is richer than the frequency component of the speech signal, and the high frequency component exists in the music signal, and the frequency component of the music signal follows the time.
  • the change is small, which is mainly reflected in the zero-crossing rate. That is to say, the zero-crossing rate of the speech signal is relatively large with the change of the daytime, and the zero-crossing rate of the music signal is relatively small with the change of the daytime.
  • the average zero-crossing rate of the music signal is relatively high, and the average zero-crossing rate of the speech signal is relatively low.
  • the zero-crossing rate average value is compared with the preset energy change maximum threshold, the zero-crossing rate change maximum threshold, and the zero-crossing rate average threshold to determine whether the signal to be tested is a music ringback tone.
  • FIG. 2 A specific example of detecting the music ringback tone flow is shown in Fig. 2.
  • step 200 the music ringback tone detection process is started, and step 205 is performed.
  • Step 205 Calculate the energy Eng (n) of the current frame on the ⁇ domain.
  • the current frame may be a signal of a signal to be tested in the communication line, and the signal obtained by the sample signal, such as every 30 ms or 40 ms, is passed to step 210.
  • Step 210 Calculate a zero-crossing rate Zc (n) of the current frame in the frequency domain, and go to step 215.
  • Step 215 Determine whether the energy of the current frame Eng (n) is greater than or equal to the energy threshold SilceThreshold (eg, 30 dB). If Eng (n) is greater than or equal to the energy threshold SilceThreshold, determine that the current frame is a voice signal or a music signal, to Step 220.
  • the energy threshold SilceThreshold eg, 30 dB
  • Eng (n) is greater than or equal to the preset energy threshold SilceThreshold
  • the inter-counter counter is judged to start counting, the frame counter is reset, and the zero-crossing rate sum is set to zero, which is the energy change maximum value EngChMax and The zero-crossing rate change maximum value ZcChMax sets an initial value; if Eng (n) is less than the energy threshold SilceThreshold, it is determined to be a mute signal, to step 225.
  • the mute signal here can mean no sound. It should be noted that, in step 215, greater than or equal to may also be greater than, and then, less than or equal to less than or equal to in step 215.
  • Step 220 the frame counter is accumulated, and the zero crossing rate is accumulated. That is, the count value of the frame counter is increased by one.
  • Step 235 Determine whether the absolute value of the adjacent frame energy change value for the current frame is greater than the maximum value of the energy change EngChMax, if EngCh>EngChMax, to step 240, otherwise, to step 245
  • step 235 may be greater than or equal to.
  • Step 240 Update the energy change maximum value EngChMax with the absolute value of the adjacent frame energy change value EngCh for the current frame, and then go to step 245.
  • Step 245 determining whether the absolute value of the zero-crossing rate change value ZcCh (n) for the current frame is greater than the zero-crossing rate change maximum value ZcChMax, if ZcCh (n) > ZcChMax, to step 250, otherwise, to step 25 5. It should be noted that the greater than or equal to the value in step 245 may be greater than or equal to.
  • Step 250 Update the zero-crossing rate change maximum value ZcChMax with the absolute value of the zero-crossing rate change value ZcCh (n) for the current frame, and then go to step 255.
  • Step 255 It is judged whether the count value of the decision daytime counter reaches the predetermined time. If the predetermined time is reached, the process goes to step 260, otherwise, the process returns to step 205.
  • Step 265 determining whether the energy change maximum value EngChMax is greater than or equal to the preset energy change maximum value threshold EngThreshold, if EngChMax is greater than or equal to EngThreshold, then to step 280, no shell I", to step 270. It should be noted that, in step 265, greater than or equal to may also be greater than.
  • Step 270 Determine whether the zero-crossing rate change maximum value ZcChMax is greater than or equal to the preset zero-crossing rate change maximum threshold ZcThreshold, if ZcChMax is greater than or equal to ZcThreshold, then go to step 280, no, and go to step 275. It should be noted that the greater than or equal to the value in step 270 may also be greater than.
  • Step 275 determining whether the zero-crossing rate average value ZcAverage is greater than or equal to the preset zero-crossing rate average threshold ZcAvgThreshold, if ZcAverage is greater than or equal to ZcAvgThreshold, then to step 280, no shell I", to step 285. It should be noted that the greater than or equal to in step 275 may also be greater than.
  • Step 280 determining that the signal in the line is a music ringback tone, so that the line state is determined to be valid, to step 285.
  • Step 285 the music ringback tone detection process ends.
  • the energy change maximum threshold, the zero-crossing rate change maximum threshold, the zero-crossing rate average threshold, and the predetermined inter-turn time corresponding to the inter-turn counter can be set according to actual needs, such as a predetermined time. Set to 3 seconds, etc.
  • the detection algorithm of the embodiment of the present invention can accurately detect the music ring back tone when the amount of calculation is low and the complexity is small. Thereby improving the normal connection accuracy of the predictive outbound system.
  • steps 220, 225, 230 can be adjusted, and steps 220, 225, and 2 30 can be performed concurrently.
  • steps 235 and 245 may be performed concurrently, or steps 245 and 250 may be performed prior to steps 235 and 240.
  • the implemented process may be:
  • Step 230 Calculate a change value of the zero-crossing rate of the adjacent frame for the current frame: ZcCh (n
  • Step 245 determining whether the absolute value of the zero-crossing rate change value ZcCh (n) for the current frame is greater than the zero-crossing rate change maximum value ZcChMax, if ZcCh (n) > ZcChMax, to step 250, otherwise, to step 23
  • Step 250 Update the zero-crossing rate change maximum value ZcChMax with the absolute value of the zero-crossing rate change value ZcCh (n) for the current frame, and then go to step 235.
  • Step 235 Determine whether the absolute value of the adjacent frame energy change value for the current frame is greater than the energy change maximum value EngChMax, if EngCh>EngChMax, to step 240, otherwise, to step 255
  • Step 240 Update the energy change maximum value EngChMax with the absolute value of the adjacent frame energy change value EngCh for the current frame, and then go to step 255.
  • the line state detecting device can be set It is placed on the calling user side, and can also be set on the network side, such as at the gateway.
  • the line state detecting device can be an independent device or can be installed in an existing device.
  • each module in the line state detecting device may be set in different devices or may be disposed in the same device.
  • the line state detecting device 300 includes: a receiving module 310, an obtaining module 320, and a first detecting module 330.
  • the line state detecting device 300 may also optionally include a second detecting module 340.
  • the receiving module 310 receives the signal to be tested in the communication line.
  • the signal to be tested received by the receiving module 310 is a response signal after dialing the called telephone number.
  • the acquisition module 320 acquires the signal characteristics of the signal to be tested received by the receiving module 310.
  • the signal characteristics acquired by the acquisition module 320 can be output to the first detection module 330 and the second detection module 340.
  • the acquiring module 320 may acquire the domain characteristic information and the frequency domain feature information of the signal to be tested to obtain signal characteristic information of the signal to be tested.
  • the obtaining module 320 may obtain only the domain information of the signal to be tested, or may acquire only the frequency domain information, and may acquire both the domain information of the signal to be tested and the frequency domain information.
  • the acquisition module 320 can include any one or two of the domain acquisition sub-module 321 and the frequency domain acquisition sub-module 322.
  • the domain acquisition sub-module 321 acquires the domain feature information of the signal to be tested received by the receiving module 310, for example, the domain acquisition sub-module 321 acquires energy change information of the adjacent signal.
  • the implementation manner may be: the ⁇ domain acquisition submodule 321 obtains the maximum value of the energy change of the adjacent signal to be tested received in the predetermined inter-turn interval.
  • the ⁇ domain acquisition sub-module 321 can also obtain other information about the energy change of the adjacent signal to be tested, such as obtaining the average value of the energy change of the received adjacent signal to be tested.
  • the embodiment of the present invention does not limit the specific implementation process of the domain acquisition sub-module 321 acquiring the energy change information of the adjacent signal to be tested.
  • the frequency domain acquisition sub-module 322 acquires frequency domain characteristic information of the signal to be tested received by the receiving module 310.
  • the frequency domain acquisition sub-module 322 can include any one or both of a zero crossing rate maximum unit 323 and an average zero crossing rate unit 324.
  • the zero-crossing rate maximum unit 323 acquires the maximum value of the zero-crossing rate change of the adjacent signal to be tested received by the receiving module 310 within the predetermined inter-turn interval.
  • the embodiment of the present invention does not limit the specific implementation process of the zero-crossing rate maximum value unit 323 to obtain the maximum value of the zero-crossing rate change of the adjacent signal to be tested.
  • the average zero-crossing rate unit 324 acquires the average zero-crossing rate of the signal to be tested received by the receiving module 310 within the predetermined inter-day interval.
  • the embodiment of the present invention does not limit the average zero-crossing rate unit 324 to obtain the average zero-crossing rate of the signal to be tested. The specific implementation process.
  • the information obtained by the above-mentioned domain acquisition sub-module 321, the zero-crossing rate maximum unit 323, and the average zero-crossing rate unit 324 can be output to the first detection module 330 and the second detection module 340.
  • the process of obtaining the sub-module 321 , the zero-crossing rate maximum unit 323 and the average zero-crossing rate unit 324 for obtaining information may be as described in steps 205 to 260 of FIG. 2 above.
  • the first detecting module 330 compares the received signal feature information with the preset music signal feature information, and determines, according to the comparison result, that the acquired signal feature meets the preset music signal feature, and outputs the line state valid information.
  • the preset music signal feature information may be stored in the first detecting module 303 or may be stored outside the first detecting module 330.
  • the preset music signal characteristic information may be any one or more of an energy change maximum value, a zero-crossing rate change maximum value, and a zero-crossing rate average value.
  • the first detecting module 330 may be at The signal characteristics of any of the received signals are in accordance with the characteristics of the music signal, and it is determined that the signal characteristics of the signal to be tested received by the receiving module 310 conform to the characteristics of the music signal, that is, all signal characteristics of the signal to be tested do not conform to the characteristics of the music signal. In other words, the first detecting module 330 determines that the signal feature of the signal to be tested received by the receiving module 310 does not meet the characteristics of the music signal. After determining that the signal characteristic of the signal to be tested received by the receiving module 310 conforms to the characteristics of the music signal, the first detecting module 330 may output information that the line status is valid, so as to perform subsequent normal connection operations.
  • the first detecting module 330 may determine that the signal in the line is not a music signal after determining that the signal characteristic of the received signal does not conform to the preset music signal characteristic. Thereafter, the first detecting module 330 can output information such as a non-music ring back tone.
  • the second detection module 340 identifies, according to the signal characteristic information acquired by the acquisition module 320, whether the signal to be tested received by the receiving module 310 is a single dual-frequency pulse signal, and after being identified as a single dual-frequency pulse signal, may be based on a single dual-frequency pulse.
  • the specific type of signal outputs information that the line state is valid or invalid, for example, information that the single-duplex pulse signal is inactive for the busy tone output line state.
  • the second detection module 340 can use the existing technology
  • the embodiment of the present invention does not limit the specific implementation process of the second detection module 340 to identify the single dual frequency pulse signal. It should be noted that the signal identification process of the first detection module 330 and the second detection module 340 may be performed simultaneously or sequentially.
  • the predictive outbound system of Figure 4 shows only the system control portion, the pager and the color ring detector, i.e., the line state detecting device 300.
  • the pager dials the phone number of the called user.
  • the line state detecting device 300 detects the signal to be tested transmitted in the telephone network, and detects the signal to be tested, and then outputs the detection result information to the system control portion, and the system control portion transmits the detection result information according to the line state detecting device 300. Decide on subsequent operations, such as whether to continue or not.
  • the line state detecting device 300 is as described in the above embodiment, and the description thereof will not be repeated here.

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Telephonic Communication Services (AREA)

Abstract

公开了线路状态检测方法、装置和预测式外呼系统。所述线路状态检测方法包括：接收通讯线路中的待测信号，获取接收的待测信号的信号特征信息，将待测信号的信号特征信息与预设置的音乐信号的时域特征和频域特征中的至少一个进行比较，如果待测信号的信号特征符合预设置的音乐信号特征，确定线路状态有效。在线路中的待测信号为音乐信号时，可准确识别出线路的状态，从而能够实现有效的接续操作，提高了滤除无效应答电话的准确率，完善了预测式外呼系统。

Description

说明书 一种线路状态检测方法、 装置和预测式外呼系统
[1] 本申请要求于 2008年 3月 10日提交中国专利局、 申请号为 200810006576.3、 发明 名称为 "一种线路状态检测方法、 装置和预测式外呼系统"的中国专利申请的优先 权， 其全部内容通过引用结合在本申请中。
[2] 技术领域
[3] 本发明涉及电话通讯技术领域， 具体涉及一种线路状态检测方法、 线路状态检 测装置和预测式外呼系统。
[4] 发明背景
[5] 预测式外呼系统在根据存储的用户电话号码进行拨号过程中， 会自动滤除一些 无效应答的电话号码， 如滤除掉空号、 电话答录机、 Modem和传真号码等错误 电话号码， 再如滤除掉被叫用户占线的电话号码等。
[6] 目前， 预测式外呼系统通过信令检测方式来滤除无效应答的电话号码， 具体实 现过程可以为： 被叫侧对被叫用户忙、 空号、 无效等信令进行检测， 并向主叫 侧播放相应的忙音、 回铃音等单双频脉冲信号； 主叫侧根据忙音、 回铃音等单 双频脉冲信号特征进行线路状态检测， 上报线路状态检测结果， 从而根据上报 的线路状态检测结果可以确定后续的操作； 如在线路状态检测结果为被叫用户 忙、 空号等无效线路状态， 则不进行后续的接续操作， 以滤除无效应答的电话 号码； 再如在线路状态检测结果为正常回铃音等有效线路状态 （即被叫电话号 码有效、 线路正常且被叫用户还没有接听） 后， 进行后续的接续操作， 如接通 座席， 使座席进行产品宣传、 客户回访等业务。
[7] 在实现本发明的过程中， 发明人发现： 悦耳的音乐回铃音正在逐渐替代传统的 单双频脉冲信号的回铃音， 有些通讯系统的被叫侧已经不再向主叫侧发送单双 频脉冲信号； 而现有的预测式外呼系统不能够针对音乐回铃音进行有效检测， 从而不能够实现正常的接续操作， 而且在回铃音为音乐回铃音的情况下会产生 错误的电话号码滤除操作。 现有的预测式外呼系统有待于进一步的完善。
[8] 发明内容 本发明实施方式提供一种线路状态检测方法、 装置和预测式外呼系统， 在线路 中的信号包含音乐信号吋， 可准确识别出线路的状态， 从而能够实现正常的接 续操作、 提高了滤除无效应答电话的准确率， 完善了预测式外呼系统。
本发明实施方式提供的线路状态检测方法， 包括：
接收通讯线路中的待测信号， 获取所述待测信号的信号特征信息；
将所述待测信号的信号特征信息与预设置的音乐信号特征信息进行比较； 如果所述比较的结果为所述待测信号的信号特征符合音乐信号特征， 确定线路 状态有效。
本发明实施方式提供的线路状态检测装置， 包括：
接收模块 310， 用于接收通讯线路中的待测信号；
获取模块 320， 用于获取所述接收模块 310接收的待测信号的信号特征信息； 第一检测模块 330， 用于将获取模块 320获取的待测信号的信号特征信息与预设 置的音乐信号特征信息进行比较， 如果所述比较的结果为待测信号的信号特征 符合音乐信号特征， 输出线路状态有效的信息。
本发明实施方式提供的预测式外呼系统包括有上述线路状态检测装置。
本发明实施方式提供的回铃音检测方法， 包括：
接收通讯线路中的待测信号， 获取所述待测信号的信号特征信息；
将所述待测信号的信号特征信息与预设置的音乐信号特征信息进行比较； 如果所述比较的结果为所述待测信号的信号特征符合音乐信号特征， 确定所述 待测信号为音乐回铃音。
通过上述技术方案的描述可知， 通过充分利用音乐的信号特征， 能够有效识别 出电话通讯线路中的音乐信号， 例如可以有效识别出音乐回铃音， 从而可在电 话通讯线路中的信号为音乐信号吋准确识别出线路的状态， 实现正常的接续操 作， 而且在回铃音为音乐回铃音的情况下不会产生错误的电话号码滤除操作， 提高了滤除无效应答电话的准确率， 完善了预测式外呼系统。
附图简要说明
图 1是语音信号吋域特征和音乐吋域特征对比示意图；
图 2是本发明实施方式的检测音乐回铃音流程图； [27] 图 3是本发明实施方式的线路状态检测装置示意图；
[28] 图 4是本发明实施方式的预测式外呼系统示意图。
[29] 实施本发明的方式
[30] 音乐信号相对于其它类型信号如语音应答信号、 单双频脉冲信号等信号来说， 具有不同的信号特征， 例如在吋域上， 音乐信号具有独特的特征； 再例如， 在 频域上， 音乐信号也具有独特的特征。 因此， 通过将通讯线路中的待测信号与 音乐信号的信号特征进行比较能够识别出待测信号是否为音乐信号， 本发明实 施方式正是利用音乐信号的信号特征实现线路状态检测的。
[31] 下面对本发明实施方式提供的线路状态检测方法进行说明。
[32] 首先， 接收通讯线路中的待测信号， 该待测信号如拨打被叫电话号码后的回应 信号。 本发明实施方式可以在主叫用户侧接收通讯线路中的待测信号， 也可以 在网络侧的某个网络设备处接收通讯线路中的待测信号， 如在网关处接收待测 信号等。
[33] 然后， 获取上述接收的待测信号的信号特征信息。 由于音乐信号在吋域和频域 方面均有独特的特征， 因此， 本发明实施方式在获取接收的待测信号的信号特 征信息吋， 可以釆用获取接收的待测信号的吋域特征信息、 频域特征信息来获 得接收的待测信号的信号特征信息。 本发明实施方式可以仅获取接收的待测信 号的吋域信息， 也可以仅获取待测信号的频域信息， 还可以既获取接收的待测 信号的吋域信息又获取频域信息。 吋域特征信息如相邻待测信号的能量变化信 息， 频域特征信息可以包括过零率信息、 频率成分变化信息中的任一项， 也可 以包括过零率信息和频率成分变化信息。
[34] 获取相邻待测信号的能量变化信息如获取预定吋间间隔内接收到的相邻待测信 号的能量变化最大值， 当然， 本发明实施方式也可以获取相邻待测信号能量变 化的其它信息， 如获取接收到的相邻待测信号的能量变化均值等。 而且， 本发 明实施方式不限制获取相邻待测信号的能量变化信息的具体实现过程。
[35] 获取接收的待测信号的过零率信息如获取预定吋间间隔内接收的相邻待测信号 的过零率变化最大值， 再如获取预定吋间间隔内接收的待测信号的过零率平均 值。 本发明实施方式中的过零率信息可以仅釆用过零率变化最大值， 也可以仅 釆用过零率平均值， 也可以同吋釆用过零率变化最大值和过零率平均值。 而且 ， 本发明实施方式不限制获取待测信号的过零率信息的具体实现过程。
[36] 在获取了待测信号的信号特征信息后， 本发明实施方式可以直接将获取的信号 特征信息与预设置的音乐信号特征信息进行比较。 考虑到目前的通讯网络中存 在单双频脉冲信号、 音乐信号和语音信号并存的情况， 因此， 本发明实施方式 可以先进行单双频脉冲信号的识别， 在识别出接收到的信号不是单双频信号后 ， 再将获取的信号特征信息与预设置的音乐信号特征信息进行比较。 当然， 识 别单双频信号的过程和与预设置的音乐信号特征信息进行比较的过程也可以同 吋进行。 本发明实施方式可以釆用现有的技术手段来识别单双频脉冲信号， 本 发明实施方式不限制识别单双频脉冲信号的具体实现过程。
[37] 本发明实施方式将待测信号的信号特征信息与预设置的音乐信号特征信息进行 比较， 以判断待测信号的信号特征是否符合音乐信号特征。 这里的预设置的音 乐信号特征信息可以是针对能量变化的， 也可以是针对过零率的， 还可以既针 对能量变化又针对过零率。 例如， 预设置的音乐信号特征信息可以为能量变化 最大值门限、 过零率变化最大值门限和过零率平均值门限中的任意一个或任意 两个或三个。
[38] 当获取的待测信号的信号特征信息包括能量变化最大值、 过零率变化最大值和 过零率平均值中的任意两个或者三个吋， 可以在其中任一信号特征符合预设置 的音乐信号特征吋， 判断出待测信号的信号特征符合音乐信号特征， 即在待测 信号的所有信号特征都不符合音乐信号特征吋， 判断出待测信号的信号特征不 符合音乐信号特征。
[39] 在判断出待测信号的信号特征符合预设置的音乐信号特征后， 可以确定线路中 的信号为音乐信号， 从而可以成功检测出音乐回铃音 （即彩铃） 。 在确定线路 中的信号为音乐信号吋， 可以确定拨打被叫用户的线路处于有效状态， 即确定 被叫电话号码有效、 线路正常且被叫用户还没有接听。 在待测信号的信号特征 符合预设置的音乐信号特征后， 可以输出线路状态有效的信息， 以便于进行后 续的正常接续操作。
[40] 在判断出待测信号的信号特征不符合预设置的音乐信号特征后， 可以确定线路 中的信号不为音乐信号， 如为语音信号等。 因此， 不能够确定拨打被叫用户的 线路是否处于有效状态， 如不能确定是否由于被叫用户不在服务区、 关机等原 因使线路无法接通， 还是其它情况。 在待测信号的信号特征不符合预设置的音 乐信号特征后， 可以输出判断结果， 而不是输出线路状态有效的信息。 输出的 判断结果可以包括原因值。 原因值如信号不为音乐信号等。
[41] 下面以预测式外呼系统釆用待测信号吋域特征和频域特征识别单双频脉冲信号 、 音乐回铃音 （即彩铃） 和语音提示音为例， 结合附图对本发明实施方式提供 的线路状态检测方法进行详细说明。
[42] 在拨打被叫用户后、 且被叫用户摘机前， 通讯线路中的信号主要有单双频脉冲 信号、 音乐回铃音 （即彩铃） 及语音信号。 单双频脉冲信号如传统的忙音、 回 铃音等。 语音信号如语音提示音等。 预测式外呼系统需要准确识别出单双频脉 冲信号、 音乐回铃音及语音信号。 预测式外呼系统利用现有的技术手段可以方 便的识别出单双频脉冲信号。 预测式外呼系统主要需要解决的技术问题为： 如 何准确的识别出音乐回铃音、 及语音信号。 预测式外呼系统是利用信号的吋域 特征和频域特征来识别音乐回铃音及语音信号的。
[43] 语音信号与音乐信号的吋域特征如附图 1所示。
[44] 图 1的上半部分为语音信号的吋域特征， 图 1的下半部分为音乐信号的吋域特征 。 比较上下两部分吋域特征可知： 语音信号的吋域特征和音乐信号的吋域特征 在能量变化上存在差异。 也就是说， 语音信号具有短吋间内平稳的特性， 例如 ， 在 30〜40ms内信号的能量变化比较小。 但是， 在较长吋间段内， 必然会出现 低能量和高能量交互的现象， 即信号能量变化比较明显。 而音乐信号则会在长 吋间内趋于稳定。
[45] 语音信号与音乐信号的频域特征存在的差异包括： 音乐信号的频率成分比语音 信号的频率成分丰富， 音乐信号中存在高频成分， 同吋， 音乐信号的频率成分 随吋间的变化很小， 这主要体现在过零率上。 也就是说， 语音信号的过零率随 吋间的变化比较大， 而音乐信号的过零率随吋间变化比较小。 而且， 音乐信号 的平均过零率比较高， 语音信号的平均过零率比较低。
[46] 从上述描述可知， 通过获取待测信号的能量变化最大值、 过零率变化最大值、 过零率平均值， 并与预设置的能量变化最大值门限、 过零率变化最大值门限和 过零率平均值门限进行比较， 即可确定出待测信号是否为音乐回铃音。
[47] 检测音乐回铃音流程的一个具体例子如附图 2所示。
[48] 图 2中， 步骤 200， 开始音乐回铃音检测过程， 到步骤 205。
[49] 步骤 205， 计算当前帧在吋域上的能量 Eng (n) 。 当前帧可以是通讯线路中待 测信号的釆样信号， 釆样信号如每隔 30ms或 40ms釆样获得的信号， 到步骤 210。
[50] 步骤 210， 计算当前帧在频域上的过零率 Zc (n) ， 到步骤 215。
[51] 步骤 215， 判断当前帧的能量 Eng (n) 是否大于等于能量门限 SilceThreshold ( 如一 30dB) ， 如果 Eng (n) 大于或等于能量门限 SilceThreshold, 则确定当前帧 为语音信号或音乐信号， 到步骤 220。 另外， 如果是第一次检测到 Eng (n) 大于 或等于预设置的能量门限 SilceThreshold, 判决吋间计数器开始计吋， 帧计数器 复位， 过零率总和设置为零， 为能量变化最大值 EngChMax和过零率变化最大值 ZcChMax设置初始值； 如果 Eng (n) 小于能量门限 SilceThreshold, 则确定为静 音信号， 到步骤 225。 这里的静音信号可以表示没有声音。 需要说明的是， 步骤 215中的大于等于也可以为大于， 此吋， 步骤 215中的小于可以为小于等于。
[52] 步骤 220， 帧计数器累加， 且过零率累加。 即帧计数器的计数值 FrameNum加一
， 过零率总和 ZcTotal = ZcTotal + Zc (n) ， 到步骤 225。
[53] 步骤 225， 计算针对当前帧的相邻帧能量变化值： EngCh = Eng (n) Eng (n — 1) ' 到步骤 230。
[54] 步骤 230， 计算针对当前帧的相邻帧过零率变化值： ZcCh (n) = Zc (n) —Zc
(n- 1) ， 到步骤 235。
[55] 步骤 235， 判断针对当前帧的相邻帧能量变化值 EngCh的绝对值是否大于能量 变化最大值 EngChMax, 如果 EngCh>EngChMax， 到步骤 240， 否则， 到步骤 245
。 需要说明的是， 步骤 235中的大于也可以为大于等于。
[56] 步骤 240， 用针对当前帧的相邻帧能量变化值 EngCh的绝对值更新能量变化最 大值 EngChMax, 然后到步骤 245。
[57] 步骤 245， 判断针对当前帧的过零率变化值 ZcCh (n) 的绝对值是否大于过零率 变化最大值 ZcChMax, 如果 ZcCh (n) >ZcChMax, 到步骤 250， 否则， 到步骤 25 5。 需要说明的是， 步骤 245中的大于也可以为大于等于。
[58] 步骤 250， 用针对当前帧的过零率变化值 ZcCh (n) 的绝对值更新过零率变化最 大值 ZcChMax, 然后到步骤 255。
[59] 步骤 255， 判断判决吋间计数器的计数值是否达到预定吋间， 如果达到预定吋 间， 到步骤 260， 否则， 返回步骤 205。
[60] 步骤 260， 计算过零率平均值 ZcAverage = ZcTotal/FrameNum， 到步骤 265。
[61] 步骤 265， 判断能量变化最大值 EngChMax是否大于等于预设置的能量变化最大 值门限 EngThreshold, 如果 EngChMax大于等于 EngThreshold, 则到步骤 280， 否 贝 I」， 到步骤 270。 需要说明的是， 步骤 265中的大于等于也可以为大于。
[62] 步骤 270， 判断过零率变化最大值 ZcChMax是否大于等于预设置的过零率变化 最大值门限 ZcThreshold, 如果 ZcChMax大于等于 ZcThreshold, 则到步骤 280， 否 贝 I」， 到步骤 275。 需要说明的是， 步骤 270中的大于等于也可以为大于。
[63] 步骤 275， 判断过零率平均值 ZcAverage是否大于等于预设置的过零率平均值门 限 ZcAvgThreshold, 如果 ZcAverage大于等于 ZcAvgThreshold, 则到步骤 280， 否 贝 I」， 到步骤 285。 需要说明的是， 步骤 275中的大于等于也可以为大于。
[64] 步骤 280， 确定出线路中的信号为音乐回铃音， 从而可以确定线路状态有效， 到步骤 285。
[65] 步骤 285， 本次音乐回铃音检测过程结束。
[66] 上述流程中的能量变化最大值门限、 过零率变化最大值门限、 过零率平均值门 限、 以及判决吋间计数器对应的预定吋间等可以根据实际需要设置， 如预定吋 间可以设置为 3秒等。
[67] 从上述检测音乐回铃音的流程可以看出， 本发明实施方式的检测算法在运算量 低、 复杂度小的情况下， 能够准确的检测出音乐回铃音。 从而提高了预测式外 呼系统的正常接续准确性。
[68] 易于思及的是， 上述步骤 260的过零率平均值的计算、 步骤 265的能量变化最大 值的比较、 步骤 270的过零率变化最大值的比较、 步骤 275的过零率平均值的比 较的顺序可以根据需要进行调整， 但是， 过零率平均值的计算一定要在过零率 平均值的比较之前执行。 [69] 同样易于思及的是， 在实际应用中， 亦可仅比对上述能量变化最大值、 过零率 变化最大值和过零率平均值中的一个， 来判断待测信号是否为音乐回铃音信号
[70] 另外， 步骤 220、 225、 230执行的先后顺序可以调整， 而且， 步骤 220、 225和 2 30可以同吋执行。 还有， 步骤 235和步骤 245也可以同吋执行， 或者步骤 245和步 骤 250在步骤 235和步骤 240之前执行。 当步骤 245和步骤 250在步骤 235和步骤 240 之前执行的情况下， 实现的过程可以为：
步骤 230， 计算针对当前帧的相邻帧过零率变化值： ZcCh (n
(n- 1) ， 到步骤 245。
步骤 245， 判断针对当前帧的过零率变化值 ZcCh (n) 的绝对值是否大于过零率 变化最大值 ZcChMax, 如果 ZcCh (n) >ZcChMax, 到步骤 250， 否则， 到步骤 23
[73] 步骤 250， 用针对当前帧的过零率变化值 ZcCh (n) 的绝对值更新过零率变化最 大值 ZcChMax, 然后到步骤 235。
[74] 步骤 235， 判断针对当前帧的相邻帧能量变化值 EngCh的绝对值是否大于能量 变化最大值 EngChMax, 如果 EngCh>EngChMax， 到步骤 240， 否则， 到步骤 255
[75] 步骤 240， 用针对当前帧的相邻帧能量变化值 EngCh的绝对值更新能量变化最 大值 EngChMax, 然后到步骤 255。
[76] 通过以上的实施方式的描述， 本领域的技术人员可以清楚地了解到本发明可借 助软件加必需的硬件平台的方式来实现， 当然也可以全部通过硬件来实施， 但 很多情况下前者是更佳的实施方式。 基于这样的理解， 本发明的技术方案对背 景技术做出贡献的全部或者部分可以以软件产品的形式体现出来， 该计算机软 件产品可以存储在存储介质中， 如 ROM/RAM、 磁碟、 光盘等， 包括若干指令用 以使得一台计算机设备 （可以是个人计算机， 服务器， 或者网络设备等） 执行 本发明各个实施例或者实施例的某些部分所述的方法。
[77] 下面结合附图对本发明实施方式提供的线路状态检测装置进行说明。
[78] 图 3为本发明实施方式提供的线路状态检测装置。 该线路状态检测装置可以设 置在主叫用户侧， 也可以设置在网络侧， 如设置在网关处等。 该线路状态检测 装置可以为独立的设备、 也可以设置在现有的设备中。 而且， 线路状态检测装 置中的各模块可以分设在不同的设备中， 也可以设置在同一设备中。
[79] 图 3中， 线路状态检测装置 300包括： 接收模块 310、 获取模块 320和第一检测模 块 330。 线路状态检测装置 300还可以可选的包括第二检测模块 340。
[80] 接收模块 310接收通讯线路中的待测信号。 接收模块 310接收的待测信号如拨打 被叫电话号码后的回应信号。
[81] 获取模块 320获取接收模块 310接收的待测信号的信号特征。 获取模块 320获取 的信号特征可以向第一检测模块 330和第二检测模块 340输出。 获取模块 320可以 获取待测信号的吋域特征信息、 频域特征信息， 以获得待测信号的信号特征信 息。 获取模块 320可以仅获取待测信号的吋域信息， 也可以仅获取频域信息， 还 可以既获取待测信号的吋域信息又获取频域信息。
[82] 获取模块 320可以包括吋域获取子模块 321和频域获取子模块 322中的任意一个 或两个。
[83] 吋域获取子模块 321获取接收模块 310接收的待测信号的吋域特征信息， 如吋域 获取子模块 321获取相邻信号的能量变化信息。 实现方式可以为： 吋域获取子模 块 321获取预定吋间间隔内接收到的相邻待测信号的能量变化最大值。 当然， 吋 域获取子模块 321也可以获取相邻待测信号能量变化的其它信息， 如获取接收到 的相邻待测信号的能量变化均值等。 而且， 本发明实施方式不限制吋域获取子 模块 321获取相邻待测信号的能量变化信息的具体实现过程。
[84] 频域获取子模块 322获取接收模块 310接收的待测信号的频域特征信息。
[85] 频域获取子模块 322可以包括过零率最大值单元 323和平均过零率单元 324中的 任意一个或两个。
[86] 过零率最大值单元 323获取预定吋间间隔内接收模块 310接收的相邻待测信号的 过零率变化最大值。 本发明实施方式不限制过零率最大值单元 323获取相邻待测 信号的过零率变化最大值的具体实现过程。
[87] 平均过零率单元 324获取预定吋间间隔内接收模块 310接收的待测信号的平均过 零率。 本发明实施方式不限制平均过零率单元 324获取待测信号的平均过零率的 具体实现过程。
[88] 上述吋域获取子模块 321、 过零率最大值单元 323和平均过零率单元 324获取的 信息均可以向第一检测模块 330和第二检测模块 340输出。 获取子模块 321、 过零 率最大值单元 323和平均过零率单元 324获取信息的过程可以如上述图 2步骤 205 至步骤 260中的描述。
[89] 第一检测模块 330将接收到的信号特征信息与预设置的音乐信号特征信息进行 比较， 根据比较结果确定出获取的信号特征符合预设置的音乐信号特征后， 输 出线路状态有效的信息。 预设置的音乐信号特征信息可以存储在第一检测模块 3 30中， 也可以存储在第一检测模块 330之外。 预设置的音乐信号特征信息可以为 能量变化最大值、 过零率变化最大值和过零率平均值中的任意一个或任意多个
[90] 当第一检测模块 330接收到的信号特征信息包括能量变化最大值、 过零率变化 最大值、 过零率平均值中的任意两个或者三个吋， 第一检测模块 330可以在其中 任一接收到的信号特征符合预设置的音乐信号特征吋， 判断出接收模块 310接收 的待测信号的信号特征符合音乐信号特征， 即在待测信号的所有信号特征都不 符合音乐信号特征吋， 第一检测模块 330判断出接收模块 310接收的待测信号的 信号特征不符合音乐信号特征。 第一检测模块 330在确定出接收模块 310接收到 的待测信号的信号特征符合音乐信号特征后， 可以输出线路状态有效的信息， 以便于进行后续的正常接续操作。
[91] 第一检测模块 330在判断出接收信号的信号特征不符合预设置的音乐信号特征 后， 可以确定线路中的信号不为音乐信号。 此吋， 第一检测模块 330可以输出非 音乐回铃声等信息。
[92] 第一检测模块 330进行信号特征比较的过程可以如附图 2步骤 265到步骤 285中的 描述， 在此不再重复说明。
[93] 第二检测模块 340根据获取模块 320获取的信号特征信息识别接收模块 310接收 的待测信号是否为单双频脉冲信号， 在识别为单双频脉冲信号后， 可以根据单 双频脉冲信号的具体类型输出线路状态有效或无效的信息， 例如， 在单双频脉 冲信号为忙音吋输出线路状态无效的信息。 第二检测模块 340可以釆用现有的技 术手段来识别单双频脉冲信号， 本发明实施方式不限制第二检测模块 340识别单 双频脉冲信号的具体实现过程。 需要说明的是， 第一检测模块 330和第二检测模 块 340的信号识别过程可以同吋进行， 也可以先后进行。
[94] 下面结合附图 4对本发明实施方式提供的预测式外呼系统进行说明。
[95] 图 4中的预测式外呼系统仅示出了系统控制部分、 呼叫器和彩铃检测器即线路 状态检测装置 300。 呼叫器拨打被叫用户的电话号码。 线路状态检测装置 300检 测电话网络中传输来的待测信号， 并对待测信号进行检测， 然后， 将检测结果 信息输出至系统控制部分， 系统控制部分根据线路状态检测装置 300传输来的检 测结果信息决定后续的操作， 如是否接续等。 线路状态检测装置 300如上述实施 方式中的描述， 在此不再重复说明。
[96] 虽然通过实施例描绘了本发明， 本领域普通技术人员知道， 本发明有许多变形 和变化而不脱离本发明的精神， 本发明的申请文件的权利要求包括这些变形和 变化。

Claims

权利要求书
[1] 一种线路状态检测方法， 其特征在于， 包括：
接收通讯线路中的待测信号， 获取所述待测信号的信号特征信息； 将所述待测信号的信号特征信息与预设置的音乐信号特征信息进行比较； 如果所述比较的结果为所述待测信号的信号特征符合音乐信号特征， 确定 线路状态有效。
[2] 如权利要求 1所述的方法， 其特征在于， 获取所述待测信号的信号特征信息 包括：
获取所述待测信号的吋域特征信息和 /或获取所述待测信号的频域特征信息 且所述预设置的音乐信号特征信息对应的包括： 预设置的音乐信号的吋域 特征信息和 /或预设置的音乐信号的频域特征信息。
[3] 如权利要求 2所述的方法， 其特征在于， 所述获取所述待测信号的吋域特征 信息包括：
获取预定吋间间隔内接收的相邻待测信号的能量变化信息。
[4] 如权利要求 2所述的方法， 其特征在于， 所述获取所述待测信号的频域特征 信息包括：
获取预定吋间间隔内待测信号的过零率信息、 和 /或频率成分变化信息。
[5] 如权利要求 4所述的方法， 其特征在于， 所述获取预定吋间间隔内待测信号 的过零率信息包括：
获取预定吋间间隔内接收的相邻待测信号的过零率变化最大值； 和 /或 获取预定吋间间隔内待测信号的平均过零率。
[6] 如权利要求 1至 5任一项所述的方法， 其特征在于：
在将所述待测信号的信号特征信息与预设置的音乐信号特征信息进行比较 吋， 如果所述待测信号的吋域特征符合预设置的音乐吋域特征、 或待测信 号的频域特征符合预设置的音乐信号的频域特征， 则所述比较的结果为所 述待测信号的信号特征符合音乐信号特征。
[7] 一种线路状态检测装置， 其特征在于， 所述装置包括： 接收模块 (310) ， 用于接收通讯线路中的待测信号；
获取模块 (320) ， 用于获取所述接收模块 (310) 接收的待测信号的信号 特征信息；
第一检测模块 (330) ， 用于将获取模块 (320) 获取的待测信号的信号特 征信息与预设置的音乐信号特征信息进行比较， 如果所述比较的结果为待 测信号的信号特征符合音乐信号特征， 输出线路状态有效的信息。
[8] 如权利要求 7所述的装置， 其特征在于， 所述装置还包括：
第二检测模块 (340) ， 用于根据获取模块 (320) 获取的信号特征信息识 别接收模块 (310) 接收的待测信号是否为单双频脉冲信号。
[9] 如权利要求 7所述的装置， 其特征在于， 所述获取模块 (320) 包括：
吋域获取子模块 (321) ， 用于获取接收模块 (310) 接收的待测信号的吋 域特征信息； 和 /或
频域获取子模块 (322) ， 用于获取接收模块 (310) 接收的待测信号的频 域特征信息。
[10] 如权利要求 9所述的装置， 其特征在于， 所述频域获取子模块 (322) 包括 过零率最大值单元 （323) ， 用于获取预定吋间间隔内接收模块 (310) 接 收的相邻待测信号的过零率变化最大值； 和 /或
平均过零率单元 (324) , 用于获取预定吋间间隔内接收模块 (310) 接收 的待测信号的平均过零率。
[11] 一种预测式外呼系统， 其特征在于， 所述系统包括如权利要求 7至 10中任一 权利要求所述的线路状态检测装置。
[12] 一种回铃音检测方法， 其特征在于， 包括：
接收通讯线路中的待测信号， 获取所述待测信号的信号特征信息； 将所述待测信号的信号特征信息与预设置的音乐信号特征信息进行比较； 如果所述比较的结果为所述待测信号的信号特征符合音乐信号特征， 确定 所述待测信号为音乐回铃音。
[13] 如权利要求 12所述的方法， 其特征在于， 获取所述待测信号的信号特征信 息包括：
获取所述待测信号的吋域特征信息和 /或获取所述待测信号的频域特征信息 且所述预设置的音乐信号特征信息对应的包括： 预设置的音乐信号的吋域 特征信息和 /或预设置的音乐信号的频域特征信息。
PCT/CN2009/070334 2008-03-10 2009-02-02 一种线路状态检测方法、装置和预测式外呼系统 WO2009111957A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200810006576A CN101534352A (zh) 2008-03-10 2008-03-10 一种线路状态检测方法、装置和预测式外呼系统
CN200810006576.3 2008-03-10

Publications (1)

Publication Number Publication Date
WO2009111957A1 true WO2009111957A1 (zh) 2009-09-17

Family

ID=41064747

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2009/070334 WO2009111957A1 (zh) 2008-03-10 2009-02-02 一种线路状态检测方法、装置和预测式外呼系统

Country Status (2)

Country Link
CN (1) CN101534352A (zh)
WO (1) WO2009111957A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108492822A (zh) * 2018-02-23 2018-09-04 济南汇通远德科技有限公司 一种基于工业应用的语音识别方法
CN112530436A (zh) * 2020-11-05 2021-03-19 联通(广东)产业互联网有限公司 通信话务状态的识别方法、系统、装置及存储介质

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105979106B (zh) * 2016-06-13 2019-06-18 北京容联易通信息技术有限公司 一种呼叫中心系统的铃音识别方法和系统
CN109151230A (zh) * 2017-06-26 2019-01-04 浙江金讯网络技术股份有限公司 一种基于数据库输出的外呼系统
CN111294467A (zh) * 2018-12-06 2020-06-16 马上消费金融股份有限公司 一种呼叫处理方法及相关装置
CN112102848B (zh) * 2019-06-17 2024-04-26 华为技术有限公司 一种用于识别音乐的方法、芯片和终端
CN110335627A (zh) * 2019-07-16 2019-10-15 柴桥子 一种基于深度学习的快速判断号码有效性的方法以及装置
CN110602332B (zh) * 2019-08-01 2021-06-15 国家计算机网络与信息安全管理中心 通信线路特征提取方法、通信线路识别方法及装置
CN110545359B (zh) * 2019-08-02 2021-06-29 国家计算机网络与信息安全管理中心 通信线路特征提取方法、通信线路识别方法及装置
CN111131628A (zh) * 2019-12-20 2020-05-08 北京淇瑀信息科技有限公司 一种线路未接通状态的语音识别方法、装置和系统

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1426563A (zh) * 2000-12-22 2003-06-25 皇家菲利浦电子有限公司 使用音频类别界定节目边界和商业广告边界的系统和方法
US20070098146A1 (en) * 1997-12-19 2007-05-03 Blake Rice Automated right-party contact telephone system

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070098146A1 (en) * 1997-12-19 2007-05-03 Blake Rice Automated right-party contact telephone system
CN1426563A (zh) * 2000-12-22 2003-06-25 皇家菲利浦电子有限公司 使用音频类别界定节目边界和商业广告边界的系统和方法

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
BAI L. ET AL.: "Research for Intelligent Call Center Building Method", MINI-MICRO SYSTEMS, vol. 26, no. 11, November 2005 (2005-11-01) *
DING X. ET AL.: "Research for Intelligent Call Center Building Method", NANJING UNIVERSITY OF SCIENCE AND TECHNOLOGY, no. 1, 2007 *
LU J. ET AL., FEATURE ANALYSIS FOR SPEECH/MUSIC AUTOMATIC CLASSIFICATION JOURNAL OF COMPUTER-AIDED DESIGN & COMPUTER GRAPHICS, vol. 14, no. 3, March 2002 (2002-03-01), pages 233 - 237 *
QI L. ET AL., RESEARCH ON OUTBOUND DIALING TECHNOLOGY OF CALL CENTER BASED ON PREDICTIVE APPROACH COMPUTER ENGINEERING AND DESIGN, vol. 27, no. 15, August 2006 (2006-08-01), pages 2799 - 2801 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108492822A (zh) * 2018-02-23 2018-09-04 济南汇通远德科技有限公司 一种基于工业应用的语音识别方法
CN112530436A (zh) * 2020-11-05 2021-03-19 联通(广东)产业互联网有限公司 通信话务状态的识别方法、系统、装置及存储介质

Also Published As

Publication number Publication date
CN101534352A (zh) 2009-09-16

Similar Documents

Publication Publication Date Title
WO2009111957A1 (zh) 一种线路状态检测方法、装置和预测式外呼系统
CA2419215C (en) Method and apparatus for detection of a telephone cpe alerting signal
JP2597817B2 (ja) 音声信号検出方法
GB2352377A (en) Noise suppression method for a mobile telephone
US6870917B2 (en) Methods and systems for call interruption services
WO2012000346A1 (zh) 电话通路故障的检测方法及装置
GB2487734A (en) An Answer Machine Detection (AMD) system using Network Level Binary Matching
CN101159891B (zh) 语音激活检测控制的方法及其控制设备
JP5597215B2 (ja) ファクシミリ装置、その制御方法、及びプログラム
WO2006063505A1 (fr) Dispositif pour detecter une tonalite de rappel et methode associee
CN111405131B (zh) 一种远端摘机信号的检测方法、系统及存储介质
WO2009135408A1 (zh) 信号音检测的方法、装置及系统
CN105516518A (zh) 一种网络语音测试方法
US20100303214A1 (en) One-way voice detection voicemail
US7844047B1 (en) Tune master internet protocol device
JP3642955B2 (ja) 自動通報装置及び自動通報方法
KR100368596B1 (ko) 언 피.비.엑스 시스템의 아날로그 호응답 신호 판단방법및 그 장치
KR100999654B1 (ko) 묵음 발생 검지 장치 및 방법
CN1085457C (zh) 一种改善用户前端设备警告信号音频误检测的方法及装置
US7095824B2 (en) Instructive out-of-service telephone lines
EP2387197A1 (en) Method, system, and media gateway for reporting dual tone multi-frequency digits
JP2004193748A (ja) Ｉｐ電話ゲートウェイ装置およびその処理方法
CN1819603B (zh) 模拟中继接口盗打电话检测方法
CN201839340U (zh) 多媒体终端
KR19980041580A (ko) 사설교환기에서 착신측 상태 음성안내방법

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 09719886

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 09719886

Country of ref document: EP

Kind code of ref document: A1