CN104819382A

CN104819382A - Self-adaptive constant false alarm rate vibration source detection method for optical fiber early warning system

Info

Publication number: CN104819382A
Application number: CN201510166995.3A
Authority: CN
Inventors: 曲洪权; 郑彤; 毕福昆; 杜栩; 薛晓鹏
Original assignee: Beijing Easted Information Technology Co ltd; North China University of Technology
Current assignee: Beijing Easted Information Technology Co ltd; North China University of Technology
Priority date: 2015-04-09
Filing date: 2015-04-09
Publication date: 2015-08-05
Anticipated expiration: 2035-04-09
Also published as: CN104819382B

Abstract

The invention provides a signal detection method based on self-adaptive constant false alarm rate (GO/SO-CFAR), which is used for realizing vibration source detection in a pipeline optical fiber early warning system and comprises the following steps: continuously receiving N pulse echo data to obtain a frame of echo data, sequentially selecting a distance unit from M distance units in the nth pulse echo data as a detection unit, removing an interference target from a reference unit sequence, calculating a detection threshold value based on the processed reference unit sampling sequence, and finally comparing the detection unit with the detection threshold value to obtain a detection result. The invention utilizes the self-adaptive constant false alarm rate detector to detect and process the optical fiber vibration signal, and the method can ensure constant false alarm rate and also can show stronger detection capability under different backgrounds.

Description

An Adaptive Constant False Alarm Rate Vibration Source Detection Method for Optical Fiber Early Warning System

技术领域technical field

本发明涉及一种用于光纤预警系统的自适应恒虚警率振源检测方法，属于光纤振动信号检测技术领域。The invention relates to an adaptive constant false alarm rate vibration source detection method for an optical fiber early warning system, belonging to the technical field of optical fiber vibration signal detection.

背景技术Background technique

随着经济的发展，石油和天然气已经成为国民经济发展的最重要能源。长输管道作为石油和天然气最理想、经济的运输方式，已经得到广泛的应用。油气管道是能源运输的大动脉，管道一旦泄漏极易发生燃烧爆炸，不仅影响管道的安全生产，还将给国家和人民群众的生命与财产造成巨大损失。With the development of the economy, oil and natural gas have become the most important energy sources for the development of the national economy. As the most ideal and economical transportation method for oil and natural gas, long-distance pipelines have been widely used. The oil and gas pipeline is the main artery of energy transportation. Once the pipeline leaks, it is easy to burn and explode, which not only affects the safe production of the pipeline, but also causes huge losses to the life and property of the country and the people.

目前应用油气管道的探测报警系统主要有以下几种：电子脉冲式围栏、微波墙式报警器、主动红外报警器、泄露电缆式周界探测报警系统、驻极体振动电缆报警系统和光纤传感器周界报警系统。与电传感器报警系统相比，光纤传感器在传感网络应用中具有非常明显的技术优势：在不需要任何户外有源器件(不需供电)的情况下能够提供长达100公里距离的安防监控，不受地形的高低、曲折、转弯、折弯等地形环境限制，打破了红外线、微波墙等只适用于视距和平坦区域使用的局限性。因此利用光纤测量振动成为管道预警系统研究的主要方法。然而如何对光纤检测信号进行合理有效的分析，建立什么样的事件模型才更为有效，成为研究中的一大热点和难点。At present, the detection and alarm systems for oil and gas pipelines mainly include the following types: electronic pulse fence, microwave wall alarm, active infrared alarm, leakage cable perimeter detection alarm system, electret vibration cable alarm system and fiber optic sensor perimeter Boundary alarm system. Compared with electrical sensor alarm systems, fiber optic sensors have very obvious technical advantages in sensor network applications: they can provide security monitoring up to 100 kilometers away without any outdoor active devices (no power supply required), It is not restricted by the terrain environment such as height, twists and turns, turns, and bends of the terrain, breaking the limitations that infrared rays and microwave walls are only applicable to line-of-sight and flat areas. Therefore, the use of optical fiber to measure vibration has become the main method of pipeline early warning system research. However, how to analyze the optical fiber detection signal reasonably and effectively, and what kind of event model to establish is more effective, has become a hot spot and difficulty in research.

目前光纤振动信号处理还存在明显不足，其设计思路一般是直接检测的方法，将检测到的信号直接送到显示器，将杂波和噪声的幅度变化同时显示出来，对目标信号的检测能力由操作员对显示器的监视决定的。但是，这种处理方法所得到的监测结果误差大、报警不准确，使长距离复杂振动检测面临严峻挑战，急需在此基础上开展长距离光纤振动检测方法研究。At present, there are still obvious deficiencies in optical fiber vibration signal processing. The design idea is generally a direct detection method, which sends the detected signal directly to the display, and displays the amplitude changes of clutter and noise at the same time. The detection ability of the target signal is controlled by the operator. determined by the monitor's monitoring. However, the monitoring results obtained by this processing method have large errors and inaccurate alarms, which make long-distance complex vibration detection face severe challenges. It is urgent to carry out research on long-distance optical fiber vibration detection methods on this basis.

现有的研究存在的主要问题是没有建立合适的模型、特别是没有建立合适的信号检测模型，从而使得已经投入生产的长距离预警系统效率较低甚至被搁置不用，光纤预警系统中的信号处理环节已成为系统和产业发展的最主要瓶颈。因此，需要提出一种合适的模型来实现振动事件的检测，以提高检测概率和虚警概率的稳定性。The main problem in the existing research is that no suitable model has been established, especially no suitable signal detection model has been established, so that the long-distance early warning system that has been put into production is inefficient or even put on hold. The signal processing in the optical fiber early warning system Links have become the main bottleneck of system and industrial development. Therefore, it is necessary to propose a suitable model to realize the detection of vibration events in order to improve the stability of detection probability and false alarm probability.

发明内容Contents of the invention

本发明基于自适应恒虚警率(GO/SO-CFAR)的光纤预警系统振源检测算法以解决对光纤振动信号检测的问题。The invention is based on an adaptive constant false alarm rate (GO/SO-CFAR) vibration source detection algorithm of an optical fiber early warning system to solve the problem of optical fiber vibration signal detection.

基于自适应恒虚警率(GO/SO-CFAR)的光纤预警系统振源检测方法，其包括：对光纤振动信号进行采集；建立自适应恒虚警率(GO/SO-CFAR)检测器模型；对无振动信号进行GO/SO-CFAR检测，比较实际虚警率与设定虚警率的关系，保证检测器正常工作；对均匀背景信号进行检测，得到在均匀背景下的检测性能；对多目标背景信号进行检测，得到在多目标背景下的检测性能；分别在两种环境下调整信噪比，再进行GO/SO-CFAR检测，画出检测性能曲线。A vibration source detection method for an optical fiber early warning system based on an adaptive constant false alarm rate (GO/SO-CFAR), which includes: collecting optical fiber vibration signals; establishing an adaptive constant false alarm rate (GO/SO-CFAR) detector model ; Conduct GO/SO-CFAR detection on the non-vibration signal, compare the relationship between the actual false alarm rate and the set false alarm rate, and ensure the normal operation of the detector; detect the uniform background signal, and obtain the detection performance under the uniform background; The multi-target background signal is detected to obtain the detection performance under the multi-target background; the signal-to-noise ratio is adjusted in the two environments respectively, and then the GO/SO-CFAR detection is performed, and the detection performance curve is drawn.

其中，检测阈值满足以下关系：Among them, the detection threshold satisfies the following relationship:

S＝TZS=TZ

其中，S是阈值；Z是功率水平，其值分为两种情况考虑，若判定结果k值大于参考单元的一半，即k>R/2时，否则T是标称化因子，其值是根据虚警概率P_fa求得，Among them, S is the threshold; Z is the power level, and its value is divided into two cases. If the judgment result k value is greater than half of the reference unit, that is, when k>R/2, otherwise T is a normalization factor, and its value is obtained according to the false alarm probability P _fa ,

${P P}_{fa fa} = = \{\begin{matrix} (\begin{matrix} R R \\ k k \end{matrix}) {Π Π}_{j j = = 11}^{k k} {[[T T + + \frac{R R - - j j + + 11}{k k - - j j + + 11}]]}^{- - 11},, k k > > R R / / 22 \\ (\begin{matrix} R R \\ k k \end{matrix}) {((11 + + T T))}^{- - ((N N - - k k - - 11))} \cdot &Center Dot; {Σ Σ}_{j j = = 00}^{k k} (\begin{matrix} k k \\ j j \end{matrix}) {((- - 11))}^{j j} {((11 + + \frac{j j}{N N - - k k} + + T T))}^{- - 11},, k k \leq \leq R R / / 22 \end{matrix}$

通过上述方案建立的自适应恒虚警率(GO/SO-CFAR)检测器，能够保证检测的虚警率恒定，并在均匀背景和多目标背景下都有较强的检测能力。The self-adaptive constant false alarm rate (GO/SO-CFAR) detector established by the above scheme can ensure a constant false alarm rate, and has strong detection ability in both uniform background and multi-target background.

根据本发明的一个方面，提供了一种基于自适应恒虚警率的信号检测方法，其特征在于包括：According to one aspect of the present invention, a kind of signal detection method based on adaptive constant false alarm rate is provided, it is characterized in that comprising:

建立自适应虚警率检测器；Build an adaptive false alarm rate detector;

将无振动实验数据经过自适应恒虚警率检测器检测，保证实际虚警率与设定的虚警率极其接近，即绝对误差保证在1e-4数量级内；The vibration-free experimental data is detected by an adaptive constant false alarm rate detector to ensure that the actual false alarm rate is extremely close to the set false alarm rate, that is, the absolute error is guaranteed to be within the order of 1e-4;

运用自适应恒虚警率分别对均匀背景下以及多目标背景下的信号进行检测，分别求得其检测概率，并更改数据的信噪比，对多组数据进行检测，得到在不同信噪比下的检测概率，表现其检测性能。The self-adaptive constant false alarm rate is used to detect the signals in the uniform background and the multi-target background respectively, and the detection probability is obtained respectively, and the signal-to-noise ratio of the data is changed, and multiple sets of data are detected, and the signal-to-noise ratio in different signal-to-noise ratios is obtained. The detection probability under , expresses its detection performance.

附图说明Description of drawings

图1本发明方法的实施以及验证过程；Implementation and verification process of the inventive method of Fig. 1;

图2自适应恒虚警率(GO/SO-CFAR)检测器方框图；Figure 2 Adaptive constant false alarm rate (GO/SO-CFAR) detector block diagram;

图3自适应恒虚警率(GO/SO-CFAR)删除算法流程图；Fig. 3 adaptive constant false alarm rate (GO/SO-CFAR) deletion algorithm flowchart;

图4均匀背景下自适应恒虚警率(GO/SO-CFAR)检测结果；Fig. 4 Detection results of adaptive constant false alarm rate (GO/SO-CFAR) in uniform background;

图5多目标背景下自适应恒虚警率(GO/SO-CFAR)检测结果；Fig. 5 Detection results of adaptive constant false alarm rate (GO/SO-CFAR) in multi-target background;

图6自适应恒虚警率(GO/SO-CFAR)在均匀背景下的检测性能曲线；Figure 6 The detection performance curve of adaptive constant false alarm rate (GO/SO-CFAR) in a uniform background;

图7自适应恒虚警率(GO/SO-CFAR)在多目标背景下的检测性能曲线；Fig. 7 The detection performance curve of adaptive constant false alarm rate (GO/SO-CFAR) in multi-target background;

具体实施方案specific implementation plan

下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述。显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部实施例。基于本发明中的实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他实施例，都属于本发明保护的范围。The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Apparently, the described embodiments are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, those of ordinary skill in the art can obtain other embodiments according to these drawings without any creative effort, which all belong to the protection scope of the present invention.

请参见图1所示，图1是本发明的基于自适应恒虚警率(GO/SO-CFAR)的光纤预警系统振源检测算法的流程图。如图1所示，本实施例所揭示的基于自适应恒虚警率(GO/SO-CFAR)的光纤预警系统振源检测算法包括：Please refer to FIG. 1 , which is a flow chart of the vibration source detection algorithm of the optical fiber early warning system based on the adaptive constant false alarm rate (GO/SO-CFAR) of the present invention. As shown in Figure 1, the optical fiber early warning system vibration source detection algorithm based on adaptive constant false alarm rate (GO/SO-CFAR) disclosed in this embodiment includes:

S101：对光纤振动信号进行采集；S101: collecting optical fiber vibration signals;

S102：建立自适应恒虚警率(GO/SO-CFAR)检测器模型；S102: Establish an adaptive constant false alarm rate (GO/SO-CFAR) detector model;

S103：对无振动信号进行自适应恒虚警率(GO/SO-CFAR)检测，比较实际虚警率与设定虚警率的关系，保证检测器正常工作；S103: Perform adaptive constant false alarm rate (GO/SO-CFAR) detection on the vibration-free signal, compare the relationship between the actual false alarm rate and the set false alarm rate, and ensure the normal operation of the detector;

S104：对均匀背景信号进行检测，得到在均匀背景下的检测性能；S104: Detecting a uniform background signal to obtain detection performance under a uniform background;

S105：对多目标背景信号进行检测，得到在多目标北京下的检测性能；S105: Detect multi-target background signals to obtain detection performance under multi-target conditions;

S106:分别在两种环境下调整信噪比，再进行自适应恒虚警率(GO/SO-CFAR)检测，画出检测性能曲线。S106: Adjust the signal-to-noise ratio in the two environments respectively, and then perform adaptive constant false alarm rate (GO/SO-CFAR) detection, and draw a detection performance curve.

在S101中对光纤振动系统的信号进行采集,连续接收N个脉冲回波数据而得到一帧回波数据，并对每个脉冲回波数据进行处理和采样。其中，每个脉冲回波数据中包含M个按照回波时间从左至右顺序排列的距离单元,不同的列数表示回波位置。具体地，将接收到的每一脉冲回波数据进行处理和采样，形成以矩阵形式排列的信号。In S101, the signal of the optical fiber vibration system is collected, N pulse echo data are continuously received to obtain a frame of echo data, and each pulse echo data is processed and sampled. Wherein, each pulse echo data includes M distance units arranged in order from left to right according to echo time, and different column numbers represent echo positions. Specifically, each received pulse echo data is processed and sampled to form a signal arranged in a matrix.

在S102中，建立自适应恒虚警率(GO/SO-CFAR)检测器模型，自适应恒虚警率(GO/SO-CFAR)检测器模型如图2所示，对已经形成矩阵的数据进行部分删除，再求出阈值，最后对检测单元与阈值进行比较，得出检测结果。其中的重点部分是删除过程，如图3所示，其揭示的方法包括：In S102, an adaptive constant false alarm rate (GO/SO-CFAR) detector model is established, and the adaptive constant false alarm rate (GO/SO-CFAR) detector model is shown in Figure 2, for the data that has formed a matrix Partial deletion is performed, and then the threshold value is calculated, and finally the detection unit is compared with the threshold value to obtain the detection result. The key part is the deletion process, as shown in Figure 3, the revealed methods include:

S301：对参考单元采样排序x₍₁₎≤x₍₂₎≤…≤x_(R)；S301: Sort reference unit sampling x ₍₁₎ ≤ x ₍₂₎ ≤... ≤ x _(R) ;

S302：然后对k个较低的有序采样求和S302: Then sum the k lower ordered samples

${Z Z}_{k k} = = {Σ Σ}_{i i = = 11}^{k k} {x x}_{((i i))}$

其中，x是参考单元的值；Z_k是前k项参考单元杂波功率；Among them, x is the value of the reference unit; Z _k is the clutter power of the first k reference units;

S303：根据给定的错误删除概率取得阈值标称化因子T_k,删除过程第k步的错误删除概率P_FC被定义为S303: Obtain the threshold normalization factor T _k according to the given error deletion probability, and the error deletion probability P _FC of the kth step of the deletion process is defined as

P_FC＝Pr{D_k>0|H_N} (1)P _FC ＝Pr{D _k >0|H _N } (1)

其中，P_FC为错误删除概率,H_N为弱杂波采样,D_k为检验统计量，Among them, P _FC is the probability of false deletion, H _N is weak clutter sampling, D _k is the test statistic,

D_k＝x_(k+1)-T_kZ_k (2)D _k = x _{(k+1) -} T _k Z _k (2)

其中，x_(k+1)为第(k+1)项采样值,T_k为第k项阈值标称化因子；Wherein, x _(k+1) is the sampling value of the (k+1)th item, and T _k is the threshold normalization factor of the kth item;

错误删除概率P_FC可以写成积分形式，The false deletion probability P _FC can be written in integral form,

${P P}_{FC FC} = = - - \frac{11}{22 πi πi} &Integral; &Integral; {w w}^{- - 11} {Φ Φ}_{{D D.}_{k k} | | {H h}_{N N}} ((w w)) dw dw - - - - - - ((33))$

其中，为在H_N假设下D_k的矩母函数，定义式为in, is the moment generating function of D _k under the assumption of H _N , the definition formula is

${Φ Φ}_{{D D.}_{k k} | | {H h}_{c c}} ((w w)) = = E E. [[exp exp ((- - {D D.}_{k k} w w))]] - - - - - - ((44))$

将(2)式代入(4)式中，得Substituting (2) into (4), we get

${Φ Φ}_{{D D.}_{k k} | | {H h}_{c c}} ((w w)) = = E E. {{exp exp [[- - (({x x}_{((k k + + 11))} - - {T T}_{k k} {Z Z}_{k k})) w w]]}} - - - - - - ((55))$

有序统计D₍₁₎,D₍₂₎，…D_(k+1)的联合矩母函数定义为The joint moment generating function of ordered statistics D ₍₁₎ , D ₍₂₎ , ... D _(k+1) is defined as

${Φ Φ}_{{D D.}_{((11))},, . . . . . .,, {D D.}_{((k k + + 11))} | | {H h}_{t t}} (({w w}_{11},, {w w}_{22},, . . . . . .,, {w w}_{k k + + 11})) = = E E. [[exp exp {{- - {Σ Σ}_{j j = = 11}^{k k + + 11} (({w w}_{j j} {D D.}_{((j j))}))}}]] - - - - - - ((66))$

比较(5)、(6)两式，可以得出Comparing the two formulas (5) and (6), we can get

w₁＝w₂＝…＝w_k＝-T_kww ₁ ＝w ₂ ＝...＝w _k ＝-T _k w

以及as well as

w_k+1＝ww _k+1 = w

因此therefore

${Φ Φ}_{{D D.}_{k k} | | {H h}_{N N}} ((w w)) = = \frac{R R!!}{((R R - - k k + + 11))!!} {Π Π}_{j j = = 11}^{k k + + 11} \frac{11}{((R R - - j j + + 11)) + + w w [[11 - - ((k k - - j j + + 11)) {T T}_{k k}]]} - - - - - - ((77))$

其中，R为参考单元数目，把式(7)代入式(3),得Among them, R is the number of reference units, and substituting formula (7) into formula (3), we get

${P P}_{FC FC} = = (\begin{matrix} R R \\ k k \end{matrix}) \frac{11}{{[[11 + + {T T}_{k k} ((R R - - k k))]]}^{k k}} - - - - - - ((88))$

即可根据公式(8)，在给定错误删除概率的情况下求出阈值标称化因子；That is, according to the formula (8), the threshold normalization factor can be obtained under the given error deletion probability;

S304：求得自适应阈值T_kZ_k，将第k+1项参考单元值与自适应阈值进行比较；S304: Obtain the adaptive threshold T _k Z _k , and compare the value of the k+1th reference unit with the adaptive threshold;

S305：如果x_(k+1)小于T_kZ_k，则对k自增1，重复进行上述步骤，否则直接停止循环，即说明x_(k+1),…,x_(R)属于强杂波区采样；S305: If x _(k+1) is less than T _k Z _k , then increment k by 1 and repeat the above steps, otherwise stop the cycle directly, which means that x _(k+1) ,…,x _(R) are strongly heterogeneous Wave zone sampling;

S306：然后对k值与R/2值进行比较；S306: Then compare the k value with the R/2 value;

S307：若k>R/2，则判定检测单元处于弱杂波区，由x₍₁₎,…,x_(k)形成杂波功率Z，即其检验统计量D的矩母函数定义式为S307: If k>R/2, it is determined that the detection unit is in the weak clutter area, and the clutter power Z is formed by x ₍₁₎ ,...,x _(k) , namely The moment generating function of the test statistic D is defined as

${Φ Φ}_{{D D.}_{k k} | | {H h}_{N N}} ((w w)) = = (\begin{matrix} R R \\ k k \end{matrix}) {Π Π}_{j j = = 11}^{k k} {[[w w + + \frac{R R - - j j + + 11}{k k - - j j + + 11}]]}^{- - 11} - - - - - - ((99))$

虚警概率的定义式为The definition of false alarm probability is

P_fa＝Φ(T) (10)P _fa ＝Φ(T) (10)

其中，P_fa是虚警概率，则联立式(9)、式(10)，得Among them, P _fa is the probability of false alarm, then the simultaneous formula (9) and formula (10), we get

${P P}_{fa fa} = = (\begin{matrix} R R \\ k k \end{matrix}) {Π Π}_{j j = = 11}^{k k} {[[T T + + \frac{R R - - j j + + 11}{k k - - j j + + 11}]]}^{- - 11} - - - - - - ((1111))$

S308：若k<R/2,则判定检测单元处于强杂波区，由x_(k+1),…,x_(R)形成杂波功率Z即其检验统计量D的矩母函数定义式为S308: If k<R/2, it is determined that the detection unit is in the strong clutter area, and the clutter power Z is formed by x _(k+1) ,...,x _(R) The moment generating function of the test statistic D is defined as

${Φ Φ}_{{D D.}_{k k} | | {H h}_{N N}} ((w w)) = = (\begin{matrix} R R \\ k k \end{matrix}) {((11 + + w w))}^{- - ((N N - - k k - - 11))} \cdot \cdot {Σ Σ}_{j j = = 00}^{k k} (\begin{matrix} k k \\ j j \end{matrix}) {((- - 11))}^{j j} {((11 + + \frac{j j}{N N - - k k} + + w w))}^{- - 11} - - - - - - ((1212))$

同样将式(12)与虚警定义式(10)联立Similarly, formula (12) and false alarm definition formula (10) are combined

${P P}_{fa fa} = = (\begin{matrix} R R \\ k k \end{matrix}) {((11 + + T T))}^{- - ((N N - - k k - - 11))} \cdot \cdot {Σ Σ}_{j j = = 00}^{k k} (\begin{matrix} k k \\ j j \end{matrix}) {((- - 11))}^{j j} {((11 + + \frac{j j}{N N - - k k} + + T T))}^{- - 11} - - - - - - ((1313))$

综上所述，标称化因子T是根据虚警概率P_fa求得，如下式To sum up, the normalization factor T is obtained according to the false alarm probability P _fa , as follows

${P P}_{fa fa} = = \{\begin{matrix} (\begin{matrix} R R \\ k k \end{matrix}) {Π Π}_{j j = = 11}^{k k} {[[T T + + \frac{R R - - j j + + 11}{k k - - j j + + 11}]]}^{- - 11},, k k > > R R / / 22 \\ (\begin{matrix} R R \\ k k \end{matrix}) {((11 + + T T))}^{- - ((N N - - k k - - 11))} \cdot \cdot {Σ Σ}_{j j = = 00}^{k k} (\begin{matrix} k k \\ j j \end{matrix}) {((- - 11))}^{j j} {((11 + + \frac{j j}{N N - - k k} + + T T))}^{- - 11},, k k \leq \leq R R / / 22 \end{matrix}$

最后计算出阈值Finally calculate the threshold

S＝TZS=TZ

其中，S是阈值。where S is the threshold.

在S103中，对光纤振动系统采集到的无振动数据进行GO/SO-CFAR检测，参数设置以及检测后的虚警概率如表1所示。In S103, GO/SO-CFAR detection is performed on the non-vibration data collected by the optical fiber vibration system. The parameter settings and false alarm probability after detection are shown in Table 1.

表1Table 1

可以看出实际虚警概率与设定虚警概率之间的误差在1e-4级内，说明自适应恒虚警率(GO/SO-CFAR)检测器正常工作。It can be seen that the error between the actual false alarm probability and the set false alarm probability is within 1e-4 level, indicating that the adaptive constant false alarm rate (GO/SO-CFAR) detector works normally.

在S104中，对同一组均匀背景下的数据进行自适应恒虚警率(GO/SO-CFAR)的重复10000次检测，参数设置如表2所示，In S104, repeated 10,000 times of self-adaptive constant false alarm rate (GO/SO-CFAR) detection is performed on the same group of data under a uniform background, and the parameter settings are shown in Table 2.

表2Table 2

分别取16幅检测结果图，如图4所示，再通过检测结果数据的记录，如表3所示，说明在均匀背景下自适应恒虚警率(GO/SO-CFAR)检测器的检测性能。Take 16 detection result pictures respectively, as shown in Figure 4, and then record the test result data, as shown in Table 3, to illustrate the detection of adaptive constant false alarm rate (GO/SO-CFAR) detector in a uniform background performance.

表3table 3

在S105中，对一组多目标背景下的数据分别进行自适应恒虚警率(GO/SO-CFAR)的重复10000次检测，参数设置如表4所示In S105, repeated 10,000 times of self-adaptive constant false alarm rate (GO/SO-CFAR) detection is performed on a group of data in a multi-target background, and the parameter settings are shown in Table 4

表4Table 4

分别取16幅检测结果图，如图5所示，再通过检测结果数据的记录，如表5所示，更能说明在多目标背景下自适应恒虚警率(GO/SO-CFAR)检测器的检测性能。Take 16 detection result pictures respectively, as shown in Figure 5, and then record the test result data, as shown in Table 5, which can better explain the adaptive constant false alarm rate (GO/SO-CFAR) detection in the multi-target background. device detection performance.

表5table 5

检测方式Detection method 平均检测概率mean probability of detection 自适应恒虚警率(GO/SO-CFAR)Adaptive constant false alarm rate (GO/SO-CFAR) 0.99500.9950

在S106中，对不同信噪比的数据分别进行检测，即可得到在信噪比不同的情况下，自适应恒虚警率(GO/SO-CFAR)检测器的检测概率值，即检测性能曲线，在均匀背景下和多目标背景下检测性能分别如图6、7所示。可以看出，自适应恒虚警率(GO/SO-CFAR)的检测能力较强。In S106, the data with different signal-to-noise ratios are detected separately, and the detection probability value of the adaptive constant false alarm rate (GO/SO-CFAR) detector under different signal-to-noise ratios can be obtained, that is, the detection performance The detection performance in the uniform background and multi-target background is shown in Fig. 6 and Fig. 7, respectively. It can be seen that the detection ability of adaptive constant false alarm rate (GO/SO-CFAR) is strong.

本发明与现有检测方法相比具有以下优点Compared with existing detection methods, the present invention has the following advantages

(1)、该方法能够达到对光纤振动信号检测；(1), the method can achieve optical fiber vibration signal detection;

(2)、恒虚警率检测方法能保证虚警概率恒定，性能稳定；(2) The constant false alarm rate detection method can ensure constant false alarm probability and stable performance;

(3)、自适应恒虚警率(GO/SO-CFAR)在均匀背景下以及多目标背景下都表现出较高的检测概率，即该自适应恒虚警率(GO/SO-CFAR)方法能保证在不同背景下的检测能力。(3), the adaptive constant false alarm rate (GO/SO-CFAR) shows a high detection probability in the uniform background and the multi-target background, that is, the adaptive constant false alarm rate (GO/SO-CFAR) The method can guarantee the detection ability in different backgrounds.

Claims

1. The signal detection method based on the self-adaptive constant false alarm rate is characterized by comprising the following steps:

establishing a self-adaptive false alarm rate detector;

detecting the vibration-free experimental data by a self-adaptive constant false alarm rate detector to ensure that the actual false alarm rate is extremely close to the set false alarm rate, namely the absolute error is ensured to be within 1e-4 orders of magnitude;

the method comprises the steps of respectively detecting signals under a uniform background and a multi-target background by using a self-adaptive constant false alarm rate, respectively obtaining detection probabilities, changing signal-to-noise ratios of data, detecting multiple groups of data, obtaining detection probabilities under different signal-to-noise ratios, and expressing detection performance.

2. The signal detection method according to claim 1, characterized by further comprising:

establishing adaptive constant false alarm rate detector model, including partial deletion of data formed into matrix, calculating threshold value, comparing detection unit with threshold value to obtain detection result,

wherein the partial deletion comprises:

(1) sample ordering x for ith row of reference cell₍₁₎≤x₍₂₎≤…≤x_(R)Wherein x is a reference unit value, and R is the total number of reference units;

(2) by x₍₁₎Representing clutter samples, comparing x₍₂₎And T₁x₍₁₎Wherein T is₁Is to satisfy a given false deletion probability P_FCIs a normalized factor of the threshold value of (x)₍₂₎<T₁x₍₁₎Determining x₍₁₎And x₍₂₎Sampling in the same distribution, and then carrying out the next step; if x₍₂₎>T₁x₍₁₎Determining x₍₂₎,…,x_(R)The method is characterized in that the method is an echo signal of a strong clutter region, and by analogy, all units in the ith row are divided into a strong clutter region and a weak clutter region;

(3) assume the division position is k, if k>R/2, wherein R is the number of reference units, the detection unit is judged to be in the weak clutter region, and x is₍₁₎,…,x_(k)Forming clutter power Z; if k is<R/2, then the detection unit is judged to be in a strong clutter region, which is determined by x_(k+1),…,x_(R)Forming clutter power Z;

(4) and solving a threshold value according to the S-TZ, wherein the S is the threshold value, the T is a normalization factor, and when the detection unit is larger than the threshold value S, the detection unit is judged to be noise, namely the signal is alarmed in the optical fiber early warning.

3. The signal detection method according to claim 2, characterized by further comprising:

obtaining a threshold normalization factor T based on a given probability of false erasure_kWherein the threshold value is normalized by a factor T_kCalculated according to the following formula

P_{FC} = (\begin{matrix} R \\ k \end{matrix}) \frac{1}{{[1 + T_{k} (R - k)]}^{k}} - - - (1)

Wherein R is the number of reference cells, P_FCProbability of false deletion in step k of deletion process

P_FC＝Pr{D_k>0|H_N} (2)

Wherein, P_FCTo the false deletion probability, H_NFor weak clutter sampling, D_kIn order to test the statistics of the test,

D_k＝x_(k+1)-T_kZ_k (3)

wherein x is_(k+1)Is the sampled value of (k +1) th term, T_kFor the k term threshold normalization factor, Z_kThe ordered sum of the first k terms.

4. A signal detection method as claimed in claim 2, characterized in that the normalization factor T is based on the probability of false alarm P_faObtained as follows