阅读说明：本技术 一种主动声纳宽带空间谱回波亮点增强及自动提取方法 (Active sonar broadband spatial spectrum echo bright spot enhancement and automatic extraction method ) 是由 邵鹏飞 邹丽娜 马启明 于 2020-04-08 设计创作，主要内容包括：本发明公开了一种主动声纳宽带空间谱回波亮点增强及自动提取方法,该回波提取方法利用了回波、混响、被动辐射噪声的分布特征差异性,通过主动声纳宽带空间谱锐化和空间谱差分两级处理有效抑制了混响和被动辐射噪声干扰,增强了回波亮点强度,可实现低虚警率的自动回波亮点提取,为发展主动声纳自主检测和跟踪、以及智能化主动探测提供了可靠的先决条件,该方法原理清晰、实现步骤简单、计算量小,通过海试数据分析验证了该方法的有效性,为当前的装备改进和新产品研制提供一定技术支撑。(The invention discloses an echo bright spot enhancement and automatic extraction method of an active sonar broadband spatial spectrum, which utilizes the distribution characteristic difference of echo, reverberation and passive radiation noise, effectively inhibits the interference of the reverberation and the passive radiation noise by two-stage processing of active sonar broadband spatial spectrum sharpening and spatial spectrum difference, enhances the intensity of the echo bright spot, can realize the automatic echo bright spot extraction with low false alarm rate, and provides reliable prerequisite conditions for developing active sonar autonomous detection and tracking and intelligent active detection.)

1. A method for enhancing and automatically extracting an echo bright spot of an active sonar broadband spatial spectrum is characterized by comprising the following steps: the method mainly comprises the following steps:

(1) firstly, performing space coherent processing on receiving array data through matched filtering and focused beam forming to obtain a corresponding active sonar azimuth-distance broadband spatial spectrum S (r, theta), wherein theta represents azimuth beam parameters, r represents distance point parameters, and the active sonar azimuth-distance broadband spatial spectrum S (r, theta) contains background noise, reverberation, passive radiation noise and target echo;

(2) secondly, carrying out sharpening processing on the broadband spatial spectrum obtained in the step (1) to reconstruct a sharpened spatial spectrum;

(3) according to the spatial sharpening spectrum S obtained in the step (2)_shap(r, theta), and performing spatial spectrum difference processing to obtain corresponding spatial difference spectrum S_diff(r, θ), the mathematical expression of which is shown below

S_diff(r,θ)＝(S_shap(r,θ)-S_shap(r-1,θ))/S_shap(r-1,θ)

Wherein S_diff(r, theta) represents the spatial difference spectrum, i.e. the wide-band spectrum result after spatial difference processing, S_shap(r, theta) represents a spatial sharpening spectrum, namely a sharpened azimuth-distance spatial broadband spectrum, theta represents an azimuth beam parameter, and r represents a distance point parameter;

(4) under the average background, combining a characteristic function partial differential value of a certain scale echo as a threshold gamma, and automatically selecting an echo bright spot from a regularized spatial differential spectrogram to carry out target judgment.

2. Active sonar wideband of claim 1The method for enhancing and automatically extracting the spatial spectrum echo bright spots is characterized by comprising the following steps: the sharpening step needs to select an azimuth beam window w_θSum distance point number window w_rReconstruction of a sharpened spatial spectrum S from data characterization within a search window_shap(r, θ), the mathematical calculation of which is shown below

S_w＝S(r-w_r:r+w_r,θ-w_θ:θ+w_θ)

Cr＝(S_w/mean(|S_w|))^H(S_w/mean(|S_w|))

S_shap(r,θ)＝Cr(w_θ+1,w_θ+1)

Where mean (-) represents the equalization process and | (-) represents the matrix data modulo process.

3. The active sonar wideband spatial spectrum echo bright spot enhancement and automatic extraction method according to claim 1, wherein: in step (3), the echo distance is calculated by compensating for the displacement of one point, i.e., L ═ r +1) c/(2 f)_s) Wherein L represents the distance calculated after compensation, r represents the echo distance point parameter, c represents the sound velocity, f_sThe numerical sampling rate of the current spectrogram is calculated.

4. The active sonar wideband spatial spectrum echo bright spot enhancement and automatic extraction method according to claim 1, wherein: in step (4), the calculation is represented as follows

Technical Field

The invention belongs to the field of active sonar broadband detection, and mainly relates to an active sonar broadband spatial spectrum echo bright spot enhancing and automatic extracting method.

Background

The common method for the active sonar to detect the underwater target broadband is to set a threshold to select an echo bright spot on an active broadband spectrum to judge whether the target exists or not. However, in actual broadband spectrum data, there are also influences of factors such as reverberation and strong passive radiation interference, which may cause a large number of detection blind areas or countless false alarms in active broadband detection; meanwhile, due to the influences of fluctuation of the depths of the sound source and the receiving array, array distortion, dynamic change of a target scattering cross section and the like, bright spots of echoes are enabled to generate a light and shade flicker phenomenon in different scanning periods, the phenomenon can cause discontinuous consistency of the echoes in the continuous scanning period process, report missing occurs in a single period with weak echoes, and the rapid reduction of the active sequential Bayes detection performance can also be caused.

Disclosure of Invention

The invention aims to overcome the defects in the prior art and provides an active sonar broadband spatial spectrum echo bright spot enhancing and automatic extracting method.

The invention is completed by the following technical scheme, which mainly comprises the following steps:

(1) firstly, performing space coherent processing on receiving array data through matched filtering and focused beam forming to obtain a corresponding active sonar azimuth-distance broadband spatial spectrum S (r, theta), wherein theta represents azimuth beam parameters, r represents distance point parameters, and the active sonar azimuth-distance broadband spatial spectrum S (r, theta) contains background noise, reverberation, passive radiation noise and target echo;

(2) secondly, carrying out sharpening processing on the broadband spatial spectrum obtained in the step (1) to reconstruct a sharpened spatial spectrum;

(3) according to the spatial sharpening spectrum S obtained in the step (2)_shap(r, theta), and performing spatial spectrum difference processing to obtain corresponding spatial difference spectrum S_diff(r, θ), the mathematical expression of which is shown below

S_diff(r,θ)＝(S_shap(r,θ)-S_shap(r-1,θ))/S_shap(r-1,θ)

Wherein S_diff(r, theta) represents the spatial difference spectrum, i.e. the wide-band spectrum result after spatial difference processing, S_shap(r, theta) represents a spatial sharpening spectrum, namely a sharpened azimuth-distance spatial broadband spectrum, theta represents an azimuth beam parameter, and r represents a distance point parameter;

(4) under the average background, combining a characteristic function partial differential value of a certain scale echo as a threshold gamma, and automatically selecting an echo bright spot from a regularized spatial differential spectrogram to carry out target judgment.

Furthermore, the sharpening process requires selecting the azimuth beam window w_θSum distance point number window w_rReconstruction of a sharpened spatial spectrum S from data characterization within a search window_shap(r, θ), the mathematical calculation of which is shown below

S_w＝S(r-w_r:r+w_r,θ-w_θ:θ+w_θ)

Cr＝(S_w/mean(|S_w|))^H(S_w/mean(|S_w|))

S_shap(r,θ)＝Cr(w_θ+1,w_θ+1)

Where mean (-) represents the equalization process and | (-) represents the matrix data modulo process.

Furthermore, in step (3), the echo distance needs to be calculated by compensating for the displacement of one point, i.e., L ═ r +1) c/(2 f)_s) Wherein L represents the distance calculated after compensation, r represents the echo distance point parameter, c represents the sound velocity, f_sThe numerical sampling rate of the current spectrogram is calculated.

Further, in step (4), the calculation is expressed as follows

The invention has the beneficial effects that: the invention utilizes the distribution characteristic difference of echo, reverberation and passive radiation noise, effectively inhibits the reverberation and passive radiation noise interference by two-stage processing of active sonar broadband spatial spectrum sharpening and spatial spectrum difference, enhances the intensity of echo bright spots, can realize automatic echo bright spot extraction with low false alarm rate, and provides reliable prerequisite conditions for developing active sonar autonomous detection and tracking and intelligent active detection.

Drawings

Fig. 1 shows a schematic flow chart of the bright spot extraction process of the active sonar wideband spectrum echo.

Fig. 2 shows a spectrogram sample after windowing in the sharpening process and a corresponding feature map.

Fig. 3 is a schematic diagram of actual sea trial data processing for a T period.

Fig. 4 is a schematic diagram of actual sea trial data processing for a T period.

Fig. 5 is a schematic diagram of actual sea trial data processing for a T period.

Fig. 6 is a schematic diagram of actual sea trial data processing for a T period.

Fig. 7 respectively shows a schematic diagram of a target azimuth single-beam result after T-period background normalization and difference processing.

FIG. 8 is a diagram of actual sea trial data processing for the T +1 cycle.

FIG. 9 is a schematic diagram of actual sea trial data processing for the T +1 cycle.

FIG. 10 is a schematic diagram of actual sea trial data processing for the T +1 cycle.

FIG. 11 is a diagram of actual sea trial data processing for the T +1 cycle.

Fig. 12 respectively shows a schematic diagram of a target azimuth single-beam result after T +1 period background normalization and difference processing.

Fig. 13 is a schematic diagram of actual sea trial data processing for the T +2 cycle.

FIG. 14 is a diagram of actual sea trial data processing for the T +2 cycle.

FIG. 15 is a schematic diagram of actual sea trial data processing for the T +2 cycle.

Fig. 16 is a schematic diagram of actual sea trial data processing for the T +2 cycle.

Fig. 17 respectively shows a schematic diagram of a target azimuth single-beam result after T +2 period background normalization and difference processing.

Detailed Description

The invention will be described in detail with reference to the following figures and examples:

as shown in fig. 1, an active sonar wideband spatial spectrum echo bright spot enhancement and automatic extraction method includes conventional spatial coherent processing, sharpening processing, differential processing, and echo bright spot automatic extraction, where the sharpening processing and the differential processing are the most core two processing steps, and after the two processing steps, reverberation and passive radiation interference can be sufficiently suppressed, so that an echo peak value is enhanced, and thus, the echo bright spot is easily extracted automatically; the method comprises the following specific steps:

(1) an active sonar wideband spectrum processing result is obtained according to a conventional wideband coherent processing method, and the energy spectrum result shows that the active sonar wideband spectrum processing result contains background noise, reverberation, passive radiation noise and target echo, and sometimes target echo which is almost not interfered mound and is difficult to distinguish is also obtained; firstly, performing space coherent processing on receiving array data through matched filtering and focused beam forming to obtain a corresponding active sonar azimuth-distance broadband space spectrum S (r, theta), wherein theta represents azimuth beam parameters, and r represents distance point parameters;

(2) secondly, carrying out sharpening processing on the broadband spatial spectrum obtained in the step (1) to reconstruct a sharpened spatial spectrum; the processing step can highlight point radial echoes and radial passive noise, has a certain suppression effect on reverberation with the characteristic of gradually-decreasing K distribution, and has a certain enhancement effect on weaker echoes and passive radiation noise.

The sharpening step needs to select an azimuth beam window w_θSum distance point number window w_rReconstruction of a sharpened spatial spectrum S from data characterization within a search window_shap(r, θ), the mathematical calculation of which is shown below

S_w＝S(r-w_r:r+w_r,θ-w_θ:θ+w_θ)

Cr＝(S_w/mean(|S_w|))^H(S_w/mean(|S_w|))

S_shap(r,θ)＝Cr(w_θ+1,w_θ+1)

Where mean (-) represents the equalization process and | (-) represents the matrix data modulo process. As shown in fig. 2, a spectrogram sample after windowing in the sharpening process and a corresponding feature map are given.

(3) The sharpened active broadband spectrum is subjected to distance-dimensional spatial spectrum differential processing, the mathematical meaning of the step is that partial differential processing is carried out on the spatial spectrum, because the echo characteristics on the broadband spectrum present pulse-shaped characteristics, while the passive radiation noise presents continuous strip-shaped characteristics, if the echo characteristics are approximated to characteristic functions, the partial derivative of the echo characteristic functions is far larger than that of the passive radiation noise characteristic functions, and for uninterested scatterers with larger radial dimension, the partial derivative of the broadband spectrum characteristic functions is also far smaller than that of underwater targets with small radial dimension. According to the spatial sharpening spectrum S obtained in the step (2)_shap(r, theta), and performing spatial spectrum difference processing to obtain corresponding spatial difference spectrum S_diff(r, θ), the mathematical expression of which is shown below

S_diff(r,θ)＝(S_shap(r,θ)-S_shap(r-1,θ))/S_shap(r-1,θ)

Wherein S_diff(r, theta) represents the spatial difference spectrum, i.e. the wide-band spectrum result after spatial difference processing, S_shap(r, theta) represents a spatial sharpening spectrum, namely a sharpened azimuth-distance spatial broadband spectrum, theta represents an azimuth beam parameter, and r represents a distance point parameter;

it should be noted that, for the first point on the original spatial spectral distance, the difference processing cannot be implemented, so the information of the first distance point is lost in the result after the difference, however, the first points of the wide-band spectral distance dimension of the active sonar are both direct waves and reverberation, so this influence can be ignored, but the displacement of one point needs to be compensated when calculating the echo distance, i.e. L ═ r +1) c/(2 f)_s) Wherein L represents the distance calculated after compensation, r represents the echo distance point parameter, c represents the sound velocity, f_sThe numerical sampling rate of the current spectrogram is calculated.

(4) Under the average background, combining the partial differential value of the characteristic function of a certain scale echo as a threshold gamma, automatically selecting an echo bright spot from a regularized spatial differential spectrogram to carry out target judgment, and calculating and expressing the following steps

Fig. 3 is actual sea test data processing for a T period, and shows an active sonar wideband spatial spectrum output after conventional coherent processing, where (a) is a three-dimensional display diagram of the spatial spectrum, (b) is a two-dimensional display diagram of the spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. It can be seen from the figure that the reverberation is suppressive with respect to the intensity level of the target, and the distribution of the reverberation shows a trend of fading-up and fading-down.

Fig. 4 is actual sea test data processing for a T period, and shows a spatial spectrum output after background normalization processing, where (a) is a three-dimensional display diagram of the background normalized spatial spectrum, (b) is a two-dimensional display diagram of the background normalized spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. Comparing fig. 1 and fig. 2, it can be seen that after the background normalization processing, the reverberation is well suppressed, however, the background intensity is significantly increased, and the intensity of the target is correspondingly faded down.

Fig. 5 is actual sea test data processing for a T period, and a spatial spectrum output after sharpening is given, and an input of the spatial spectrum is a broadband spatial spectrum after conventional processing, where (a) is a three-dimensional display diagram of the sharpened spatial spectrum, (b) is a two-dimensional display diagram of the sharpened spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. As can be seen by comparing fig. 2 and 3, the sharpening process not only suppresses reverberation well, but also highlights bright spots of echoes.

Fig. 6 is actual sea test data processing for a T period, and a spatial spectrum output after differential processing is given, and an input of the spatial spectrum is a sharpened spatial spectrum, where (a) is a three-dimensional display diagram of the differential spatial spectrum, (b) is a two-dimensional display diagram of the differential spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. As can be seen by comparing fig. 2 and 3, the sharpening process not only suppresses reverberation well, but also highlights bright spots of echoes.

Fig. 7 shows the target azimuth single-beam results after T period background normalization and difference processing, respectively, and it can be seen from the figure that the single-beam effect after difference processing is obviously better than the background normalization output single-beam effect.

Fig. 8 is actual sea test data processing for a period T +1, and shows an active sonar wideband spatial spectrum output after conventional coherent processing, where (a) is a three-dimensional display diagram of the spatial spectrum, (b) is a two-dimensional display diagram of the spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. It can be seen from the figure that the reverberation is suppressive with respect to the intensity level of the target, and the distribution of the reverberation shows a trend of fading-up and fading-down.

Fig. 9 is actual sea test data processing for a period T +1, and shows a spatial spectrum output after background normalization processing, where (a) is a three-dimensional display diagram of the background normalized spatial spectrum, (b) is a two-dimensional display diagram of the background normalized spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. Comparing fig. 1 and fig. 2, it can be seen that after the background normalization processing, the reverberation is well suppressed, however, the background intensity is significantly increased, and the intensity of the target is correspondingly faded down.

Fig. 10 is actual sea test data processing for a period T +1, and shows a spatial spectrum output after sharpening, and an input of the spatial spectrum is a broadband spatial spectrum after conventional processing, where (a) is a three-dimensional display diagram of the sharpened spatial spectrum, (b) is a two-dimensional display diagram of the sharpened spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. As can be seen by comparing fig. 2 and 3, the sharpening process not only suppresses reverberation well, but also highlights bright spots of echoes.

Fig. 11 is actual sea test data processing for a period T +1, and a spatial spectrum output after differential processing is given, and an input of the spatial spectrum is a sharpened spatial spectrum, where (a) is a three-dimensional display diagram of the differential spatial spectrum, (b) is a two-dimensional display diagram of the differential spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. As can be seen by comparing fig. 2 and 3, the sharpening process not only suppresses reverberation well, but also highlights bright spots of echoes.

Fig. 12 shows the target azimuth single-beam results after T +1 period background normalization and difference processing, respectively, and it can be seen from the figure that the single-beam effect after difference processing is significantly better than the background normalization output single-beam effect.

Fig. 13 is actual sea test data processing for a period T +2, and shows an active sonar wideband spatial spectrum output after conventional coherent processing, where (a) is a three-dimensional display diagram of the spatial spectrum, (b) is a two-dimensional display diagram of the spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. It can be seen from the figure that the reverberation is suppressive with respect to the intensity level of the target, and the distribution of the reverberation shows a trend of fading-up and fading-down.

Fig. 14 is actual sea test data processing for a T +2 period, and shows a spatial spectrum output after background normalization processing, where (a) is a three-dimensional display diagram of the background normalized spatial spectrum, (b) is a two-dimensional display diagram of the background normalized spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. Comparing fig. 1 and fig. 2, it can be seen that after the background normalization processing, the reverberation is well suppressed, however, the background intensity is significantly increased, and the intensity of the target is correspondingly faded down.

Fig. 15 is actual sea test data processing for a period T +2, and shows a spatial spectrum output after sharpening, and the input of the spatial spectrum is a broadband spatial spectrum after conventional processing, where (a) is a three-dimensional display diagram of the sharpened spatial spectrum, (b) is a two-dimensional display diagram of the sharpened spatial spectrum, and (b) a position indicated by a red circle is an echo bright point of a target under test. As can be seen by comparing fig. 2 and 3, the sharpening process not only suppresses reverberation well, but also highlights bright spots of echoes.

Fig. 16 is actual sea test data processing for a period T +2, and a spatial spectrum output after differential processing is given, and an input of the spatial spectrum is a sharpened spatial spectrum, where (a) is a three-dimensional display diagram of the differential spatial spectrum, (b) is a two-dimensional display diagram of the differential spatial spectrum, and a position indicated by a red circle in (b) is an echo bright point of a target in a test. As can be seen by comparing fig. 2 and 3, the sharpening process not only suppresses reverberation well, but also highlights bright spots of echoes.

Fig. 17 shows the target azimuth single-beam results after T +2 period background normalization and difference processing, respectively, and it can be seen from the figure that the single-beam effect after difference processing is obviously better than the background normalization output single-beam effect.

It should be understood that the technical solutions and inventive concepts of the present invention may be equally replaced or changed by those skilled in the art, and the technical solutions and inventive concepts should also belong to the scope of the appended claims.