Multi-beam backscattering intensity jump correction method

Abstract

The invention discloses a multi-beam backscattering intensity jump correction method, which takes a multi-beam measurement strip as a unit, automatically detects backscattering intensity jump generated by factors such as equipment parameter setting through a Gaussian weighting function, and selects a moving window to perform difference compensation by using a weighted average value. And carrying out iterative detection on the repaired intensity data and repairing the jump point until the jump point converges to 0. An example calculation result shows that the method can effectively detect and repair the intensity jump in the multi-beam strip, improve the processing efficiency and quality of the backscattering intensity data of the multi-beam strip, and improve the analysis accuracy of the submarine landform of the multi-beam intensity data.

Description

Technical Field

[0001] The invention relates to the field of ocean detection technology, and in particular to the field of a multi-beam backscatter intensity jump correction method. Background Art

[0002] With the rapid development of marine engineering construction, marine resource development, submarine pipeline detection and other fields, it is very important to accurately grasp the seabed geological types. As one of the core technologies of marine detection, the multi-beam sonar system can generate backscatter intensity data and three-dimensional terrain images of the seabed through the transmission and reception of sonar signals. These data are widely used in underwater measurement, marine exploration, environmental monitoring, submarine cable laying and other fields. However, due to the complexity of the marine environment and the limitations of the sonar system hardware equipment, the backscatter intensity data collected in actual applications often have deviations and jumps, which seriously affect the accuracy of the data and subsequent analysis and processing. Therefore, how to effectively correct these data has become a technical problem that needs to be solved urgently.

[0003] In practical applications, the deviation and jump of backscatter intensity data are mainly caused by the following factors: Terrain complexity: The propagation path of sound waves in complex terrain will be affected by various interferences of the underwater environment, such as the undulation of the seabed terrain, the change of geological type, etc., which causes the sound wave reflection signal to show different intensity distribution at different angles and positions, thereby causing data deviation. System parameter error: The parameter settings of the sonar equipment itself, such as sonar frequency, the response of the receiving array, the transmission power, time gain compensation (TVG), etc., will affect the collected backscatter intensity data. For example, too high or too low gain will cause the echo signal to be over-enhanced or weakened, affecting the authenticity of the measurement data. Environmental factors: Environmental factors such as suspended matter, temperature, salinity, and noise in the water body will cause energy loss and changes in reflection intensity during signal propagation, thereby affecting the collection quality of sonar data. These factors cannot be corrected by simple parameter adjustment. Data jump: The common strip jump phenomenon in the data collected by multi-beam sonar, that is, between consecutive Pings (a signal transmission and reception), the backscatter intensity data suddenly changes, forming data discontinuity. This jump may be caused by equipment switching, signal interference, etc., which directly affects the overall quality of the sonar image.

[0004] In response to the above problems, domestic and foreign scholars have conducted a lot of research and tried to correct the errors in the backscattering intensity data through different methods. However, the existing technology still has great limitations in solving the problem of data jump.

[0005] Some solutions have studied the various factors that affect the parameters of multi-beam sonar systems and their specific effects. Studies have shown that errors in system parameters, such as gain settings of sonar equipment and time gain compensation, can lead to deviations in backscatter intensity data. However, this study is limited to analyzing the causes of the errors and does not provide an effective correction method. Some other solutions have proposed a method to reduce system setting errors by measuring the responses of the receiving and transmitting arrays, combining linear dynamic response correction and complex weighted correction to address the correction problem of multi-beam system parameters. This method effectively reduces the signal deviation caused by inaccurate device parameter settings by accurately measuring the device response characteristics. However, this method is mainly based on correcting hardware parameters, and cannot completely solve the problem of backscatter intensity data jumps caused by the environment or other complex factors. Some solutions have proposed an empirical correction method for the angle response problem in backscatter intensity data, which can reduce angle-dependent artifacts and improve image uniformity. By calculating and removing angle artifacts in backscatter intensity, this method can improve the accuracy of sonar images to a certain extent. Some other solutions have adopted a comprehensive processing and analysis technology, especially to correct the angle dependency of backscatter intensity. However, these methods cannot effectively deal with the problem of jumps in data. There are also some methods that reprocess the Snippet echo intensity. These studies have made different treatments for discontinuities in the data, but most of them focus on improving the overall grayscale uniformity of the image and eliminating local anomalies, and cannot solve the core problem of backscatter intensity data jumps within the strip.

[0006] At present, the correction of multi-beam backscatter intensity system mainly focuses on the correction of hardware parameters and the processing of angle response. Although these methods have achieved certain results in reducing system errors and improving image uniformity, there are still many deficiencies in the processing of strip intensity jump phenomenon:

[0007] Limitations of system parameter correction: Most existing methods rely on adjusting system hardware parameters, such as gain, transmit power, etc. However, the hardware characteristics of the device are not the only cause of the jump phenomenon. Complex environmental factors and device switching can also cause jumps. The current technical means to correct these factors are relatively limited.

[0008] Angle dependence: Although correction methods for angle-dependent artifacts effectively reduce the impact of geometric artifacts on images, these methods are mainly suitable for smoothing signal fluctuations caused by angle changes in images, and do not significantly improve sudden jump phenomena.

[0009] Lack of automated processing mechanism for jumps: Most existing methods rely on manually set parameters or manual intervention to process jump data, lacking an automated and intelligent jump detection and repair mechanism. This makes the existing technology less adaptable in practical applications, especially in the changing marine environment, where manual adjustments are difficult to cope with complex situations. Summary of the invention

[0010] In response to the problems existing in the prior art, this patent proposes a multi-beam backscatter intensity jump correction method, aiming to solve the following technical problems: Automatic detection of jump points: By introducing Gaussian weighted functions and dynamic threshold detection algorithms, the jump points in the data strips can be automatically and accurately detected, avoiding the impact of manual intervention on accuracy and efficiency. Intelligent repair of jump data: Based on Gaussian weighted difference compensation technology, weighted processing is performed on the data segments before and after the jump point to ensure the smoothness and consistency of the data after repair, and it can effectively adapt to the jump repair needs in different environments. Enhance the adaptability and robustness of data: This method can adapt to different sonar working environments, whether it is complex terrain or harsh marine environment, it can effectively identify and repair the jump phenomenon in the data, and improve the quality and consistency of backscatter data. In order to achieve the above purpose, the present invention discloses a multi-beam backscatter intensity jump correction method, which specifically includes the following steps:

[0011] S1: data file import;

[0012] S2: extracting intensity data, preprocessing the single-strip multi-beam backscatter intensity data, wherein the preprocessing includes time gain compensation (TVG) and sonar parameter correction to obtain the corrected backscatter intensity data;

[0013] The backscatter intensity data is collected by a multi-beam bathymetric system and is usually recorded in units of Ping. Each Ping records the backscatter signals of hundreds or thousands of seabed points obtained by a complete signal transmission and reception cycle of a multi-beam transducer at a certain moment. Each strip of data contains multiple continuous Pings. The backscatter intensity data of each Ping is used to reflect the signal echo reflection strength of the seabed within a certain range in the vertical track direction.

[0014] S3: intensity jump detection, detecting jumps based on a Gaussian weighted function, specifically including the following steps:

[0015] S3-1: Calculate the Gaussian weighted average value, input the backscattering intensity data after preprocessing in step S2, and use the Gaussian weighted function method to detect jumps. The Gaussian weighted average value corresponding to each Ping is:

[0016]

[0017] Among them, x i represents the intensity value of the i-th data, ω i is the Gaussian weight;

[0018] S3-2: Calculate the Gaussian weighted mean difference. For a series of Pings whose Gaussian weighted mean has been calculated, record it as Calculate the difference in Gaussian weighted averages between adjacent pings. The specific calculation formula is as follows:

[0019]

[0020] in, Indicates the intensity value of the jth data, ΔI j represents the difference of Gaussian weighted means;

[0021] S3-3: Dynamic threshold detection jump point, calculate the average of the intensity difference between all pings and the standard deviation σ of all intensity differences ΔI , and accordingly set the dynamic threshold to detect the jump point. The specific calculation formula is as follows:

[0022]

[0023] Obtained through and σ ΔI Define the dynamic threshold as T:

[0024]

[0025] When ΔI>T, it means that the Ping has a jump, and the Ping is marked;

[0026] S4: Determine the data segment that does not need to be compensated for the jump, specifically including the following steps:

[0027] S4-1: Calculate the weighted average value within the window W before and after the jump point. Specifically, for each jump point, take the strength values ​​within the window W forward and backward with the jump point as the center, and calculate the weighted average value as the reference value for repair;

[0028] S4-2: all the jump points are processed in sections in order, and every two adjacent jump points determine a jump data segment. For each jump data segment, the weighted average value of the start and end positions is calculated to determine the compensation difference of the jump data segment;

[0029] S5: Intensity jump repair: All Pings in the jump data segment are repaired based on the above compensation value, and the data is adjusted to be consistent with the fluctuation range of the data at both ends. After compensation, the weighted average value of each data in the first two jump segments is and Then the weighted average value of the Ping starting from the third jump segment after compensation is

[0030] Steps S3-S5 are performed again on the strength data after jump repair until the iteration stops when there is no jump point.

[0031] S6: Export the repaired strength data, that is, all the Pings repaired in step S5.

[0032] As a technical preferred solution of the present invention, the preprocessing step in step S2 includes removing the influencing factors of the system parameters and the human or operating system to ensure the authenticity of the backscattering intensity data, specifically including compensating for the time gain (TVG), and the calculation formula is:

[0033] TVG=Sp·log 10 (R)+2·α·R (6)

[0034] Among them, Sp is the propagation loss coefficient, R is the distance between the target and the sonar, and α is the attenuation coefficient in the propagation medium. By applying this formula, the echo intensity value of each Ping is compensated to correct the energy loss caused by distance and medium attenuation.

[0035] As a technical preferred solution of the present invention, the sonar parameter correction in the preprocessing step in step S2 includes correction of system parameters, specifically correcting the received signal by the system gain constant. The calculation formula of the sonar parameter correction is:

[0036] I corrected =I received +G sonar (7)

[0037] In the formula, I received is the original signal strength received, G sonar It is the system gain constant, which is used to correct the signal gain value increased manually or automatically by the system.

[0038] As a technical preferred solution of the present invention, the process of setting the dynamic threshold in step S3-3 further includes noise smoothing processing and dynamic adjustment of the standard deviation weight factor to adapt to the needs of different sonar working environments. The dynamic threshold calculation formula is:

[0039]

[0040] Among them, k is a dynamically adjusted weight factor, which is adjusted according to the actual working environment of the sonar and the characteristics of data changes; N fIt is a noise smoothing factor used to remove the influence of random noise in the data. This factor is calculated by the Kalman filter algorithm.

[0041] As a technical preferred solution of the present invention, the jump repair in step S5 includes iterative detection and compensation repair of the jump point, specifically including the following steps:

[0042] T1: After completing a jump point compensation repair, repeat steps S3-S5 and record the newly detected jump point;

[0043] T2: For the newly detected jump points, weighted average calculation and data compensation are performed again to gradually correct all jump points;

[0044] T3: Iteratively repeat the repair process of step S5 to gradually reduce the repair error until the number of jump points converges to zero.

[0045] As a technical preferred solution of the present invention, the processing results of step S1 to step S2 also include preset conditions, and the preset conditions are specifically:

[0046] The multi-beam intensity is processed in units of a single strip, and the data is corrected for system parameters to remove the influencing factors introduced by human or operating systems, and a complete strip data is selected as the source of correction data;

[0047] In the process of using Gaussian weighted automatic detection of jumps, the statistical calculation of intensity differences is introduced to set the dynamic threshold based on it. The dynamic threshold changes with the data and can maintain good adaptability in different situations.

[0048] All the jump points are segmented in sequence, and every two adjacent jump points determine a jump data segment. The intensity values ​​in the window W are taken before and after the jump point, and the weighted average value is calculated as the compensation difference.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] Automated jump detection and repair: By introducing Gaussian weighted function and dynamic threshold detection algorithm, automatic jump point detection can be achieved, eliminating the tedious steps of relying on manual setting and adjustment in the existing technology, and improving efficiency and accuracy.

[0051] Enhance data smoothness and consistency: Based on Gaussian weighted difference compensation technology, intelligent repair is performed before and after the jump point to ensure smooth transition of data, effectively eliminate the discontinuity caused by the jump, and improve the quality of sonar images.

[0052] Adapt to different environments and equipment: This method can dynamically adjust noise processing and threshold parameters to adapt to sonar data collection in different ocean environments. It has strong applicability and can still operate stably under changeable and complex ocean conditions to ensure data quality.

[0053] Improved data accuracy and system robustness: Through system parameter correction and time gain compensation, the intensity deviation caused by equipment errors and environmental factors is corrected, and the overall accuracy and robustness of the backscatter data are improved.

[0054] Reducing human intervention: The present invention reduces the need for human intervention, reduces operational complexity, and enhances the intelligence level of the system through automated jump detection and repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flow chart of a multi-beam backscatter intensity jump correction method of the present invention;

[0056] Figure 2 is a schematic diagram of a transition point ad in an original image according to an embodiment of the present invention;

[0057] Figure 3 is a cross-sectional view along the direction of the jump point ad of an embodiment of the present invention;

[0058] Figure 4 is a jump marking result of an embodiment of the present invention;

[0059] Figure 5 is the position of the partial jump marked in the embodiment of the present invention;

[0060] Figure 6 This is a comparison diagram of images before and after restoration of the jump position in an embodiment of the present invention;

[0061] Figure 7 This is a comparison diagram before and after the repair of the data segment at the jump point abcd in an embodiment of the present invention;

[0062] Figure 8 It is a statistical diagram of the change range before and after the jump point is repaired in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The present invention is further described below in conjunction with the accompanying drawings and examples. However, the present invention can be implemented in many different ways and should not be construed as being limited to the embodiments shown; on the contrary, these embodiments provide those skilled in the art with implementation methods that meet applicable legal requirements.

[0064] Example 1: Figure 1 As shown, the multi-beam backscatter intensity jump correction method described in the present invention specifically includes the following steps:

[0065] S1: data file import;

[0066] S2: extracting intensity data, preprocessing the single-strip multi-beam backscatter intensity data, wherein the preprocessing includes time gain compensation (TVG) and sonar parameter correction to obtain the corrected backscatter intensity data;

[0067] The backscatter intensity data is collected by a multi-beam bathymetric system and is usually recorded in units of Ping. Each Ping records the backscatter signals of hundreds or thousands of seabed points obtained by a complete signal transmission and reception cycle of a multi-beam transducer at a certain moment. Each strip of data contains multiple continuous Pings. The backscatter intensity data of each Ping is used to reflect the signal echo reflection strength of the seabed within a certain range in the vertical track direction.

[0068] S3: intensity jump detection, detecting jumps based on a Gaussian weighted function, specifically including the following steps:

[0069] S3-1: Calculate the Gaussian weighted average value, input the backscattering intensity data after preprocessing in step S2, and use the Gaussian weighted function method to detect jumps. The Gaussian weighted average value corresponding to each Ping is:

[0070]

[0071] Among them, x i represents the intensity value of the i-th data, ω i is the Gaussian weight;

[0072] S3-2: Calculate the Gaussian weighted mean difference. For a series of Pings whose Gaussian weighted mean has been calculated, record it as Calculate the difference in Gaussian weighted averages between adjacent pings. The specific calculation formula is as follows:

[0073]

[0074] in, Indicates the intensity value of the jth data, ΔI j represents the difference of Gaussian weighted means;

[0075] S3-3: Dynamic threshold detection jump point, calculate the average of the intensity difference between all pings and the standard deviation σ of all intensity differences ΔI , and accordingly set the dynamic threshold to detect the jump point. The specific calculation formula is as follows:

[0076]

[0077] Obtained through and σ ΔI Define the dynamic threshold as T:

[0078]

[0079] When ΔI>T, it means that the Ping has a jump, and the Ping is marked;

[0080] S4: Determine the jump compensation data segment, specifically including the following steps:

[0081] S4-1: Calculate the weighted average value within the window W before and after the jump point. Specifically, for each jump point, take the strength values ​​within the window W forward and backward with the jump point as the center, and calculate the weighted average value as the reference value for repair;

[0082] S4-2: all the jump points are processed in sections in order, and every two adjacent jump points determine a jump data segment. For each jump data segment, the weighted average value of the start and end positions is calculated to determine the compensation difference of the jump data segment;

[0083] S5: Transition repair: All Pings in the transition data segment are repaired based on the above compensation value, and the data is adjusted to be consistent with the fluctuation range of the data at both ends. After compensation, the weighted average value of each data in the first two transition segments is and Then the weighted average value of the Ping starting from the third jump segment after compensation is

[0084] Steps S3-S5 are performed again on the strength data after jump repair until the iteration stops when there is no jump point.

[0085] S6: Export the repaired strength data, that is, all the Pings repaired in step S5.

[0086] The preprocessing step in step S2 includes removing the influencing factors of system parameters and human or operating system to ensure the authenticity of the backscattering intensity data, specifically including compensating for the time gain (TVG), which is calculated as follows:

[0087] TVG=Sp·log 10 (R)+2·α·R (6)

[0088] Among them, Sp is the propagation loss coefficient, R is the distance between the target and the sonar, and α is the attenuation coefficient in the propagation medium. By applying this formula, the echo intensity value of each Ping is compensated to correct the energy loss caused by distance and medium attenuation.

[0089] The single-strip multi-beam backscatter intensity data was corrected in terms of TVG and sonar parameter correction. After processing, the backscatter intensity data obtained is basically considered to have eliminated or weakened the influence of system parameters. The backscatter intensity data obtained after preprocessing the multi-beam backscatter intensity data is basically considered to have eliminated or weakened the influence of system parameters. However, the original image shows that there are obvious gray value difference jump points in the processed data strip image. Some jump points and their profiles along the jump point direction are shown in the figure. Figure 2 , Figure 3 shown.

[0090] Simple mean calculation is difficult to reflect the changing trend of backscatter intensity data, making it difficult to effectively detect and repair jumps. To solve this problem, the Gaussian weighted function method is tried. The principle of detecting jumps based on the Gaussian weighted function is as follows:

[0091] Given a series of Pings containing intensity values, assume that each Ping contains n intensity values, let {x1,x2,…,x n}, first calculate the mean of the intensity differences between all Pings and the standard deviation σ of all intensity differences ΔI , and accordingly set the dynamic threshold to detect the jump point. The specific calculation formula is as follows:

[0092]

[0093] Where: x i Represents the intensity value of the i-th data. Calculate the Gaussian weighted average of each Ping based on the obtained μ and σ, and define the Gaussian weight as ω i , then the Gaussian weighted average corresponding to each Ping is

[0094] After obtaining the Gaussian weighted average of each Ping in the data, transition detection and marking are required. For a series of Pings for which the Gaussian weighted average has been obtained: Calculate the difference ΔI between the Gaussian weighted averages of adjacent pings j , the specific calculation formula is as follows:

[0095]

[0096] in, Indicates the intensity value of the jth data, ΔI j represents the difference of Gaussian weighted means;

[0097] The intensity difference between adjacent Pings reflects the magnitude of change between adjacent data points, but it is difficult to determine whether these changes are normal fluctuations or abnormal jumps based on this degree of change alone. Therefore, the statistical calculation of intensity differences is introduced, and a dynamic threshold is set accordingly. Since the dynamic threshold changes with the data, it can maintain good adaptability in different situations and always remain sensitive to jumps. In contrast, a fixed threshold may increase the probability of false detection or leak detection. In order to statistically analyze the significance of the intensity change difference, it is necessary to calculate the mean of the intensity difference between all Pings. and the standard deviation σ of all intensity differences ΔI , and set the dynamic threshold to detect the jump point. and σ ΔI Define the dynamic threshold as T:

[0098]

[0099] When ΔI>T, it means that there is a jump in the Ping, and the Ping is marked.

[0100] Based on the above theory of jump detection, when the multi-beam strip intensity data is imported and jump detection is performed, the detected jump Ping will be marked and displayed in the window (see Figure 4 ).

[0101] In order to verify the correctness of the algorithm, the saved jump Ping is displayed and compared with the original data to visually observe the detection effect. After comparison, it is found that the marked jump Ping is accurately located at the position where the jump exists in the original image (see Figure 5 ). This result proves the accuracy of the algorithm.

[0102] The jump repair is achieved based on Gaussian weighted difference compensation. The marked jump is defined as {y i}, in order to select the appropriate data segment to compensate the data segment starting from the jump point, a moving window W is defined, and the size of W determines the width of the selected window. In order to calculate the compensation value, it is necessary to first calculate μ p and μ n , the calculation formula is as follows:

[0103]

[0104]

[0105] Where: y i is the number of the i-th jump point; μ p is the jump point y i The weighted average of the data segment of size W in the previous window; μ n is the jump point yi The weighted average of the data segments of size W in the next window.

[0106] After calculating the weighted average of the previous and next data segments, it is necessary to calculate the weighted average difference between the two data segments as the compensation value. However, considering that the entire jump segment is convex or concave in the entire strip, the directionality needs to be considered when selecting a moving window to use the weighted average for difference compensation. By comparison, it is found that according to the order of the jump points, regardless of whether the jump segment is "convex" or "concave", the previous jump point always uses the previous window data to compensate for the next window data, while the next jump point is the opposite. Therefore, all the jump points must first be segmented in order. Every two adjacent jump points determine a jump data segment. In each jump data segment, the compensation difference ΔV is calculated. i and ΔV i+1 , the specific calculation formula is:

[0107] ΔV i =μ p -μ n (14)

[0108] ΔV i+1 =μ n -μ p (15)

[0109] Where: ΔV i is the compensation value of a data segment of size W after the i-th jump point; ΔV i+1 It is the compensation value of the W-size data segment before the i+1th jump point.

[0110] In order to ensure the continuity of compensation and the stability of intensity data, the accumulation of compensation values ​​needs to be considered. Therefore, starting from the third jump segment, the compensation value of the previous segment needs to be added to obtain the final compensation value. Therefore, the compensation value of the third jump point is defined as ΔV k (k≥3):

[0111]

[0112] Based on the above compensation value, the part with jump in the selected window is compensated. After the first two jump segments are compensated, the weighted average value of each data is: and Then the weighted average value of the Ping starting from the third jump segment after compensation is

[0113]

[0114] The detection method based on the above-mentioned Gaussian weighted dynamic threshold setting can accurately detect adjacent Pings that exceed the threshold limit between different data, and mark these jump Pings in real time. Through the Gaussian weighted moving window compensation method, the difference in the weighted average value of the data segment of the window size selected before and after the jump point is calculated for compensation and repair, and the data of the jump segment is adjusted to the same fluctuation range as the data at both ends of the jump segment. After the first repair is completed, iterative detection and repair begins, gradually reducing the number of jump points, and finally making the jump points completely converge to 0, indicating that the repair of the strip has been completed.

[0115] Based on the jump repair theory, the window size is set to 50 Ping. When all jump points are repaired, it is found by comparing with the original image that the strip image has been repaired and balanced. Figure 6 shown.

[0116] In order to verify the accuracy of the repair, the weighted average value of the jump segment before and after the repair was analyzed. The above four jumps were selected, and the data of 25 Ping before and after the jump point were taken. Some cross-sections were drawn for comparative analysis before and after the repair, as shown in the figure. Figure 7 By comparing the line graph of the weighted average values ​​before and after the repair of the three obvious jump points, it can be seen that before the repair, there was a huge difference between the two sections before and after the jump point, and the data change abnormality showed an obvious jump. After the Gaussian weighted average difference compensation repair was achieved through the moving window, the difference between the front and back sections was significantly reduced, and the jump section and the overall data section tended to be flat, indicating that the data was effectively smoothed. Figure 7 It shows that the detection of jump points and the repair of jump segments have achieved significant results overall.

[0117] The standard deviation of a piece of data can better reflect the degree of dispersion of this piece of data. Therefore, the standard deviation can reflect the change range and trend of the data before and after the repair. Figure 8 This is a statistical chart of the change range before and after the jump point repair. For the repaired data sequence Calculate the overall mean value before repair of this selected section and record it as The overall mean before repair is recorded as The standard deviation of the selected data segment before processing is σ 前 The standard deviation after processing is σ 后 :

[0118]

[0119] Where: is the weighted average of the iPing before repair; It is the weighted average of the iPing after repair.

[0120] Table 1: Standard deviation statistics of the data segments before and after the jump point before and after repair

[0121] <![CDATA[Standard deviation σ before repair 前 > <![CDATA[Standard deviation σ after repair 后 > Data segments before and after the jump point a 1.724 0.4536 Data segments before and after the jump point b 1.9203 0.3711 Data segment before and after the jump point c 1.755 0.5235 Data segment before and after the jump point d 1.8564 0.5902

[0122] As shown in Table 1, before the repair, there are large fluctuations in the data segments before and after the jump point, resulting in a large standard deviation of the segments before and after, indicating that the overall data tends to be discrete. After compensation and repair, the overall data tends to fluctuate smoothly, and the standard deviation of the segments before and after is small. The repair effectively reduces the fluctuation amplitude of the jump point and achieves smooth repair of the overall jump point.

[0123] A Gaussian weighted method is used to detect and repair jumps in multi-beam backscatter intensity data. The jump point can be effectively detected by setting the dynamic threshold through weighted average. During repair, the difference compensation of the weighted average values ​​of the data segments before and after the jump point is used, and the jump data segment is compensated by the part without jump. Automatic iterative detection and repair processing is achieved through programming.

[0124] Through the above embodiments, compared with the existing methods for correcting multi-beam system parameters at home and abroad, this method is based on the weighted average of the backscatter intensity data using statistical methods, and performs better smoothing repair on the intensity jump data, removing the jump part in the strip to the maximum extent. The obtained image effect is more balanced and of higher quality than the correction of the hardware parameters and other aspects of the multi-beam system. In summary, this method can effectively correct the intensity jump caused by system parameters in the multi-beam backscatter intensity, and can serve as an important reference for this type of correction.

[0125] The above embodiments only express the implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention. It should be pointed out that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention.

Claims

1. A multi-beam backscatter intensity jump correction method, characterized in that: The method comprises the following steps: S1: data file import; S2: Extraction of intensity data, pre-processing of single-strip multi-beam backscatter intensity data, the pre-processing includes time gain compensation TVG and sonar parameter correction, to obtain corrected backscatter intensity data; wherein the backscatter intensity data is collected by the multi-beam bathymetric system, usually recorded in units of Ping, each Ping records hundreds or thousands of seabed point backscatter signals obtained by a complete signal transmission and reception cycle of the multi-beam transducer at a certain moment, each strip of data contains multiple continuous Pings, and the backscatter intensity data of each Ping is used to reflect the signal echo reflection strength of the seabed within a certain range perpendicular to the track direction; S3: intensity jump detection, detecting jumps based on a Gaussian weighted function, specifically including the following steps: S3-1: Calculate the Gaussian weighted average value, input the backscattering intensity data after preprocessing in step S2, and use the Gaussian weighted function method to detect jumps. The Gaussian weighted average value corresponding to each Ping is: Among them, x i represents the intensity value of the i-th data, ω i is the Gaussian weight, n is the total number of Pings; S3-2: Calculate the Gaussian weighted mean difference. For a series of Pings whose Gaussian weighted mean has been calculated, record it as Calculate the difference in Gaussian weighted averages between adjacent pings. The specific calculation formula is as follows: in, Indicates the intensity value of the jth data, ΔI j represents the difference of Gaussian weighted means; S3-3: Dynamic threshold detection jump point, calculate the average of the intensity difference between all pings and the standard deviation σ of all intensity differences ΔI , and accordingly set the dynamic threshold to detect the jump point. The specific calculation formula is as follows: Obtained through and σ ΔI Define the dynamic threshold as T: When ΔI>T, it means that the Ping has a jump, and the Ping is marked. m represents the total number of differences in the Gaussian weighted average. S4: Determine the jump compensation data segment, specifically including the following steps: S4-1: Calculate the weighted average value within the window W before and after the jump point. Specifically, for each jump point, take the strength values ​​within the window W forward and backward with the jump point as the center, and calculate the weighted average value as the reference value for repair; S4-2: all the jump points are processed in sections in order, and every two adjacent jump points determine a jump data segment. For each jump data segment, the weighted average value of the start and end positions is calculated to determine the compensation difference of the jump data segment; S5: Intensity jump repair: All Pings in the jump data segment are repaired based on the above compensation value, and the data is adjusted to be consistent with the fluctuation range of the data at both ends. After compensation, the weighted average value of each data in the first two jump segments is and Then the weighted average value of the Ping starting from the third jump segment after compensation is Steps S3-S5 are performed again on the strength data after jump repair until the iteration stops when there is no jump point. S6: Export the repaired strength data, that is, all the Pings repaired in step S5.

2. A multi-beam backscatter intensity jump correction method according to claim 1, characterized in that: The preprocessing step in step S2 includes removing the influencing factors of system parameters and human or operating system to ensure the authenticity of the backscattering intensity data, specifically including compensating for the time gain TVG, and the calculation formula is: TVG=Sp log 10 (R)+2·α·R (6) Among them, Sp is the propagation loss coefficient, R is the distance between the target and the sonar, and α is the attenuation coefficient in the propagation medium. By applying this formula, the echo intensity value of each Ping is compensated to correct the energy loss caused by distance and medium attenuation.

3. A multi-beam backscatter intensity jump correction method according to claim 1, characterized in that: The sonar parameter correction in the preprocessing step in step S2 includes the correction of system parameters, specifically correcting the received signal by the system gain constant. The calculation formula of the sonar parameter correction is: I corrected =I received +G sonar (7) Among them, I received is the original signal strength received, G sonar It is the system gain constant, which is used to correct the signal gain value increased manually or automatically by the system.

4. A multi-beam backscatter intensity jump correction method according to claim 1, characterized in that: The process of setting the dynamic threshold in step S3-3 further includes noise smoothing processing and dynamic adjustment of the standard deviation weight factor to adapt to the needs of different sonar working environments. The dynamic threshold calculation formula is: Among them, k is a dynamically adjusted weight factor, which is adjusted according to the actual working environment of the sonar and the characteristics of data changes; N f It is a noise smoothing factor used to remove the influence of random noise in the data. This factor is calculated by the Kalman filter algorithm.

5. A multi-beam backscatter intensity jump correction method according to claim 1, characterized in that: The jump repair in step S5 includes iterative detection and compensation repair of the jump point, which specifically includes the following steps: T1: After completing a jump point compensation repair, repeat steps S3-S5 and record the newly detected jump point; T2: For the newly detected jump points, weighted average calculation and data compensation are performed again to gradually correct all jump points; T3: Iteratively repeat the repair process of step S5 to gradually reduce the repair error until the number of jump points converges to zero.

6. A multi-beam backscatter intensity jump correction method according to claim 1, characterized in that: The processing results of step S1 to step S2 also include preset conditions, which are specifically: The multi-beam intensity is processed in units of a single strip, and the data is corrected for system parameters to remove the influencing factors introduced by human or operating systems, and a complete strip data is selected as the source of correction data; In the process of using Gaussian weighted automatic detection of jumps, the statistical calculation of intensity differences is introduced to set the dynamic threshold based on it. The dynamic threshold changes with the data and can maintain good adaptability in different situations. All the jump points are segmented in sequence, and every two adjacent jump points determine a jump data segment. The intensity values ​​within the window W are taken before and after the jump point, and the weighted average value is calculated as the compensation difference.

Images