An airborne marine lidar detection method and system for weak underwater echo signals

Abstract

The present invention belongs to the technical field of marine surveying and mapping, and discloses a method and system for detecting weak underwater echo signals of an airborne marine lidar. This method constructs a standard deviation curve to determine the rising edge of the waveform, accurately extracts the water surface echo time, and calculates two water surface geometric elements, namely the water surface point and the refraction vector, on this basis; constructs a waveform superposition model with spatio-temporal constraints, takes the Euclidean distance between the two water surface geometric elements as the nearest neighbor discrimination criterion, and uses the K-nearest neighbor judgment method to screen and superpose the nearest neighbor waveforms to obtain a superposed pseudo-waveform with high signal-to-noise ratio; finally, uses an improved normalized significant peak detection method to gradually judge the underwater echo position and obtain the final underwater echo time. After being processed by the model constructed by the present invention, the weak underwater echo detection of the airborne marine lidar is effectively realized, the maximum sounding performance of the airborne marine lidar is improved, and an effective technical support and solution are provided for sounding data processing and application.

Description

Technical Field

[0001] The present invention belongs to the technical field of marine surveying and mapping, and particularly relates to a method and system for detecting weak underwater echo signals of an airborne marine lidar. Background Technique

[0002] The airborne marine lidar sounding technology has the characteristics of high measurement accuracy, large measuring point density, strong mobility and amphibious use. It is particularly suitable for the rapid detection of complex terrains such as shallow waters and surrounding waters in coastal zones and reefs, and can obtain integrated underwater and above-water topographic and geomorphic data, providing a data basis and important reference for aspects such as marine resource development, marine ecological protection, coastal zone surveys and water resource management. The airborne marine lidar sounding system can emit green laser with a wavelength of 532 nm that can penetrate water, and receive waveform data information including surface, water body and underwater echo. By extracting each echo component, the travel time of the laser can be obtained. Combining with positioning and attitude data, the three-dimensional coordinates of underwater terrain points can be calculated. Among them, the detection of underwater echo is the basis for obtaining accurate laser travel time. The detection ability of weak underwater echo directly determines the maximum sounding performance of the system and is the key content of airborne marine lidar sounding data processing.

[0003] The existing methods for detecting weak underwater echoes of airborne marine lidar mainly include two categories: one is the detection method based on a single waveform; the other is the detection method that combines multiple neighboring waveforms. Among them, the former extracts the position of the underwater weak echo by using time-domain, frequency-domain or time-frequency joint analysis methods according to the characteristic differences between the target echo and the noise signal in a single waveform, but this type of method does not fully consider the correlation between spatially neighboring waveforms; the latter extracts the weak underwater echo by using methods such as waveform superposition on the basis of considering the correlation of neighboring waveforms, but the discrimination criteria for neighboring waveforms mostly adopt a grid form, which limits the flexibility of this method. In view of the above problems, the present invention proposes a more general method for detecting weak underwater echo signals of airborne marine lidar to improve the maximum sounding performance of airborne marine lidar. Summary of the Invention

[0004] To overcome the problems existing in the related technologies, the disclosed embodiments of the present invention provide a method and system for detecting weak underwater echo signals of an airborne marine lidar. The purpose is to solve the problems of insufficient utilization of the neighboring relationship between waveforms, and the need to improve the detection flexibility and generality in the existing detection technologies.

[0005] The technical solutions are as follows: A method for detecting weak underwater echo signals of an airborne marine lidar, including the following steps:

[0006] S1, determining the position of the surface echo based on the waveform standard deviation curve, and obtaining two geometric elements, namely the surface point and the refraction vector;

[0007] S2. Construct a waveform superposition model with spatio-temporal constraints to obtain a superposition pseudo-waveform with high signal-to-noise ratio;

[0008] S3. Based on an improved normalized significant peak detection method, gradually judge and obtain the position of the weak bottom echo.

[0009] In step S1, determine the water surface echo position based on the waveform standard deviation curve, and obtain two geometric elements of the water surface point and the refraction vector, including:

[0010] S101. According to the standard deviation difference between the noise and the rising edge of the echo in the sounding waveform, use a sliding window to construct a waveform standard deviation curve, and determine the rising edge position as the water surface echo position;

[0011] S102. According to the determined water surface echo position and the positioning and attitude information, calculate the water surface point cloud , and based on the principal component analysis method, obtain the refraction vector corresponding to each water surface point .

[0012] In step S101, determine the rising edge position as the water surface echo position, and the expression is:

[0013] (1)

[0014] In the formula, is the water surface echo position, is the position to find the data that meets the conditions; is the standard deviation curve, , is the original echo signal at position in the standard deviation value, is the sliding window size, is the standard deviation function; is the extreme point position of the standard deviation curve , is the noise threshold; is the differential operation, is the sign function;

[0015] In step S102, obtain the refraction vector corresponding to each water surface point , and the expression is:

[0016] (2)

[0017] In the formula, are the refraction vector and the incident vector respectively, respectively represent the refractive indices of laser in air and water, respectively represent the incident angle and the refraction angle, is the water surface point Psi and the normal vector at this point.

[0018] In step S2, a waveform superposition model with spatio-temporal constraints is constructed to obtain a superposed pseudo-waveform with high signal-to-noise ratio, including:

[0019] S201, traverse each water surface point and the corresponding refraction vector , construct a waveform superposition model with spatio-temporal constraints, including two parts: near-neighbor waveform selection and pseudo-waveform acquisition; and calculate the Euclidean distance between all water surface points and refraction vectors to obtain the water surface point distance vector set and the refraction distance vector set ;

[0020] Among them, , is the total number of sea surface points, is the two-norm operation, is the and the Euclidean distance between the and the sea surface points, is the Euclidean distance between the and the refraction vectors, is the coordinates of the

[0021] S202, according to the near-neighbor criterion, obtain the near-neighbor water surface points and the corresponding refraction vectors ; from the selected near-neighbor refraction vectors , according to the near-neighbor criterion again, obtain the near-neighbor refraction vectors and the corresponding water surface points; finally determine the near-neighbor waveforms;

[0022] S203, superimpose the near-neighbor waveform data, filter out Gaussian white noise, and obtain a superimposed pseudo-waveform with a higher signal-to-noise ratio.

[0023] In step S202, obtain the near-neighbor water surface points and the corresponding refraction vectors , and the expression is:

[0024] (3)

[0025] In the formula, are the nearest neighbor water surface points and refraction vectors obtained in the first screening, is the index of the nearest neighbor water surface points obtained in the first screening, is the sorting index of the array elements from small to large, is to obtain values from the sea surface point set, is from the refraction vector set R to obtain values, is the interval vector from 0 to ; Again, according to the nearest neighbor criterion to obtain the refraction vectors corresponding to each water surface point and the nearest neighbor refraction vectors and the corresponding water surface points, the expression is:

[0026] (4)

[0027] In the formula, are the nearest neighbor water surface points and refraction vectors obtained in the second screening, is the index of the nearest neighbor water surface points obtained in the second screening, is to obtain values from the set to obtain values, is to obtain values from the set to obtain values, is the refraction vector R i subtracted from all vectors in the set correspondingly, is the interval vector from 0 to ;

[0028] Finally, the determined nearest neighbor waveform set , where, are the waveform data of the 1st to the th nearest neighbors respectively.

[0029] In step S203, obtain the superimposed pseudo waveform with higher signal-to-noise ratio, and the expression is:

[0030] (5)

[0031] In the formula, is the superimposed pseudo waveform, is the water surface echo position of the th nearest neighbor waveform, is the length of the waveform taken, is the waveform data interval taken, is the truncation vector a certain paragraph of is the finally determined number of neighboring waveforms.

[0032] In step S3, based on the improved normalized significant peak detection method, the position of the weak bottom echo is judged and obtained step by step, including:

[0033] S301, obtain the amplitudes , separation degrees and prominences of all peaks in the superimposed pseudo waveform, and normalize the three indexes respectively to calculate the significance of each peak;

[0034] S302, in the superimposed pseudo waveform , start searching backward from moment. If the search value is greater than , continue to search backward, otherwise stop; according to the moment position when stopping, obtain the search interval of the weak bottom echo;

[0035] S303, within the echo search interval, calculate the significance values of the peak positions in the waveforms corresponding to each water surface point , and determine the peak position with the maximum significance as the final bottom echo position .

[0036] In step S301, the significance of each peak is calculated, and the expression is:

[0037] (6)

[0038] In the formula, is the significance value; respectively represent the normalized amplitude, separation degree and prominence values; the peak position with a prominence greater than the noise threshold and the maximum significance value is used as the approximate moment of the bottom echo;

[0039] The noise threshold is expressed as:

[0040] (7)

[0041] In the formula, is the end noise in the superimposed pseudo waveform; is to obtain the values at the end of the superimposed pseudo waveform; is the standard deviation function; is the mean function.

[0042] In step S302, the search interval of the weak bottom echo is , is the starting time position of the interval, is the time position at stop.

[0043] Another object of the present invention is to provide a method for detecting weak underwater echo signals of an airborne marine lidar. The system implements the method for detecting weak underwater echo signals of the airborne marine lidar. The system includes:

[0044] A geometric element acquisition module, configured to determine the water surface echo position based on the waveform standard deviation curve, and acquire two geometric elements of the water surface point and the refraction vector;

[0045] A superimposed pseudo-waveform acquisition module, configured to construct a waveform superimposition model with spatio-temporal constraints, and acquire a superimposed pseudo-waveform with high signal-to-noise ratio;

[0046] A weak underwater echo position acquisition module, configured to gradually judge and acquire the weak underwater echo position based on an improved normalized significant peak detection method.

[0047] Combining all the above technical solutions, the beneficial effects of the present invention are as follows: The present invention constructs a standard deviation curve to determine the waveform rising edge, accurately extracts the water surface echo time, and calculates two water surface geometric elements of the water surface point and the refraction vector on this basis; constructs a waveform superimposition model with spatio-temporal constraints, uses the Euclidean distance between the two water surface geometric elements as the nearest neighbor discrimination criterion, and uses the K-nearest neighbor judgment method to screen and superimpose the nearest neighbor waveforms to obtain a superimposed pseudo-waveform with high signal-to-noise ratio; finally, uses an improved normalized significant peak detection method to gradually judge the underwater echo position and obtain the final underwater echo time. Through the method of the present invention, the detection of weak underwater echoes of the airborne marine lidar is effectively realized, the maximum sounding performance of the airborne marine lidar is improved, and an effective technical support and solution are provided for sounding data processing and application.

[0048] The present invention aims to innovatively propose the acquisition of spatio-temporal elements, the construction of a waveform superimposition model with spatio-temporal constraints, and the gradual detection of submarine echoes by using the K-nearest neighbor judgment method to screen and superimpose the nearest neighbor waveforms to obtain a superimposed pseudo-waveform with high signal-to-noise ratio. The present invention uses an improved normalized significant peak detection method to define the significance score jointly according to the prominence, separation degree, and amplitude of the signal.

[0049] The present invention can effectively detect the weak bottom echo signal of an airborne marine lidar, improve the maximum sounding performance of the airborne marine lidar, and provide accurate and reliable topographic data for shallow waters in the fields of ocean, surveying and mapping, transportation, and military. The present invention proposes a waveform superposition model based on spatio-temporal constraints to achieve accurate detection of weak bottom echoes. Since the present invention makes full use of the spatio-temporal neighborhood relationship of waveforms, compared with the existing technologies, the present invention has stronger applicability and higher detection accuracy. The method of spatio-temporal constraint waveform superposition in the present invention eliminates noise interference and improves the recognizability of bottom echo signals. Description of the Drawings

[0050] The drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure;

[0051] Figure 1 It is a schematic diagram of the principle of the method for detecting weak bottom echo signals of an airborne marine lidar provided by an embodiment of the present invention;

[0052] Figure 2 It is a waveform standard deviation curve diagram constructed by using a sliding window in the determination of the water surface echo position based on the waveform standard deviation curve provided by an embodiment of the present invention;

[0053] Figure 3 It is a water surface echo position diagram determined in the determination of the water surface echo position based on the waveform standard deviation curve provided by an embodiment of the present invention;

[0054] Figure 4 It is a schematic diagram of the selection of neighboring waveforms provided by an embodiment of the present invention;

[0055] Figure 5 It is a schematic diagram of the construction of a pseudo waveform provided by an embodiment of the present invention;

[0056] Figure 6 It is the amplitude , separation and prominence of all peaks of the waveform obtained in the determination of weak bottom echoes based on normalized significant peak detection provided by an embodiment of the present invention;

[0057] Figure 7 It is a schematic diagram in the determination of weak bottom echoes based on normalized significant peak detection provided by an embodiment of the present invention, where the peak position with the prominence of the superimposed pseudo waveform greater than the noise threshold and the maximum significance value is used as the approximate time of the bottom echo ;

[0058] Figure 8It is a schematic diagram for obtaining the search range of weak underwater echoes in the determination of weak underwater echoes based on normalized significant peak detection provided by an embodiment of the present invention. Schematic diagram;

[0059] Figure 9 It is a schematic diagram for determining the position of the final underwater echo by determining the peak position with the maximum significance in the determination of weak underwater echoes based on normalized significant peak detection provided by an embodiment of the present invention. Schematic diagram;

[0060] Figure 10 It is a schematic diagram of underwater terrain points calculated by the method of the present invention;

[0061] Figure 11 It is a map of underwater terrain points calculated by the existing SPD algorithm;

[0062] Figure 12 For the present invention Figure 10 Schematic diagram of the first bathymetric profile corresponding to the solid line in the present invention;

[0063] Figure 13 For the present invention Figure 11 Schematic diagram of the second bathymetric profile corresponding to the solid line in the present invention;

[0064] Figure 14 For the present invention Figure 10 Schematic diagram of the first bathymetric profile corresponding to the dashed line in the present invention;

[0065] Figure 15 For the present invention Figure 11 Schematic diagram of the second bathymetric profile corresponding to the dashed line in the present invention. Detailed implementation manners

[0066] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following describes the detailed implementation manners of the present invention with reference to the accompanying drawings. Many specific details are set forth in the following description to fully understand the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific implementations disclosed below.

[0067] To solve the problem that the signal-to-noise ratio of the weak underwater echo signal of the airborne ocean lidar is low and it is difficult to extract, the present invention proposes a method for detecting weak underwater echoes based on a spatio-temporal constraint waveform superposition model. First, by constructing a standard deviation curve to locate the rising edge of the waveform, the sea surface echo time is determined, and two elements, namely the sea surface point and the corresponding refraction vector, are calculated therefrom; then, the Euclidean distance between the two elements is used as the K-nearest neighbor judgment criterion to construct a spatio-temporal constraint waveform superposition model to obtain a superposition pseudo-waveform with a higher signal-to-noise ratio; finally, a normalized significant peak detection method is used to determine the final position of the weak underwater echo.

[0068] The research object of the present invention is the weak echo from the seabed detected by airborne marine laser sounding. And the present invention improves the waveform superposition according to the waveform characteristics and proposes a waveform superposition model with spatio-temporal constraints.

[0069] Example 1, as Figure 1 shown, the method for detecting the weak echo signal from the seabed by the airborne marine lidar provided by the embodiment of the present invention includes:

[0070] S1, determining the water surface echo position based on the waveform standard deviation curve and obtaining two geometric elements, namely the water surface point and the refraction vector;

[0071] S2, constructing a waveform superposition model with spatio-temporal constraints and obtaining a superposed pseudo-waveform with high signal-to-noise ratio;

[0072] S3, based on the improved normalized significant peak detection method, gradually judging and obtaining the position of the weak echo from the seabed.

[0073] Exemplarily, in step S1, determining the water surface echo position based on the waveform standard deviation curve and obtaining two geometric elements, namely the water surface point and the refraction vector, includes:

[0074] S101, according to the standard deviation difference between the noise and the rising edge of the echo in the sounding waveform, using a sliding window to construct a waveform standard deviation curve (such as Figure 2 ), and determining the exact rising edge position as the water surface echo position ( Figure 3 ), the present invention innovatively proposes as shown in formula (1):

[0075] (1)

[0076] In the formula, is the water surface echo position, is the data position for finding the condition; is the standard deviation curve, , is the original echo signal at the position of the standard deviation value, is the sliding window size, is the standard deviation function; is the extreme point position of the standard deviation curve , is the noise threshold; is the difference operation, is the sign function;

[0077] S102, according to the determined water surface echo position and the position and attitude information, reducing the water surface point cloud , and based on the principal component analysis method, obtaining each water surface point The corresponding refraction vector .

[0078] Obtain each water surface point The corresponding refraction vector , and the expression is:

[0079] (2)

[0080] In the formula, are the refraction vector and the incident vector respectively, represent the refractive indices of laser in air and water respectively, represent the incident angle and the refraction angle respectively, is the water surface point Psi The normal vector at the position.

[0081] Exemplarily, in step S2, constructing a waveform superposition model with spatio-temporal constraints to obtain a superposed pseudo-waveform with high signal-to-noise ratio includes:

[0082] S201, traverse each water surface point and its corresponding refraction vector , and calculate the Euclidean distance between other water surface points and refraction vectors to obtain a set of water surface point distance vectors and a set of refraction distance vectors ; among them, , represents the two-norm operation, , is the two-norm operation, , is the total number of sea surface points, is the two-norm operation, is the th Euclidean distance between the th and the th sea surface points, is the th sea surface point coordinate, is the th refraction vector corresponding to the sea surface point.

[0083] On this basis, a waveform superposition model with spatio-temporal constraints is constructed. The essence of this model is to improve the signal-to-noise ratio level of the waveform by considering the correlation of neighboring waveforms to achieve the detection of weak bottom echo signals. This model mainly includes S202 neighboring waveform selection (such as Figure 4 ) and S203 pseudo-waveform acquisition (such as Figure 5 ) in two parts.

[0084] S202. First, according to the K-nearest neighbor criterion, obtain the nearest neighboring water surface points and the corresponding refraction vectors . The present invention innovatively proposes as shown in Equation (3):

[0085] (3)

[0086] In the formula, are respectively the initially screened neighboring water surface points and refraction vectors, is the index of the initially screened neighboring water surface points, is the sorting index of the array elements from small to large, takes values from the sea surface point set, is from the refraction vector set R to take values, is the interval vector from 0 to ;

[0087] From the screened neighboring refraction vectors , again according to the nearest neighbor criterion, obtain the corresponding refraction vectors nearest neighboring refraction vectors and the corresponding water surface points. The present invention innovatively proposes as shown in Equation (4):

[0088] (4)

[0089] In the formula, are respectively the re-screened neighboring water surface points and refraction vectors, is the index of the re-screened neighboring water surface points, takes values from the set , takes values from the set , is the refraction vector R i subtracted from all vectors in the set correspondingly, is the interval vector from 0 to ;

[0090] Through the nearest neighbor constraints of the above two elements, the finally determined nearest neighbor waveform . Among them, are respectively the waveform data of the 1st to the nearest neighbors.

[0091] S203. From the waveform pair Superimpose the adjacent waveform data to filter out Gaussian white noise and obtain a superimposed pseudo-waveform with a higher signal-to-noise ratio. The present invention innovatively proposes the following as shown in Equation (5):

[0092] (5)

[0093] In the formula, is the superimposed pseudo-waveform, is the water surface echo position of the th adjacent waveform, is the length of the waveform taken, is the waveform data interval taken, is a certain segment of the intercept vector , is the finally determined number of adjacent waveforms.

[0094] Exemplarily, in step S3, based on the improved normalized significant peak detection method, gradually judging and obtaining the weak bottom echo position includes:

[0095] S301, obtain the amplitudes , separation degrees and prominences of all peaks in the superimposed pseudo-waveform (such as Figure 6 ), and normalize the three indicators respectively to calculate the significance of each peak. The present invention innovatively proposes the following formula (6):

[0096] (6)

[0097] In the formula, is the significance value; respectively represent the normalized amplitude, separation degree and prominence values; the peak position with a prominence greater than the noise threshold and the largest significance value is used as the approximate time of the bottom echo (such as Figure 7 ); among them, the noise threshold is expressed as:

[0098] (7)

[0099] In the formula, is the end noise in the superimposed pseudo-waveform; is to obtain the values at the end of the superimposed pseudo-waveform; is the standard deviation function; is the mean function.

[0100] S302, in the superimposed pseudo-waveform , starting from the moment and searching backward, if the search value is greater than Then continue to search backward, otherwise stop. According to the time position at the stop , the present invention innovatively obtains the search range of weak underwater echoes , where is the starting time position of the range, is the time position at the stop (such as Figure 8 ).

[0101] S303. Within the echo search range, calculate the significance values of the peak positions in the waveforms corresponding to each water surface point , and determine the peak position with the maximum significance as the final underwater echo position , as shown in Figure 9 .

[0102] As can be seen from the above embodiments, the present invention proposes a method for detecting weak underwater echo signals of an airborne marine lidar. Compared with the prior art, it determines the water surface echo position based on the waveform standard deviation curve, and obtains two geometric elements: the water surface point and the refraction vector; constructs a waveform superposition model with spatio-temporal constraints, and respectively calculates the Euclidean distance between the two geometric elements. According to the K-nearest neighbor judgment method, screen the neighboring waveforms and perform superposition to obtain a superposed pseudo-waveform with high signal-to-noise ratio; based on an improved normalized significance peak detection method, gradually judge and determine the final weak underwater echo position. After processing with the model constructed by the present invention, the weak underwater echo detection of the airborne marine lidar is effectively realized, the maximum sounding performance of the airborne marine lidar is improved, and effective technical support and solutions are provided for sounding data processing and applications.

[0103] Another object of the present invention is to provide a system for detecting weak underwater echo signals of an airborne marine lidar, the system comprising:

[0104] A geometric element acquisition module, configured to determine the water surface echo position based on the waveform standard deviation curve, and obtain two geometric elements: the water surface point and the refraction vector;

[0105] A superposed pseudo-waveform acquisition module, configured to construct a waveform superposition model with spatio-temporal constraints, and obtain a superposed pseudo-waveform with high signal-to-noise ratio;

[0106] A weak underwater echo position acquisition module, configured to gradually judge and obtain the weak underwater echo position based on an improved normalized significance peak detection method.

[0107] To verify the effectiveness of the proposed algorithm, the present invention conducts experiments and analyzes based on the sounding data of an airborne marine lidar measured in a certain sea area of a certain region. And the traditional significance-based peak detection (SPD) algorithm is selected as the comparison method to analyze the improvement of the maximum sounding performance of the method of the present invention.Figure 10 and Figure 11 respectively show the underwater terrain points calculated by the method of the present invention and the SPD algorithm. It can be intuitively seen that the sounding points calculated by the algorithm of the present invention extend deeper into the water depth, and can better restore the underwater characteristic landforms in deeper areas. In addition, to further analyze the performance of the method, two sounding profiles parallel to the flight strip direction, sounding profile 1 and sounding profile 2 in the direction of the flight strip, corresponding to Figure 10 the dotted line and the solid line in Figure 10 are taken. Under different processing methods, the sounding profile 1 and sounding profile 2 corresponding to the solid line in Figure 12 and Figure 13 respectively. It can be seen that the terrain of this profile has large fluctuations, which is caused by underwater reefs. And the underwater terrain trend obtained by the method of the present invention is continuous. Relative to the average water surface, the maximum sounding depth is 6.93m. While the underwater terrain obtained by the SPD algorithm is discontinuous, and terrain discontinuities appear around the reefs and in areas with large water depths. In addition, Figure 10 the sounding profile 1 and sounding profile 2 corresponding to the dotted line in Figure 14 and Figure 15 respectively. The underwater terrain of this profile is relatively gentle. The maximum sounding depth obtained by the method of the present invention is 8.67m, and the maximum sounding depth obtained by the SPD algorithm is 5.68m. Therefore, the present invention can effectively realize the detection of weak underwater echoes and significantly improve the maximum sounding depth ability of airborne marine lidar.

[0108] As mentioned above, only the relatively preferred specific embodiments of the present invention are described, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for detecting weak underwater echo signals by airborne ocean laser radar, characterized in that: The method comprises the following steps: S1, determine the water surface echo position based on the waveform standard deviation curve, and obtain two geometric elements: water surface point and refraction vector; S2, constructing a waveform superposition model with time and space constraints to obtain a superposition pseudo waveform with a high signal-to-noise ratio; S3, based on the improved normalized significant peak detection method, the position of weak bottom echo is judged and obtained step by step; In step S2, a waveform superposition model with time and space constraints is constructed to obtain a superposition pseudo waveform with a high signal-to-noise ratio, including: S201, traverse each water surface point P si The corresponding refraction vector R i , construct a waveform superposition model with time and space constraints, including neighbor waveform selection and pseudo waveform acquisition; and calculate the Euclidean distance between all water surface points and refraction vectors to obtain the water surface point distance vector set ΔP si = {ΔP si1 ,ΔP si2 …ΔP sim } and the refraction distance vector set ΔR i = {ΔR i1 , ΔR i2 …ΔR sm } Where ΔP sij =||P si -P sj || 2 , ΔR ij =||R i -R j ||2, j = 1, 2, ... m, m is the total number of sea surface points, || ||2 is the two-norm operation, ΔP sij is the Euclidean distance between the i-th and j-th sea surface points, ΔR ij is the Euclidean distance between the i-th and j-th refraction vectors, P sj is the coordinate of the jth sea surface point, R j is the refraction vector corresponding to the jth sea surface point; S202, according to the K nearest neighbor criterion, obtain the si The K1 nearest water surface points and the corresponding refraction vectors The refraction vector from the selected K1 nearest neighbors In the process, we use the K nearest neighbor criterion to obtain the si The corresponding refraction vector R i The nearest K2 refraction vectors and the corresponding water surface points; the final nearest waveform; S203, superimposing K2 neighboring waveform data from the waveform, filtering out Gaussian white noise, and obtaining a superimposed pseudo waveform with a higher signal-to-noise ratio.

2. The method for detecting weak underwater echo signals of an airborne ocean laser radar according to claim 1, characterized in that: In step S1, the water surface echo position is determined based on the waveform standard deviation curve, and two geometric elements, the water surface point and the refraction vector, are obtained, including: S101, constructing a waveform standard deviation curve using a sliding window according to the standard deviation difference between the noise in the sounding waveform and the rising edge of the echo, and determining the rising edge position as the water surface echo position; S102, based on the determined water surface echo position and positioning information, calculate the water surface point cloud P s , and based on the principal component analysis method, obtain each water surface point P si The corresponding refraction vector R i .

3. The method for detecting weak underwater echo signals of an airborne ocean laser radar according to claim 2, characterized in that: In step S101, the rising edge position is determined as the water surface echo position, and the expression is: Where, t s is the water surface echo position, find() is to find the data position that meets the conditions; S is the standard deviation curve, S=[s1,s2…s m ],s i =std(ω[i:i+n]) is the standard deviation value at position i in the original echo signal ω, n is the sliding window size, and std() is the standard deviation function; T e is the extreme point position of the standard deviation curve S, ε s is the noise threshold; diff() is the difference operation, sign() is the sign function; In step S102, each water surface point P is obtained si The corresponding refraction vector R i , the expression is: In the formula, R i ,I i are the refraction vector and the incident vector, respectively, n a ,n w Represent the refractive index of laser in air and water, θ i ,θ r Respectively represent the incident angle and the refraction angle, N i is the normal vector at the water surface point Psi.

4. The method for detecting weak underwater echo signals of an airborne ocean laser radar according to claim 1, characterized in that: In step S202, obtain the si The K1 nearest water surface points and the corresponding refraction vectors The expression is: In the formula, are the nearest water surface points and refraction vectors screened for the first time, IDX1 is the index of the nearest water surface points screened for the first time, argsort() is the sorting index of the array elements from small to large, P s [ ] is a value taken from the sea surface point set, R[ ] is a value taken from the refraction vector set R, and [:K1] is an interval vector from 0 to K1; Again, according to the K nearest neighbor criterion, obtain the si The corresponding refraction vector R i The expression of the K2 nearest refraction vectors and the corresponding water surface points is: In the formula, are the neighboring water surface points and refraction vectors that are screened again, IDX2 is the index of the neighboring water surface points that are screened again, For the collection Take the value in For the collection Take the value in is the refraction vector R i With Collection Subtract all corresponding vectors in [:K2] from 0 to K2; Finally, the determined neighbor waveform set in, They are the waveform data of the 1st to K2th nearest neighbors respectively.

5. The method for detecting weak underwater echo signals of an airborne ocean laser radar according to claim 1, characterized in that: In step S203, a superimposed pseudo waveform with a higher signal-to-noise ratio is obtained, and the expression is: In the formula, ω acc is the superimposed pseudo waveform, t si is the water surface echo position of the i-th neighboring waveform, l is the length of the waveform taken, [t si :t si +l] is an interval, ω i [] is the intercept vector ω i K2 is the number of neighboring waveforms finally determined.

6. The method for detecting weak underwater echo signals of an airborne ocean laser radar according to claim 1, characterized in that: In step S3, based on the improved normalized significant peak detection method, the position of the weak bottom echo is judged and obtained step by step, including: S301, obtaining the amplitude A, separation I and prominence P of all peaks in the superimposed pseudo waveform, and normalizing the three indicators respectively to calculate the significance of each peak; S302, superimposing the pseudo waveform ω acc In the equation, t′ b Start searching backwards from time. If the search value is greater than If the search fails, the search continues; otherwise, the search stops; according to the position at the time of stopping t t , obtain the weak bottom echo search range; S303, calculate each water surface point P in the echo search interval si The significance value of each peak position in the corresponding waveform is used to determine the peak position with the greatest significance as the final bottom echo position t b .

7. The method for detecting weak underwater echo signals of an airborne ocean laser radar according to claim 6, characterized in that: In step S301, the significance of each peak is calculated, and the expression is: Sig=A'×I'×P'(6) Where Sig is the significance value; A′, I′, and P′ represent the normalized amplitude, separation, and prominence values, respectively; the prominence value greater than the noise threshold ε s The peak position with the largest significance value is taken as the approximate time t′ of the bottom echo b ; Noise threshold ε s It is expressed as: In the formula, is the tail noise in the superimposed pseudo waveform; ω acc [-s:] is to obtain s values ​​of the tail end of the superimposed pseudo waveform; std() is the standard deviation function; mean() is the mean value function.

8. The method for detecting weak underwater echo signals of an airborne ocean laser radar according to claim 6, characterized in that: In step S302, the weak echo search interval is: [2×t′ b -t t , t t ],2×t′ b -t t is the starting time of the interval, t t is the position at the time of stopping.

9. An airborne ocean laser radar underwater weak echo signal detection system, characterized in that: The system implements the method for detecting weak underwater echo signals of an airborne ocean laser radar as claimed in any one of claims 1 to 8, and the system comprises: A geometric element acquisition module is used to determine the water surface echo position based on the waveform standard deviation curve and obtain two geometric elements: the water surface point and the refraction vector; A superposition pseudo waveform acquisition module is used to construct a waveform superposition model with time and space constraints to obtain a superposition pseudo waveform with a high signal-to-noise ratio; The module for acquiring the position of weak bottom echoes is used to judge and acquire the position of weak bottom echoes step by step based on an improved normalized significant peak detection method.

Images