Detection method of side-scan sonar submarine line

Abstract

The invention relates to the technical field of sonar, in particular to a side-scan sonar seabed line detection method. The method comprises the following steps: firstly, acquiring real-time echo signal data of the left and right sides of the side-scan sonar, then updating detection parameters of the left and right sides according to the echo signal data, and then respectively detecting the off-bottom heights of the left and right sides; judging whether the left and right terrain clearances are symmetrical or not, and correcting error data; and obtaining the ground clearance of the starboard and the starboard of the current frame, smoothing the result, and outputting a submarine line tracking result, and the smoothing method adopts a median filtering method or a Kalman filtering method or a delay filtering method. According to the method, a seabed line result can be accurately detected in real time instead of post-processing based on a side-scan sonar image, the method can be used for wet end grounding early warning, and a basis is provided for side-scan sonar image real-time slope distance correction, image equalization, target real size measurement and sound image real-time inlaying.

Description

Technical Field

[0001] The invention relates to the technical field of sonar, and in particular to a method for detecting a side-scan sonar seabed line. Background Art

[0002] Side scan sonar is a slant range imaging mode. There is a "water column area" on the sound map. The obvious boundary lines on both sides of the "water column area" are the seabed lines. The seabed lines reflect the height changes of the side scan sonar wet end during movement and the undulating terrain of the seabed. They can be used for wet end bottoming warning, and are also the basis for subsequent slant range correction, image equalization, target geometric feature measurement and sound map mosaic. The existing seabed line detection methods are usually performed during image post-processing, mainly including edge detection based on image edge technology, neural network model detection method and threshold method. Side scan sonar is a type of image sonar and the main equipment for landform measurement. It has the advantages of high resolution, continuous acquisition of two-dimensional seabed images, and low price. It is widely used in marine surveying and mapping, marine geological surveys, marine engineering exploration, underwater search and rescue, underwater archaeology, and marine military. The side scan sonar first emits a short sound pulse, which propagates outward in the form of a spherical wave. When it hits the seabed or the target object in the water, it will be scattered. The backscattered waves are arranged in sequence according to the return time to form a frame of scanning lines. The intensity and time of the reflected scattered waves reflect the location, material, topography and other seabed information of the scattering point. The scanning lines of each frame are arranged vertically one by one to form a seabed topography sound map. The sound map and the seabed plane are mapped point by point, and the brightness of the sound map contains the characteristics of the seabed topography and the target. At present, the detection of the seabed line of the side-scan sonar is mainly carried out during the post-processing of the sonar image. There are mainly three schemes: one is to detect the seabed line based on the edge detection technology of the image. This detection method has high requirements on the quality of the acoustic image, and has problems such as large amount of calculation and inability to detect in real time; the other is the seabed line detection based on the neural network, which realizes the real-time detection of the seabed line through the pre-trained model, but due to the complex seabed conditions and the difficulty in obtaining the measured data of the side-scan sonar, the amount of data used to train the model is small, which reduces the accuracy of the model detection result, and this method also has the problem of large calculation; the last one is the threshold method, which determines the position of the seabed line by comparing the echo signal with the threshold. The threshold is usually set by engineers based on experience in the early days, and semi-automatic threshold setting has also been developed later, but they are all based on post-processing. The threshold method can realize real-time detection of the seabed line, but it is easily affected by wake, waves, suspended matter in the water, fish schools and noise in the water, and has the problem of low accuracy of the detection result.

[0003] In summary, the existing seabed line detection methods have the problems of poor real-time performance and low precision, and there is an urgent need for a side-scan sonar seabed line detection method that can be performed in real time and with high precision. Summary of the invention

[0004] The problem to be solved by the present invention is a real-time high-precision seabed line detection method.

[0005] In view of the shortcomings of the prior art, the present invention solves the technical problems by adopting a technical solution: a method for detecting a seabed line by side-scan sonar, comprising the following steps:

[0006] 1. Obtain real-time echo signal data of the port and starboard sides of the side scan sonar, including the orthogonal demodulated I-channel and Q-channel signals and ADC sampling signals of the port and starboard sides;

[0007] 2. Update the left and right detection parameters according to the working parameters of the ADC sampling signal and the detection results of the previous frame;

[0008] 3. Detect the bottom height of the left and right sides respectively;

[0009] 4. Determine whether the height from the ground on the left and right sides is symmetrical and correct the wrong data;

[0010] 5. Get the left and right heights above the ground in the current frame, smooth the results, and output the seabed line tracking results.

[0011] Preferably, the process of initializing the detection parameters in step 2 is:

[0012] start_point=pluse_width*fs;

[0013] end_point=(work_range-5) / c*fs;

[0014] threshold1=threshold0;

[0015] Where start_point is the starting point of the seabed line search, end_point is the end point of the seabed line search, pluse_width is the pulse width of the side scan sonar signal, fs is the sampling rate, work_range is the working range of the side scan sonar, c is the sound speed, and threshold0 is the initial threshold;

[0016] Detection parameter update:

[0017] start_point=(edge0-d0)*2 / c*fs;

[0018] end_point=(edge0+d0)*2 / c*fs;

[0019]

[0020] Where edge0 is the height above the ground detected in the previous frame, d0 is the maximum fluctuation range of the height above the ground between two adjacent frames, d0 is preset, and S0 is the echo signal of the previous frame of the current detection side (left and right sides).

[0021] Preferably, the specific contents of step 3 are as follows:

[0022] 3.1 Obtain detection parameters, sliding window size N, amplitude threshold threshold1, continuous parameter threshold threshold2, seabed height fluctuation threshold threshold3, left and right parameters are independent;

[0023] 3.2 Search the seabed line point by point from the search starting point start_point until the seabed line height is found. If edge=edge0 is not found, edge0 is the detection result of the side in the previous frame;

[0024] 3.3 The average echo intensity data_ave of the echo signal in the sliding window is compared with the threshold threshold1. The calculation formula for the average echo intensity of the n-th point signal is:

[0025]

[0026] Among them, S(n) is the echo signal of the current detection side (left and right sides), N is the size of the sliding window,

[0027] If data_ave(n)>threshold1, go to step 3.4, otherwise slide the sliding window a little and repeat step 3.3;

[0028] 3.4 Determine the number of points n2 whose echo intensity among the N sampling points in the sliding window is greater than threshold2. The method of obtaining n2 is as follows:

[0029]

[0030] If n2 is greater than N2, go to step 3.5, otherwise the sliding window slides a little and repeats step 3.3. N2 is the threshold of the number of sampling points within the sliding window that is greater than threshold1. N2 is preset in advance.

[0031] 3.5 Determine the difference between the current point n and the last detected seabed line edge0 and threshold3. If the difference between the current point n and the last detected seabed line edge0 is greater than threshold3, it is considered that the seabed line detection result of the side of the frame is edge=n.

[0032] Preferably, in step 4, the left and right sides ground clearance data correction sets the error threshold4, which is set based on the wet end attitude, noise and seabed topography factors, threshold4 = 0.03*(edge_left0+edge_right0), edge_left0 and edge_right0 are the average values ​​of the left seabed line detection results, respectively, which can be updated by sliding average.

[0033] Preferably, the smoothing method in step 5 is a median filter method, a Kalman filter method, or a delay filter method.

[0034] The beneficial effects of the present invention are as follows: the seabed line results can be detected in real time and accurately instead of based on the post-processing of the side-scan sonar image, which can be used for wet-end bottom-touching warning, and provide a basis for real-time slant range correction, image equalization, target true size measurement and real-time mosaicking of acoustic images of side-scan sonar images. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flow chart of the method of the present invention;

[0036] Figure 2 is a process diagram of step 3 in the method of the present invention;

[0037] Figure 3 is a seafloor line detection map without using the method of the present invention;

[0038] Figure 4 The present invention is a seabed line detection diagram using the method of the present invention. DETAILED DESCRIPTION

[0039] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention are provided for the purpose of illustration and description, and are not intended to be exhaustive or to limit the present invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention, and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

[0040] The present invention aims at the problems of poor real-time performance and low precision of existing seabed line detection methods, and proposes a real-time and high-precision seabed line detection method. The seabed line echo signal is the intersection of the water column area and the seabed area, and has the characteristic of mutation. However, the echo signal of the "water column area" may also have mutations due to factors such as suspended matter, fish schools, waves and underwater noise. The echo signal of the water column area quickly falls back after the mutation. The average energy of the sliding window and the number of points greater than the threshold in the sliding window are used to detect the seabed line. And based on the fact that the seabed line will not mutate, abnormal points are eliminated. After the detection of the seabed lines on the left and right sides is completed, the abnormal points are further eliminated according to the principle of left-right symmetry of the seabed line and combined with the posture of the side-scan sonar, and finally the results are smoothed by using methods such as Kalman filtering and median filtering.

[0041] The method of the present invention comprises the following process:

[0042] 1. Obtain the real-time echo signal data of the port and starboard sides of the side scan sonar. The echo signal data is usually the orthogonal demodulated I and Q signals of the port and starboard sides, denoted as I L , Q L ,I R and Q R , or it can be an ADC sampling signal, and synchronously obtain the sonar working frequency, signal form, pulse width, bandwidth and working range and other working parameters; roll, pitch and other attitude data, for the transmission signal, its form is CW signal:

[0043]

[0044] For the transmitted signal, the signal form is LFM signal:

[0045] S R =conv(I R +i*Q R ,S lfm )

[0046] S L =conv(I L +i*Q L ,S lfm )

[0047] Among them, S R is the right echo signal, S L is the left echo signal, i is the imaginary unit, I R is the I signal of the orthogonal demodulation on the starboard side, Q R is the Q-path signal of the orthogonal demodulation on the starboard side, Q L The Q-path signal of the orthogonal demodulation on the port side, I L is the I-channel signal of the orthogonal demodulation on the port side, where S lfm is the transmitted signal, S lfmIt can be generated based on parameters such as pulse width, operating frequency, bandwidth, etc., or it can be obtained through measured data. Conv means to find the convolution of two signals;

[0048] 2. Update the left and right detection parameters, specifically update the detection parameters according to the working parameters of the side scan sonar and the detection results of the previous frame:

[0049] Detection parameter initialization:

[0050] start_point=pluse_width*fs;

[0051] end_point=(work_range-5) / c*fs;

[0052] threshold1=threshold0;

[0053] Where start_point is the starting point of the seabed line search, end_point is the end point of the seabed line search, pluse_width is the pulse width of the side scan sonar signal, fs is the sampling rate, work_range is the working range of the side scan sonar, c is the sound speed, and threshold0 is the initial threshold, which is pre-set. For example, an 8-bit signal can be set to 5-20.

[0054] Detection parameter update:

[0055] start_point=(edge0-d0)*2 / c*fs;

[0056] end_point=(edge0+d0)*2 / c*fs;

[0057]

[0058] Where edge0 is the height above the ground detected in the previous frame, d0 is the maximum fluctuation range of the height above the ground between two adjacent frames, d0 is preset, and S0 is the echo signal of the previous frame of the current detection side (left and right sides);

[0059] 3. Detect the bottom height of the left and right sides respectively. The detection process is as follows: Figure 2 As shown, the specific steps are as follows:

[0060] 3.1 Obtain detection parameters, sliding window size N, amplitude threshold threshold1, continuous parameter threshold threshold2, seabed height fluctuation threshold threshold3, left and right parameters are independent;

[0061] 3.2 Search the seabed line point by point from the search starting point start_point until the seabed line height is found. If edge=edge0 is not found, edge0 is the detection result of the side in the previous frame;

[0062] 3.3 Comparison of the average echo intensity data_ave of the echo signal in the sliding window with the threshold threshold1. In order to avoid the target in the water column area being misidentified as the seabed line in the threshold method, the average echo intensity data_ave of the echo signal in the sliding window is compared with the threshold threshold1. The calculation formula for the average echo intensity of the n-th point signal is:

[0063]

[0064] Among them, S(n) is the echo signal of the current detection side (left and right sides), N is the size of the sliding window,

[0065] If data_ave(n)>threshold1, go to step 3.4, otherwise slide the sliding window a little and repeat step 3.3;

[0066] 3.4 Determine the number of points n2 whose echo intensity is greater than threshold2 among the N sampling points in the sliding window. In order to avoid misjudgment caused by the presence of a target with a large target intensity in the sliding window, further determine the number of points n2 whose echo intensity is greater than threshold2 among the N sampling points in the sliding window. The method for obtaining n2 is as follows:

[0067]

[0068] If n2 is greater than N2, go to step 3.5, otherwise the sliding window slides a little and repeats step 3.3. N2 is the threshold of the number of sampling points within the sliding window that is greater than threshold1. N2 is preset in advance.

[0069] 3.5 Determine the difference between the current point n and the last detected seabed line edge0 and threshold3. The seawater depth and the side scan sonar towing depth are both gradual changes, so the seabed line height is also a gradual change. Determine whether the difference between the current point n and the last detected seabed line edge0 is greater than threshold3. If it is greater than threshold3, it is considered that the seabed line detection result of this frame is edge=n.

[0070] 4. Determine whether the height above the ground on the left and right sides is symmetrical, and correct the erroneous data: The transducers on the left and right sides of the side scan sonar wet end are installed symmetrically and the spacing is small. The echo signals of the seabed line all come from directly below the wet end. The height of the seabed lines on both sides should be equal in theory, but considering the influence of the wet end attitude, noise and seabed topography, a certain error threshold4 is allowed between the two. The error can be taken as 0.03*(edge_left0+edge_right0), edge_left0 and edge_right0 are the average values ​​of the left seabed line detection results, which can be updated by sliding average. The influence of the wet end attitude on the error can also be further considered. If the left and right detection results abs(edge_left-edge_right)>threshold4, edge_left and edge_right are the left and right seabed lines detected in step 3, respectively, then the left and right detection results are compared with the average values ​​of the left and right detection results of the previous frame, and the result on the side with the large deviation is corrected to the average value of the previous frame detection result;

[0071] 5. Get the left and right heights above the ground in the current frame, smooth the results, and output the seabed line tracking results: Smooth the seabed line detection results of the left and right sides respectively, remove abnormal points to improve the accuracy of the detection results. The smoothing method can be median filtering, Kalman filtering, delay filtering and other methods. Taking Kalman filtering as an example, the detection result is updated by the following formula:

[0072] kalman_edge k|k-1 =A*kalman_edge k-1 +w k-1

[0073] P k|k-1 =AP k-1 A T +Q

[0074] K k =P k|k-1 H T (HP k|k-1 H T +R) -1

[0075] kalman_edge k =kalman_edge k|k-1 +K k (edge k -H*kalman_edge k|k-1 )

[0076] P k =(IK k H)P k|k-1

[0077] Where kalman_edge k|k-1 is the seabed state of the current frame predicted based on the previous state, kalman_edge k is the result of Kalman filtering of the kth frame, A is the transfer matrix, P k|k-1 is the system covariance of the predicted k-th frame, P k is the system covariance of the kth frame, the matrix H is the prediction matrix of the object, and w k-1 is the state noise, K k is the Kalman gain, R is the covariance matrix of the measurement noise, Q is the covariance of the system process noise, I is the identity matrix, edge k is the result of the seafloor detected in step 4, and T represents the matrix transpose.

[0078] Figure 3 The threshold method is used and the results are smoothed. Due to the existence of targets in the water column area, there are errors in the detection of the seabed line. Figure 4 The detection method provided by the present invention has the advantages of small computational complexity and high detection accuracy. It can detect the seabed line in real time and accurately, can be used for wet end bottom contact warning, and provides a basis for real-time slant range correction, image equalization, target true size measurement and real-time mosaicking of acoustic images of side scan sonar images.

Claims

1. A method for detecting a seabed line using a side-scan sonar, characterized in that: The following steps are included:

1. Obtain real-time echo signal data of the port and starboard sides of the side scan sonar, including the orthogonal demodulated I-channel and Q-channel signals and ADC sampling signals of the port and starboard sides; 2. Update the left and right detection parameters according to the working parameters of the ADC sampling signal and the detection results of the previous frame; 3. Detect the bottom height of the left and right sides respectively; 4. Determine whether the height from the ground on the left and right sides is symmetrical and correct the wrong data; 5. Get the left and right heights above the ground in the current frame, smooth the results, and output the seabed line tracking results.

2. The method for detecting a seabed line by side-scan sonar according to claim 1, characterized in that: The process of initializing the detection parameters in step 2 is: start_point=pluse_width*fs; end_point=(work_range-5) / c*fs; threshold1=threshold0; Where start_point is the starting point of the seabed line search, end_point is the end point of the seabed line search, pluse_width is the pulse width of the side scan sonar signal, fs is the sampling rate, work_range is the working range of the side scan sonar, c is the sound speed, and threshold0 is the initial threshold; Detection parameter update: start_point=(edge0-d0)*2 / c*fs; end_point=(edge0+d0)*2 / c*fs; Where edge0 is the height above the ground detected in the previous frame, d0 is the maximum fluctuation range of the height above the ground between two adjacent frames, d0 is preset, and S0 is the echo signal of the previous frame of the current detection side (left and right sides).

3. The method for detecting the seabed line by side-scan sonar according to claim 2, characterized in that: The specific contents of step 3 are as follows: 3.1 Obtain detection parameters, sliding window size N, amplitude threshold threshold1, continuous parameter threshold threshold2, seabed height fluctuation threshold threshold3, left and right parameters are independent; 3.2 Search the seabed line point by point from the search starting point start_point until the seabed line height is found. If edge=edge0 is not found, edge0 is the detection result of the side in the previous frame; 3.3 The average echo intensity data_ave of the echo signal in the sliding window is compared with the threshold threshold1. The calculation formula for the average echo intensity of the n-th point signal is: Among them, S(n) is the echo signal of the current detection side (left and right sides), N is the size of the sliding window, If data_ave(n)>threshold1, go to step 3.4, otherwise slide the sliding window a little and repeat step 3.3; 3.4 Determine the number of points n2 whose echo intensity among the N sampling points in the sliding window is greater than threshold2. The method of obtaining n2 is as follows: If n2 is greater than N2, go to step 3.5, otherwise the sliding window slides a little and repeats step 3.

3. N2 is the threshold of the number of sampling points within the sliding window that is greater than threshold1. N2 is preset in advance. 3.5 Determine the difference between the current point n and the last detected seabed line edge0 and threshold3. If the difference between the current point n and the last detected seabed line edge0 is greater than threshold3, it is considered that the seabed line detection result of the side of the frame is edge=n.

4. The method for detecting the seabed line by side-scan sonar according to claim 3, characterized in that: In step 4, the left and right sides ground clearance data correction setting error threshold4, threshold4 is set based on the wet end attitude, noise and seabed topography factors, threshold4 = 0.03 * (edge_left0 + edge_right0), edge_left0, edge_right0 are the average values ​​of the left seabed line detection results, which can be updated by sliding average.

5. The method for detecting seabed lines by side-scan sonar according to claim 1, characterized in that: The smoothing method in step 5 is a median filter method, a Kalman filter method, or a delay filter method.

Images