A seabed position detection and tracking method based on sonar image features

Abstract

The present invention discloses a method for detecting and tracking the seabed position based on sonar image features, comprising the following steps: S1: automatic detection of the initial bottom position interval: at startup, the position of the bottom area is estimated by analyzing and processing in a multi-cycle accumulation manner. This step can automatically detect the initial bottom position interval and provide an accurate starting point for subsequent bottom detection; S2: at the initial position, the sonar remains stationary and accumulates multiple Ping data, recorded as S i , i = 1, 2, ..., M, M is the number of accumulated data; S3: Process the above data, using the data of the bottom-sounding sensor, or using the acoustic data for calculation, assuming that the acquired seabed position is I i , i=1,2,...,M, through the steps of automatic detection of initial bottom position interval, bottom detection, bottom detection data quality assessment and seabed position tracking and prediction, it can effectively solve difficult problems such as acoustic image noise, bottom echo difference and interference of suspended targets in the water.

Description

Technical Field

[0001] The present invention relates to the technical field of seabed position detection based on sonar image features, and in particular to a seabed position detection and tracking method based on sonar image features. Background Art

[0002] At present, accurate detection and tracking of seabed positions are of great significance for marine resource development, submarine pipeline laying, etc. However, due to noise interference in acoustic images, differences in echo intensity of different bottom materials, and interference from suspended targets in the water, traditional seabed position detection and tracking methods have certain difficulties.

[0003] The main problems of traditional seabed position detection and tracking include:

[0004] (1) Acoustic image noise is large: Due to the complexity of the seabed environment, there is a lot of noise in the acoustic image, which affects the accurate detection and tracking of the seabed position.

[0005] (2) Different seabed materials produce different echo intensities: Different seabed materials produce different echo intensities, which leads to errors in the detection and tracking results of the seabed position.

[0006] (3) Interference from suspended targets in water: Objects suspended in water can cause echo interference in the acoustic image, thus affecting the accuracy of the seabed position.

[0007] In summary, there is an urgent need to provide a seabed position detection and tracking method that can effectively solve the above problems. Summary of the invention

[0008] In view of the deficiencies in the prior art, the present invention provides a method for detecting and tracking the seabed position based on sonar image features to solve the technical problems in the prior art: "(1) large acoustic image noise: due to the complexity of the seabed environment, there is a large amount of noise in the acoustic image, which affects the accurate detection and tracking of the seabed position; (2) large differences in echo intensity due to different bottom materials: different seabed materials will produce different echo intensities, resulting in errors in the detection and tracking results of the seabed position; (3) interference from suspended targets in the water: suspended objects in the water will cause echo interference in the acoustic image, thereby affecting the accuracy of the seabed position."

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for detecting and tracking seabed positions based on sonar image features, comprising the following steps:

[0010] S1: Automatic detection of the initial bottom position interval: At startup, the position of the bottom area is estimated by analyzing and processing through multi-cycle accumulation. This step can automatically detect the initial bottom position interval and provide an accurate starting point for subsequent bottom detection;

[0011] S2: At the initial position, the sonar remains stationary and accumulates multiple Ping data, recorded as S i (i=1,2,...,M), M is the number of accumulated data;

[0012] S3: Process the above data, using the data of the bottom-sounding sensor or using the acoustic data for calculation, assuming that the acquired seabed position is I i (i=1,2,...,M), then the initial seafloor position is:

[0013]

[0014] S4: Start sonar movement, perform continuous imaging, and process continuous single Ping data;

[0015] S5: Use three feature extraction methods to perform bottom detection and identify the location of the seabed;

[0016] S6: Median filtering, the bottom positions calculated by the above 5 methods are recorded as I1, I2, I3, I4. Considering that the seabed changes slowly, the data of the previous Ping is introduced as I0. When the first Ping data is initially calculated,

[0017] I0=I org

[0018] Perform median filtering to remove points with large fluctuations. The calculation method is as follows

[0019] Note as I mid is the median of the five values, and judges I1, I2, I3, and I4

[0020]

[0021] Among them I Δ To set the threshold;

[0022] S7: weighted average, perform weighted average on the I retained in the above 6 to obtain the true bottom position.

[0023]

[0024] where a i As weight, add the bottom position I0 of the previous Ping to ensure the stability of the result;

[0025] S8: Bottom baseline smoothing, the bottom position calculated above is recorded as I n (n=1,2,3,...), n is the Ping number.

[0026] The result is further smoothed to remove noise points. The final calculated bottom position is

[0027]

[0028] The window width is 2w.

[0029] As the preferred technical solution of the present invention, when the seabed echo arrives, the energy characteristics before and after are very different, and this characteristic can be used for bottom detection. Single Ping data is a one-dimensional array, whose subscripts represent the time sequence and the numerical value represents the sound intensity. When processing it, set a time window (length), divide the time window into two parts, and calculate the energy difference between the two parts before and after the time window; assuming that the time window length is 2k, the energy of the first half of the time window is recorded as P1, and the energy of the second half is recorded as P2, and different methods are used to evaluate the window energy difference.

[0030] As a preferred technical solution of the present invention, the different methods include: Perform window sliding calculation to find the position where P is the largest and record it as the bottom position, where

[0031]

[0032] y represents an element in the Ping data.

[0033] As a preferred technical solution of the present invention, the different methods also include: Perform window sliding calculation to find the position where P is the largest and record it as the bottom position, where y represents an element in the Ping data.

[0034] As a preferred technical solution of the present invention, the different methods also include: Perform window sliding calculation to find the position where P is the largest and record it as the bottom position, where

[0035]

[0036] y represents an element in the Ping data.

[0037] As a preferred technical solution of the present invention, the different methods also include: P = P2-P1, performing window sliding calculation, finding the position where P is the largest, and recording it as the bottom position, where

[0038]

[0039] y represents an element in the Ping data.

[0040] The present invention provides a method for detecting and tracking seabed positions based on sonar image features, which has the following features:

[0041] Beneficial effects:

[0042] The present invention proposes a method for seabed position detection and tracking based on image sonar features. Through the steps of automatic detection of initial bottom position interval, bottom detection, bottom detection data quality assessment and seabed position tracking and prediction, it can effectively solve difficult problems such as acoustic image noise, bottom echo difference and interference from suspended targets in the water. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 The present invention provides a flow chart of a method for detecting and tracking seabed positions based on sonar image features. DETAILED DESCRIPTION

[0044] The following embodiments of the present invention are described in further detail in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0045] Embodiment 1

[0046] refer to Figure 1 , a method for detecting and tracking seabed position based on sonar image features, comprising the following steps:

[0047] S1: Automatic detection of the initial bottom position interval: At startup, the position of the bottom area is estimated by analyzing and processing through multi-cycle accumulation. This step can automatically detect the initial bottom position interval and provide an accurate starting point for subsequent bottom detection;

[0048] S2: At the initial position, the sonar remains stationary and accumulates multiple Ping data, recorded as S i (i=1,2,...,M), M is the number of accumulated data;

[0049] S3: Process the above data, using the data of the bottom-sounding sensor or using the acoustic data for calculation, assuming that the acquired seabed position is I i (i=1,2,...,M), then the initial seafloor position is:

[0050]

[0051] S4: Start sonar movement, perform continuous imaging, and process continuous single Ping data;

[0052] S5: Use three feature extraction methods to perform bottom detection and identify the location of the seabed;

[0053] S6: Median filtering, the bottom positions calculated by the above 5 methods are recorded as I1, I2, I3, I4. Considering that the seabed changes slowly, the data of the previous Ping is introduced as I0. When the first Ping data is initially calculated,

[0054] I0=I org

[0055] Perform median filtering to remove points with large fluctuations. The calculation method is as follows

[0056] Note as I mid is the median of the five values, and judges I1, I2, I3, and I4

[0057]

[0058] Among them I Δ To set the threshold;

[0059] S7: weighted average, perform weighted average on the I retained in the above 6 to obtain the true bottom position.

[0060]

[0061] where a i As weight, add the bottom position I0 of the previous Ping to ensure the stability of the result;

[0062] S8: Bottom baseline smoothing, the bottom position calculated above is recorded as I n (n=1,2,3,...), n is the Ping number.

[0063] The result is further smoothed to remove noise points. The final calculated bottom position is

[0064]

[0065] The window width is 2w.

[0066] Among them, when the seabed echo arrives, the energy characteristics before and after are very different, and this feature can be used for bottom detection. Single Ping data is a one-dimensional array, the subscript represents the time sequence, and the value represents the sound intensity. When processing it, set a time window (length), divide the time window into two parts, and calculate the energy difference between the two parts before and after the time window; assuming that the time window length is 2k, the energy of the first half of the time window is recorded as P1, and the energy of the second half is recorded as P2. The method used to evaluate the window energy difference is: Perform window sliding calculation to find the position where P is the largest and record it as the bottom position, where

[0067]

[0068] y represents an element in the Ping data.

[0069] Embodiment 2

[0070] A method for detecting and tracking seabed positions based on sonar image features comprises the following steps:

[0071] S1: Automatic detection of the initial bottom position interval: At startup, the position of the bottom area is estimated by analyzing and processing through multi-cycle accumulation. This step can automatically detect the initial bottom position interval and provide an accurate starting point for subsequent bottom detection;

[0072] S2: At the initial position, the sonar remains stationary and accumulates multiple Ping data, recorded as S i (i=1,2,...,M), M is the number of accumulated data;

[0073] S3: Process the above data, using the data of the bottom-sounding sensor or using the acoustic data for calculation, assuming that the acquired seabed position is I i (i=1,2,...,M), then the initial seafloor position is:

[0074]

[0075] S4: Start sonar movement, perform continuous imaging, and process continuous single Ping data;

[0076] S5: Use three feature extraction methods to perform bottom detection and identify the location of the seabed;

[0077] S6: Median filtering, the bottom positions calculated by the above 5 methods are recorded as I1, I2, I3, I4. Considering that the seabed changes slowly, the data of the previous Ping is introduced as I0. When the first Ping data is initially calculated,

[0078] I0=I org

[0079] Perform median filtering to remove points with large fluctuations. The calculation method is as follows

[0080] Note as I mid is the median of the five values, and judges I1, I2, I3, and I4

[0081]

[0082] Among them I Δ To set the threshold;

[0083] S7: weighted average, weighted average is performed on the I retained in the above 6 to obtain the true bottom position.

[0084]

[0085] where a i As weight, add the bottom position I0 of the previous Ping to ensure the stability of the result;

[0086] S8: Bottom baseline smoothing, the bottom position calculated above is recorded as I n (n=1,2,3,...), n is the Ping number.

[0087] The result is further smoothed to remove noise points. The final calculated bottom position is

[0088]

[0089] The window width is 2w.

[0090] Among them, when the seabed echo arrives, the energy characteristics before and after are very different, and this feature can be used for bottom detection. Single Ping data is a one-dimensional array, the subscript represents the time sequence, and the value represents the sound intensity. When processing it, set a time window (length), divide the time window into two parts, and calculate the energy difference between the two parts before and after the time window; assuming that the time window length is 2k, the energy of the first half of the time window is recorded as P1, and the energy of the second half is recorded as P2. The method used to evaluate the window energy difference is: Perform window sliding calculation to find the position where P is the largest and record it as the bottom position, where y represents an element in the Ping data.

[0091] Embodiment 3

[0092] A method for detecting and tracking seabed positions based on sonar image features comprises the following steps:

[0093] S1: Automatic detection of the initial bottom position interval: At startup, the position of the bottom area is estimated by analyzing and processing through multi-cycle accumulation. This step can automatically detect the initial bottom position interval and provide an accurate starting point for subsequent bottom detection;

[0094] S2: At the initial position, the sonar remains stationary and accumulates multiple Ping data, recorded as S i (i=1,2,...,M), M is the number of accumulated data;

[0095] S3: Process the above data, using the data of the bottom-sounding sensor or using the acoustic data for calculation, assuming that the acquired seabed position is I i (i=1,2,...,M), then the initial seafloor position is:

[0096]

[0097] S4: Start sonar movement, perform continuous imaging, and process continuous single Ping data;

[0098] S5: Use three feature extraction methods to perform bottom detection and identify the location of the seabed;

[0099] S6: Median filtering, the bottom positions calculated by the above 5 methods are recorded as I1, I2, I3, I4. Considering that the seabed changes slowly, the data of the previous Ping is introduced as I0. When the first Ping data is initially calculated,

[0100] I0=I org

[0101] Perform median filtering to remove points with large fluctuations. The calculation method is as follows

[0102] Note as I mid is the median of the five values, and judges I1, I2, I3, and I4

[0103]

[0104] Among them I Δ To set the threshold;

[0105] S7: weighted average, weighted average is performed on the I retained in the above 6 to obtain the true bottom position.

[0106]

[0107] where a i As weight, add the bottom position I0 of the previous Ping to ensure the stability of the result;

[0108] S8: Bottom baseline smoothing, the bottom position calculated above is recorded as I n (n=1,2,3,...), n is the Ping number.

[0109] The result is further smoothed to remove noise points. The final calculated bottom position is

[0110]

[0111] The window width is 2w.

[0112] Among them, when the seabed echo arrives, the energy characteristics before and after are very different, and this feature can be used for bottom detection. Single Ping data is a one-dimensional array, the subscript represents the time sequence, and the value represents the sound intensity. When processing it, set a time window (length), divide the time window into two parts, and calculate the energy difference between the two parts before and after the time window; assuming that the time window length is 2k, the energy of the first half of the time window is recorded as P1, and the energy of the second half is recorded as P2. The method used to evaluate the window energy difference is: Perform window sliding calculation to find the position where P is the largest and record it as the bottom position, where

[0113]

[0114] y represents an element in the Ping data.

[0115] Embodiment 4

[0116] A method for detecting and tracking seabed positions based on sonar image features comprises the following steps:

[0117] S1: Automatic detection of the initial bottom position interval: At startup, the position of the bottom area is estimated by analyzing and processing through multi-cycle accumulation. This step can automatically detect the initial bottom position interval and provide an accurate starting point for subsequent bottom detection;

[0118] S2: At the initial position, the sonar remains stationary and accumulates multiple Ping data, recorded as S i (i=1,2,...,M), M is the number of accumulated data;

[0119] S3: Process the above data, using the data of the bottom-sounding sensor or using the acoustic data for calculation, assuming that the acquired seabed position is I i (i=1,2,...,M), then the initial seafloor position is:

[0120]

[0121] S4: Start sonar movement, perform continuous imaging, and process continuous single Ping data;

[0122] S5: Use three feature extraction methods to perform bottom detection and identify the location of the seabed;

[0123] S6: Median filtering, the bottom positions calculated by the above 5 methods are recorded as I1, I2, I3, I4. Considering that the seabed changes slowly, the data of the previous Ping is introduced as I0. When the first Ping data is initially calculated,

[0124] I0=I org

[0125] Perform median filtering to remove points with large fluctuations. The calculation method is as follows

[0126] Note as I mid is the median of the five values, and judges I1, I2, I3, and I4

[0127]

[0128] Among them I Δ To set the threshold;

[0129] S7: weighted average, perform weighted average on the I retained in the above 6 to obtain the true bottom position.

[0130]

[0131] where a i As weight, add the bottom position I0 of the previous Ping to ensure the stability of the result;

[0132] S8: Bottom baseline smoothing, the bottom position calculated above is recorded as I n (n=1,2,3,...), n is the Ping number.

[0133] The result is further smoothed to remove noise points. The final calculated bottom position is

[0134]

[0135] The window width is 2w.

[0136] Among them, when the seabed echo arrives, there is a big difference in the energy characteristics before and after, and this feature can be used for bottom detection. Single Ping data is a one-dimensional array, the subscript represents the time sequence, and the value represents the sound intensity. When processing it, set a time window (length), divide the time window into two equal parts, and calculate the energy difference between the two parts before and after the time window; assuming that the time window length is 2k, the energy of the first half of the time window is recorded as P1, and the energy of the second half is recorded as P2. The method used to evaluate the window energy difference is: P = P2-P1, perform window sliding calculation, find the position where P is the largest, and record it as the bottom position, where

[0137]

[0138] y represents an element in the Ping data.

[0139] The above is only a preferred specific implementation of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field within the technical scope disclosed by the present invention, according to the technical solution and the inventive concept of the present invention, shall be covered within the protection scope of the present invention. For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention.

Claims

1. A method for detecting and tracking seabed position based on sonar image features, characterized in that: The following steps are included: S1: Automatic detection of the initial bottom position interval: At startup, the position of the bottom area is estimated by analyzing and processing through multi-cycle accumulation. This step can automatically detect the initial bottom position interval and provide an accurate starting point for subsequent bottom detection; S2: At the initial position, the sonar remains stationary and accumulates multiple Ping data, recorded as S i , i=1,2,...,M, M is the number of accumulated data; S3: Process the above data, using the data of the bottom-sounding sensor or using the acoustic data for calculation, assuming that the acquired seabed position is I i , i=1,2,...,M, then the initial seafloor position is: S4: Start sonar movement, perform continuous imaging, and process continuous single Ping data; S5: Use three feature extraction methods to perform bottom detection and identify the location of the seabed; S6: Median filtering, the bottom position calculated by multiple methods in the above step S5 is recorded as I1, I2, I3, I 4； Considering that the seabed changes slowly, the data of the previous ping is introduced as I0. When the first ping data is initially calculated, I0=I org Perform median filtering to remove points with large fluctuations; The calculation method is as follows: Note as I mid is the median of the five values, and judges I1, I2, I3, and I4 Among them I Δ To set the threshold; S7: weighted average, weighted average the I retained in the above step S6 to obtain the true bottom position: where a i As weight, add the bottom position I0 of the previous Ping to ensure the stability of the result; S8: Bottom baseline smoothing, the bottom position calculated above is recorded as I n , n=1,2,3,..., n is the Ping number; The result is further smoothed to remove noise points, and the final calculated bottom position is: The window width is 2w.

2. The method for detecting and tracking seabed position based on sonar image features according to claim 1, characterized in that: When the seabed echo arrives, the energy characteristics before and after are very different. This characteristic is used to detect the bottom. Single Ping data is a one-dimensional array, the subscript represents the time sequence, and the value represents the sound intensity. When processing it, a time window length is set, and the time window is divided into two parts, and the energy difference between the two parts before and after the time window is calculated; Assuming that the time window length is 2k, the energy of the first half of the time window is recorded as P1, and the energy of the second half is recorded as P2. Different methods are used to evaluate the window energy difference.

3. The method for detecting and tracking seabed position based on sonar image features according to claim 2, characterized in that: The different methods include: Perform window sliding calculation to find the position where P is the largest and record it as the bottom position, where y represents an element in the Ping data.

4. The method for detecting and tracking seabed position based on sonar image features according to claim 2, characterized in that: The different methods also include: Perform window sliding calculation to find the position where P is the largest and record it as the bottom position, where y represents an element in the Ping data.

5. The method for detecting and tracking seabed position based on sonar image features according to claim 2, characterized in that: The different methods also include: Perform window sliding calculation to find the position where P is the largest and record it as the bottom position, where y represents an element in the Ping data.

6. The method for detecting and tracking seabed position based on sonar image features according to claim 2, characterized in that: The different methods also include: P = P2-P1, performing window sliding calculation, finding the position where P is the largest, and recording it as the bottom position, where y represents an element in the Ping data.

Images