An underwater target recognition method based on contrast calculation of acoustic imaging

Abstract

The present application relates to a kind of underwater target identification methods of acoustic imaging based on contrast calculation, the steps of the method include: step 1, the scale and position information of all suspected targets are obtained according to intelligent identification algorithm, including real target and false alarm determined by intelligent identification algorithm;Step 2, acoustic imaging background normalization;Step 3, using mean filter to remove the pepper and salt noise in acoustic imaging data.The present application combines the processing method with intelligent identification method, reduces false alarm through multi-threshold setting, improves target identification accuracy.

Description

Technical fields:

[0001] This invention belongs to the field of underwater acoustic engineering technology, specifically relating to an acoustic imaging underwater target recognition method based on contrast calculation. Background technology:

[0002] Underwater target identification and localization is becoming increasingly important in the field of underwater acoustics. Imaging sonar, especially side-scan sonar and synthetic aperture sonar, occupies a very important position in the identification of small underwater targets. When identifying static underwater targets, imaging sonar mainly analyzes the acoustic imaging reflection characteristics of the target.

[0003] In underwater target acoustic imaging, the transmitted signal, after being reflected by the target and compensated for time delay and spatial motion errors, can be used to construct the acoustic imaging results of the target. Based on the sonar's imaging range and azimuth resolution, the target's position and imaging target scale can be calculated. However, when the underwater environment is complex, target identification in imaging sonar becomes difficult, as many terrain features similar to small targets are easily misidentified as targets. Currently, intelligent underwater acoustic imaging target identification suffers from a high rate of false alarms in complex terrain. Summary of the Invention:

[0004] The technical problem to be solved by the present invention is to provide an acoustic imaging underwater target recognition method based on contrast calculation. This method utilizes the high reflection intensity of the target signal and uses a target characteristic value with a similar signal-to-noise ratio to suppress false alarms in complex terrain and improve the target recognition accuracy.

[0005] The technical solution of this invention is to provide a method for underwater target identification based on acoustic imaging with contrast calculation, the method comprising the following steps:

[0006] Step 1: Obtain the scale and location information of all suspected targets based on the intelligent recognition algorithm, including real targets and false alarms determined by the intelligent recognition algorithm.

[0007] Step 2: Acoustic Imaging Background Normalization. Due to the energy attenuation of the signal reflection process, the acoustic imaging background needs to be normalized. Assuming the acoustic imaging area is M×N, a separation window of m×n is set according to the target size. The background equalization processing function is as follows:

[0008] Weight = conv(data) M×N ,Window m×n )

[0009] Data = data.*(1 / Weight)

[0010] Where data is the input data, Window is the split window, Weight is the calculated weight value, and Data is the background equalization data.

[0011] Step 3: Use a mean filter to remove salt-and-pepper noise from the acoustic imaging data. Set the filter size to k×l. The processing function is as follows:

[0012]

[0013] Data = conv(Data,kernal) k×l stride=1)

[0014] Step 4: Extract the target information obtained in Step 1, integrate the target location information, and calculate the target intensity. To reduce the amount of calculation, based on the selected target information, set windows on the upper, middle, and lower sides of the target area's central axis, calculate the average value, and then take the maximum value among the average values ​​as the target intensity. The specific calculation process is as follows:

[0015] Target=Data target

[0016] T = max(mean(Target) top ),mean(Target mid ),mean(Target down ))

[0017] Where T represents the intensity of the target.

[0018] Step 5: Based on the target information in Step 3, calculate the intensity of the surrounding background. The specific calculation process is as follows:

[0019] Background=Data background

[0020] B_mean = mean(Background)

[0021]

[0022] When calculating the Background, the target area needs to be removed, and Num represents the pixel value of the surrounding background.

[0023] Step 6: Calculate the contrast between the target and its surrounding environment. The specific calculation formula is as follows:

[0024] TBR = abs(T - B_mean) / B_inf

[0025] Step 7: Set the target confidence weights according to the target scale obtained in Step 1. The specific weight settings are as follows, in meters:

[0026]

[0027] Step 8: Weight the contrast ratio according to the weight value, and set a threshold for the weighted contrast ratio to suppress false alarms.

[0028] That is, TBR = 20 * lg(TBR * c).

[0029] Compared with the prior art, the present invention has the following advantages:

[0030] This invention addresses the problem of high false alarm rates in underwater target acoustic imaging recognition in complex terrain. It proposes an underwater target recognition method based on contrast calculation, combining this processing method with intelligent recognition methods. By setting multiple thresholds, false alarms are reduced and the target recognition accuracy is improved. Attached image description:

[0031] Figure 1 This is a flowchart of the process of this invention;

[0032] Figure 2 The target contrast result is calculated based on actual synthetic aperture acoustic imaging data, which is the result of this invention. Detailed implementation method:

[0033] The present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0034] like Figure 1 As shown, the present invention provides an acoustic imaging underwater target recognition method based on contrast calculation. Based on imaging sonar data, a large amount of target location and scale information is obtained through an intelligent recognition and positioning algorithm. Based on the obtained target information, target intensity information and background information are obtained. The contrast calculation method proposed in the method is used to perform preliminary screening of a large number of targets.

[0035] Specifically, the steps of this underwater target size prediction method include:

[0036] Step 1: Obtain the scale and location information of all suspected targets based on the intelligent recognition algorithm, including real targets and false alarms determined by the intelligent recognition algorithm.

[0037] Step 2: Acoustic Imaging Background Normalization. Due to the energy attenuation of the signal reflection process, the acoustic imaging background needs to be normalized. Assuming the acoustic imaging area is M×N, a separation window of m×n is set according to the target size. The background equalization processing function is as follows:

[0038] Weight = conv(data) M×N ,Window m×n )

[0039] Data = data.*(1 / Weight)

[0040] Where data is the input data, M is 512, N is 1024, Window is the split window, m is 50, n is 200, Weight is the calculated weight value, and Data is the background equalization data.

[0041] Step 3: Use a mean filter to remove salt-and-pepper noise from the acoustic imaging data. Set the filter size to k×l. The processing function is as follows:

[0042]

[0043] Data = conv(Data,kernal) k×l stride=1)

[0044] Where k is 5 and l is 12.

[0045] Step 4: Extract the target information obtained in Step 1, integrate the target location information, and calculate the target intensity. To reduce the amount of calculation, based on the selected target information, set windows j×k on the upper, middle, and lower sides of the target region's central axis, calculate the average value, and then take the maximum value of the average value as the target intensity. The specific calculation process is as follows:

[0046] Target=Data target

[0047] T = max(mean(Target) top ),mean(Target mid ),mean(Target down ))

[0048] Where T represents the intensity of the target, and j is 6 and k is 24 in the window for extracting target information.

[0049] Step 5: Based on the target information in Step 3, calculate the intensity of the surrounding background. The specific calculation process is as follows:

[0050] Background=Data background

[0051] B_mean = mean(Background)

[0052]

[0053] When calculating the Background, the target area needs to be removed, and Num represents the pixel value of the surrounding background.

[0054] Step 6: Calculate the contrast between the target and its surrounding environment. The specific calculation formula is as follows:

[0055] TBR = abs(T - B_mean) / B_inf

[0056] Step 7: Set the target confidence weights according to the target scale obtained in Step 1. The specific weight settings are as follows, in meters:

[0057]

[0058] Step 8: Weight the contrast ratio according to the weight value, and set a threshold for the weighted contrast ratio to suppress false alarms.

[0059] TBR = 20 * lg(TBR * c)

[0060] like Figure 2 As shown, this invention calculates the target contrast based on target information obtained from acoustic imaging data and intelligent identification and positioning algorithms, and further filters a large number of targets based on threshold information to reduce false alarms in the underwater target identification process.

[0061] The above description only illustrates preferred embodiments of the present invention and should not be construed as limiting the scope of the claims. Any equivalent procedural modifications made using this specification are included within the patent protection scope of this invention.

Claims

1. A method for underwater target recognition based on acoustic imaging with contrast calculation, characterized in that: Includes the following steps, Step 1: Obtain the scale and location information of all suspected targets based on the intelligent recognition algorithm; Step 2: Acoustic imaging background normalization. Assuming the acoustic imaging region is M×N, a separation window of m×n is set according to the target scale. The background equalization processing function is as follows: Weight=conv(data M×N ,Window m×n ) Data = data.*(1 / Weight) Where data is the input data, Window is the split window, Weight is the calculated weight value, and Data is the background equalization data; Step 3: Use a mean filter to remove salt-and-pepper noise from the acoustic imaging data. Set the filter size to k×l, and the processing function is as follows: Data=conv(Data,kernal k×l ,stride=1) Step 4: Extract the target information obtained in Step 1, integrate the target location information, and calculate the target intensity. The specific calculation process is as follows: Target=Data target T=max(mean(Target top ),mean(Target mid ),mean(Target down )) Where T is the intensity of the target; Step 5: Based on the target information in Step 3, calculate the intensity of the surrounding background. The specific calculation process is as follows: Background=Data background B_mean = mean(Background) Step 6: Calculate the contrast between the target and its surrounding environment. The specific calculation formula is as follows: TBR = abs(T - B_mean) / B_inf Step 7: Set the target confidence weights according to the target scale obtained in Step 1. The specific weight settings are as follows, in meters: Step 8: Weight the contrast ratio according to the weight value, and set a threshold for the weighted contrast ratio to suppress false alarms.

2. The acoustic imaging underwater target recognition method based on contrast calculation according to claim 1, characterized in that: In step 1, suspected targets include real targets and false alarms identified by intelligent recognition algorithms.

3. The acoustic imaging underwater target recognition method based on contrast calculation according to claim 1, characterized in that: In step 4, to reduce the amount of calculation, windows are set on the upper, middle and lower sides of the central axis of the target area based on the selected target information, the average value is calculated, and the maximum value of the average value is taken as the intensity of the target.

4. The acoustic imaging underwater target recognition method based on contrast calculation according to claim 1, characterized in that: When calculating the Background, the target area needs to be removed, and Num represents the pixel value of the surrounding background.

Images