阅读说明：本技术 水下目标检测识别装置及方法 (Underwater target detection and identification device and method ) 是由 谢翔 邹少锋 李国林 王志华 于 2020-03-30 设计创作，主要内容包括：本发明实施例提供一种水下目标检测识别装置及方法,其中装置包括配准单元、检测识别单元,以及装设在水下载体每一侧的至少两个具有不同扫描方向的单视侧扫声呐；所述配准单元用于对水下载体单侧所有单视侧扫声呐针对同一水下目标采集的单视声图进行图像配准,得到图像配准结果；所述检测识别单元用于基于所述图像配准结果,确定所述水下目标的检测识别结果。本发明实施例提供的装置及方法,提高了水下目标的检测率和识别率,降低了虚警率,同时,实现了一次航行路径对水下目标的多次检测和识别,提高了水下载体的能效比,降低了成本。(The embodiment of the invention provides an underwater target detection and identification device and a method, wherein the device comprises a registration unit, a detection and identification unit and at least two single-view side-scan sonars which are arranged on each side of an underwater carrier and have different scanning directions; the registration unit is used for carrying out image registration on single-view sonograms acquired by all single-view side-scan sonars on one side of the underwater carrier aiming at the same underwater target to obtain an image registration result; the detection and identification unit is used for determining the detection and identification result of the underwater target based on the image registration result. The device and the method provided by the embodiment of the invention improve the detection rate and the recognition rate of the underwater target, reduce the false alarm rate, realize the repeated detection and recognition of the underwater target by one navigation path, improve the energy efficiency ratio of the underwater carrier and reduce the cost.)

1. An underwater target detection and identification device is characterized by comprising a registration unit, a detection and identification unit and at least two single-view side-scan sonars which are arranged on each side of an underwater carrier and have different scanning directions;

the registration unit is used for carrying out image registration on single-view sonograms acquired by all single-view side-scan sonars on one side of the underwater carrier aiming at the same underwater target to obtain an image registration result;

the detection and identification unit is used for determining the detection and identification result of the underwater target based on the image registration result.

2. The underwater object detection and recognition device according to claim 1, wherein the detection recognition unit includes a first detection recognition subunit;

the first detection and identification subunit is used for carrying out target detection processing on each single-view sonogram to obtain an underwater target image area in each single-view sonogram and a corresponding confidence coefficient thereof, and determining the detection and identification result of the underwater target based on the image registration result, the underwater target image area in each single-view sonogram and the corresponding confidence coefficient thereof.

3. The underwater object detection and recognition device according to claim 1, wherein the detection recognition unit includes a second detection recognition subunit;

the second detection and identification subunit is used for inputting the image registration result into a joint detection and identification model to obtain a detection and identification result of the underwater target output by the joint detection and identification model;

and the joint detection and identification model is obtained by training based on the sample image registration result and the detection and identification result of the corresponding sample underwater target.

4. The underwater object detection and identification device according to any one of claims 1 to 3, wherein the registration unit is specifically configured to:

and carrying out image registration based on multi-information fusion on the single-view sonograms acquired by all the single-view side-scan sonars on the single side of the underwater carrier aiming at the same underwater target, wherein the multi-information fusion comprises fusion of at least two of the navigation speed of the underwater carrier, the position of the underwater carrier, the posture of the single-view side-scan sonar and the sonogram.

5. The underwater object detection and recognition device according to any one of claims 1 to 3, further comprising a feature enhancement unit:

the feature enhancement unit is used for performing target feature enhancement on each single-view sound image by adopting a super-resolution algorithm and/or an image enhancement algorithm.

6. The underwater object detection and recognition device according to any one of claims 1 to 3, wherein each single view side scan sonar is provided with a backflow enclosure cover.

7. An underwater target detection and identification method is characterized by comprising the following steps:

carrying out image registration on single-view sonograms collected by all single-view side-scan sonars on one side of the underwater carrier aiming at the same underwater target to obtain an image registration result;

determining a detection and identification result of the underwater target based on the image registration result;

wherein, each side of the underwater carrier is provided with at least two single-view side-scan sonars with different scanning directions.

8. The method according to claim 7, wherein the determining the detection and identification result of the underwater target based on the image registration result specifically includes:

carrying out target detection processing on each single-vision sound image to obtain an underwater target image area in each single-vision sound image and a corresponding confidence coefficient thereof;

and determining the detection and identification result of the underwater target based on the image registration result, the underwater target image area in each single-visual sound image and the corresponding confidence coefficient of the underwater target image area.

9. The method according to claim 7, wherein the determining the detection and identification result of the underwater target based on the image registration result specifically includes:

inputting the image registration result into a joint detection and identification model to obtain a detection and identification result of the underwater target output by the joint detection and identification model;

and the joint detection and identification model is obtained by training based on the sample image registration result and the detection and identification result of the corresponding sample underwater target.

10. The underwater target detection and identification method according to any one of claims 7 to 9, wherein the image registration of all the single-view side-scan sonars on the single side of the underwater vehicle on the single-view sound image acquired by the same underwater target comprises:

and carrying out image registration based on multi-information fusion on the single-view sonograms acquired by all the single-view side-scan sonars on the single side of the underwater carrier aiming at the same underwater target, wherein the multi-information fusion comprises fusion of at least two of the navigation speed of the underwater carrier, the position of the underwater carrier, the posture of the single-view side-scan sonar and the sonogram.

Technical Field

The invention relates to the technical field of computer vision, in particular to an underwater target detection and identification device and method.

Background

The frequency of sonar is very low relative to the frequency of light, and strong noise (such as reverberation noise) exists in the actual sea state, so that the resolution of the obtained sound image is low, and the target edge is blurred. Therefore, efficient detection and identification of underwater targets based on sonograms is a hotspot and difficulty of research in academic and industrial fields nowadays.

The existing underwater target detection and identification system utilizes a side scan sonar to detect underwater targets from a single position and a single visual angle, and when a certain scanning direction or a certain scanning direction has larger noise and/or the targets are shielded and partially shielded by small non-target objects, the target detection rate is low, the identification rate is low and the false alarm rate is high.

At present, the solution to this problem is to perform multi-view scanning on an underwater target by using methods such as multi-path navigation or UUV (unmanned underwater vehicle) formation, however, the implementation cost of the above method is high, and the UUV energy efficiency ratio is low.

Disclosure of Invention

The embodiment of the invention provides an underwater target detection and identification device and method, which are used for solving the problems of low target detection rate, low identification rate and high false alarm rate of the existing underwater target detection and identification system.

In a first aspect, an embodiment of the present invention provides an underwater target detection and identification device, including a registration unit, a detection and identification unit, and at least two single-view side-scan sonars installed on each side of an underwater carrier and having different scanning directions;

the registration unit is used for carrying out image registration on single-view sonograms acquired by all single-view side-scan sonars on one side of the underwater carrier aiming at the same underwater target to obtain an image registration result;

the detection and identification unit is used for determining the detection and identification result of the underwater target based on the image registration result.

Optionally, the detection identification unit comprises a first detection identification subunit;

the first detection and identification subunit is used for carrying out target detection processing on each single-view sonogram to obtain an underwater target image area in each single-view sonogram and a corresponding confidence coefficient thereof, and determining the detection and identification result of the underwater target based on the image registration result, the underwater target image area in each single-view sonogram and the corresponding confidence coefficient thereof.

Optionally, the detection identification unit comprises a second detection identification subunit;

the second detection and identification subunit is used for inputting the image registration result into a joint detection and identification model to obtain a detection and identification result of the underwater target output by the joint detection and identification model;

and the joint detection and identification model is obtained by training based on the sample image registration result and the detection and identification result of the corresponding sample underwater target.

Optionally, the registration unit is specifically configured to:

and carrying out image registration based on multi-information fusion on the single-view sonograms acquired by all the single-view side-scan sonars on the single side of the underwater carrier aiming at the same underwater target, wherein the multi-information fusion comprises fusion of at least two of the navigation speed of the underwater carrier, the position of the underwater carrier, the posture of the single-view side-scan sonar and the sonogram.

Optionally, the apparatus further comprises a feature enhancement unit:

the feature enhancement unit is used for performing target feature enhancement on each single-view sound image by adopting a super-resolution algorithm and/or an image enhancement algorithm.

Optionally, each single view side scan sonar is configured with a backflow enclosure housing.

In a second aspect, an embodiment of the present invention provides an underwater target detection and identification method, including:

carrying out image registration on single-view sonograms collected by all single-view side-scan sonars on one side of the underwater carrier aiming at the same underwater target to obtain an image registration result;

determining a detection and identification result of the underwater target based on the image registration result;

wherein, each side of the underwater carrier is provided with at least two single-view side-scan sonars with different scanning directions.

Optionally, the determining, based on the image registration result, the detection and identification result of the underwater target specifically includes:

carrying out target detection processing on each single-vision sound image to obtain an underwater target image area in each single-vision sound image and a corresponding confidence coefficient thereof;

and determining the detection and identification result of the underwater target based on the image registration result, the underwater target image area in each single-visual sound image and the corresponding confidence coefficient of the underwater target image area.

Optionally, the determining, based on the image registration result, the detection and identification result of the underwater target specifically includes:

inputting the image registration result into a joint detection and identification model to obtain a detection and identification result of the underwater target output by the joint detection and identification model;

and the joint detection and identification model is obtained by training based on the sample image registration result and the detection and identification result of the corresponding sample underwater target.

Optionally, the image registration of the single-view sonograms acquired by all the single-view side-scan sonars on the single side of the underwater vehicle for the same underwater target includes:

and carrying out image registration based on multi-information fusion on the single-view sonograms acquired by all the single-view side-scan sonars on the single side of the underwater carrier aiming at the same underwater target, wherein the multi-information fusion comprises fusion of at least two of the navigation speed of the underwater carrier, the position of the underwater carrier, the posture of the single-view side-scan sonar and the sonogram.

According to the underwater target detection and identification device and method provided by the embodiment of the invention, the single-view side-scan sonars with different scanning angles are respectively arranged at a plurality of different positions on each side of the underwater carrier, so that a plurality of sound images of the same target can be obtained from different positions and different visual angles, the probability of large noise interference at the positions is greatly reduced, the phenomenon that the target is shielded is avoided at a high probability, the detection rate and the identification rate of the underwater target are improved, the false alarm rate is reduced, meanwhile, the underwater target is detected and identified for a plurality of times by one navigation path, the energy efficiency ratio of the underwater carrier is improved, and the cost is reduced.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.

Fig. 1 is a schematic structural diagram of an underwater target detection and identification device provided in an embodiment of the present invention;

FIG. 2 is an imaging schematic diagram of an underwater target detection and identification device provided by an embodiment of the present invention;

FIG. 3 is an imaging schematic diagram of another underwater target detection and identification device provided by the embodiment of the invention;

fig. 4 is a schematic flow chart of the underwater target detection and identification method provided by the embodiment of the invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The current underwater target detection and identification mainly focuses on the ideal condition, namely, the detection and identification of the underwater target presenting a significantly highlighted and shaded area in the acoustic image, but even under the ideal condition, the target identification rate is still low (not more than 80%), and the missing rate is still high (about 30%). Under the actual complex sea condition, the highlight and shadow contours corresponding to underwater targets in the acquired high-frequency sound image often become very fuzzy, and even the corresponding highlight or shadow areas are submerged by reverberation noise, so that the loss phenomenon occurs. The target detection and identification under the conditions of fuzzy bright-shadow contour or loss of bright-shadow part and the like are difficult points and challenges in the development trend of the target detection technology under the actual complex sea condition.

The existing side scan sonar comprises two transducer linear arrays, namely transducer linear arrays which are respectively arranged on the left side and the right side of an underwater carrier. The difficulty of underwater target detection based on the existing side scan sonar is embodied as follows: when a certain sonar or a certain section of scanning direction has relatively large noise, the boundary outline of a light and dark area of an underwater target can be extremely fuzzy, the light and dark areas are not in one-to-one correspondence, and only the light area or the shadow area can exist, so that the missing detection or the false alarm rate is high; because the position and the posture of the target relative to the sonar are random, the shape of a target or a dark area is random, the target and a non-target (such as rock and the like) are difficult to distinguish, meanwhile, the view angle of the sonar is fixed, when the target is shielded and partially shielded by a small non-target object, the target is difficult to detect and recognize, and the missing detection or the false alarm rate is high.

At present, although the above problems exist in detecting underwater targets by side scan sonar at home and abroad, a multi-view underwater target detection technology implemented by methods such as multi-path navigation or formation has been proposed to improve the recognition rate and reduce the false alarm rate, the following problems still exist: due to the adoption of complex multi-path navigation, the energy efficiency ratio of the UUV is low; the use of UUV formation to detect underwater targets results in very high costs.

Fig. 1 is a schematic structural diagram of an underwater target detection and identification device provided in an embodiment of the present invention, and as shown in fig. 1, the device includes a registration unit 102, a detection and identification unit 103, and at least two single-view side-scan sonars 101 with different scanning directions, which are installed on each side of an underwater vehicle;

specifically, the apparatus includes a single-view side scan sonar 101, a registration unit 102, and a detection recognition unit 103. The single-view side-scan sonar only comprises a single transducer linear array, and is different from the existing side-scan sonar comprising two transducer linear arrays.

The underwater carrier is used for installing a single-view side-scan sonar to detect and identify an underwater target, the underwater carrier can be a UUV or a fish towing vehicle, and the embodiment of the invention is not particularly limited. The single-view side scan sonar 101 is used for detecting underwater targets and collecting sonograms of the underwater targets. A plurality of single-view side-scan sonars are arranged on each side of the underwater carrier, and the specific number can be configured according to actual needs. Because whole detection identification process is realized under the circumstances that the carrier navigation under water, consequently, can acquire the sonogram to same target in different moments, different visual angles and different positions through installing the at least two single-vision side scan sonar that have different scanning direction in every side of the carrier under water.

The following is illustrated by way of example. Fig. 2 is an imaging schematic diagram of the underwater target detection and recognition device provided by the embodiment of the present invention, and as shown in fig. 2, two single-view side-scan sonars are respectively installed on two sides of the underwater vehicle. When the underwater target appears on the right side of the navigation direction of the underwater carrier, two single-view side-scan sonars on the right side of the underwater carrier scan and detect the underwater target. Because the scanning directions of the right single-view side-scan sonars 1 and 2 are different, the underwater target can be respectively imaged at two different visual angles. When the underwater carrier walks and navigates, the single-view side-scan sonar 1 and the single-view side-scan sonar 2 on the right side respectively emit acoustic pulses of fan-shaped beams to the lower side, and the fan-shaped beam angles of the two beams form a fixed view angle difference and can be represented by a scanning included angle alpha.

In order to avoid interference between sonar signals, the right single-view side-scan sonar 1 is deployed at the front part of the carrier, and the right single-view side-scan sonar 2 is deployed at the rear part of the carrier. After the target is imaged by the right single-view side-scan sonar 1, the carrier moves forward by a certain distance, the right single-view side-scan sonar 2 can image the same target at another angle, and observation at different moments, different visual angles (the visual angle difference is the scanning included angle alpha) and different positions can be realized to greatly reduce the false alarm rate of underwater detection and the problem of missed detection caused by shielding to a certain degree. Preferably, the scanning direction of the single-view side-scan sonar can be adjusted so that the scanning included angle α is 90 ° to obtain a better detection and identification result. The embodiment of the invention does not specifically limit the adjustment mode of the scanning direction of the single-view side scan sonar.

Fig. 3 is an imaging schematic diagram of another underwater target detection and identification device provided by an embodiment of the present invention, and as shown in fig. 3, a single-view side scan sonar is respectively added to the middle positions of two sides of an underwater carrier, and each side forms 3 different viewing angles, so as to further improve the detection capability of the device. Under the condition that the space of the underwater carrier allows, more single-view side-scan sonars with different scanning angles can be arranged on two sides of the underwater carrier.

The method comprises the steps that single-view side-scan sonars with different scanning angles are respectively installed at a plurality of different positions on each side of an underwater carrier, so that two or more sound images of the same target are obtained from different positions and different visual angles, the probability that large noise interference exists at a plurality of positions is greatly reduced, the probability that the underwater target with a relatively obvious bright-dark area is obtained by scanning at least from one position or a plurality of positions is greatly improved, and the detection rate and the recognition rate of the underwater target are improved; meanwhile, the size, shape information and space information of the underwater target can be extracted with higher probability from the sound images at different positions and visual angles, and the phenomenon that the target is shielded can be avoided with higher probability.

The registration unit 102 is used for performing image registration on single-view sonograms acquired by all single-view side-scan sonars on a single side of the underwater carrier aiming at the same underwater target to obtain an image registration result;

specifically, each single-view side scan sonar of the single side of the underwater carrier has serious distortion and strong background noise interference aiming at the single-view sound image acquired by the same underwater target, so that the accurate registration is required to be carried out aiming at the sound-view images acquired at a plurality of different moments, different positions and different visual angles, namely, the registration and fusion are carried out on the same area in the plurality of sound-view images, so that the observation of the same target from a plurality of different positions and visual angles is realized, the probability of large noise interference existing at a plurality of positions is reduced, and the phenomenon that the target is shielded is avoided.

Before image registration, in order to obtain a high-quality image registration result, an acoustic image preprocessing operation may be performed on each single-view acoustic image. The sonogram pre-processing operation includes at least one of distortion correction, filtering, and enhancement. The preprocessing method may use image processing algorithms such as Gamma (Gamma) transform and histogram transform, which is not specifically limited in this embodiment of the present invention.

The image registration operation may be performed by using a method for extracting stable characteristic corners, such as a corner detection algorithm based on a Harris operator, or may be performed by using a method based on a relevant phase, which is not specifically limited in the embodiment of the present invention.

The image registration result provides a basis for the detection and identification of the underwater target. The image registration result may be directly output in an image form after image fusion, or may be output in other forms, which is not specifically limited in the embodiment of the present invention.

The detection and identification unit 103 is used for determining the detection and identification result of the underwater target based on the image registration result.

Specifically, according to the image registration result, the detection and identification unit 103 performs final joint detection and identification on the region which is determined to be the underwater target in the acoustic images with different viewing angles, and finally determines the detection and identification result of the underwater target.

According to the underwater target detection and identification device provided by the embodiment of the invention, the single-view side scan sonar with different scanning angles is respectively arranged at the plurality of different positions on each side of the underwater carrier, so that the acquisition of a plurality of sound images of the same target from different positions and different visual angles is realized, the probability of large noise interference at the plurality of positions is greatly reduced, the phenomenon that the target is shielded is avoided at a high probability, the detection rate and the identification rate of the underwater target are improved, the false alarm rate is reduced, meanwhile, the repeated detection and identification of the underwater target by one navigation path are realized, the energy efficiency ratio of the underwater carrier is improved, and the cost is reduced.

Based on the above embodiment, the detection identification unit includes a first detection identification subunit;

the first detection and identification subunit is used for carrying out target detection processing on each single-vision acoustic image to obtain an underwater target image area in each single-vision acoustic image and the corresponding confidence coefficient thereof, and determining the detection and identification result of the underwater target based on the image registration result, the underwater target image area in each single-vision acoustic image and the corresponding confidence coefficient thereof.

Specifically, when a plurality of single-view side-scan sonars are used to detect a target, the worst case that signal noise is large in the sound images acquired by the plurality of single-view side-scan sonars still needs to be considered, that is, the case that the actual target resolution of each single-view sound image is low and strong noise causes the edge information of the target feature (the bright area and the shadow area) to be blurred or lost needs to be considered. Meanwhile, the randomness of the shape of the target and the situations of being blocked and partially blocked by non-target objects need to be considered.

Therefore, it is necessary to perform target detection processing on each single-view sound image, that is, to perform segmentation and detection on an underwater target image region on each single-view sound image, so as to improve the detection rate and the recognition rate of an underwater target and reduce the false alarm rate. And when the detection of the underwater target image area in each single-view sound image is completed, the confidence coefficient of the underwater target image area is given, and a basic basis is provided for the detection and identification of the underwater target in the subsequent multi-view sound images.

The segmentation of the underwater target image region is performed on each single-view acoustic image, and algorithms such as threshold segmentation, region growing, markov, level set and the like may be adopted, which is not specifically limited in the embodiment of the present invention. The detection of the underwater target image region and the determination of the confidence coefficient may adopt algorithms such as an SVM (Support Vector Machine), AdaBoost, significance test, template matching and the like, or may use algorithms such as YOLOv3, ssd (single Shot multiple boxdetector) and fast R-CNN based on a neural network, which is not specifically limited in the embodiment of the present invention.

And establishing a probability judgment model of joint detection and identification by adopting a random forest algorithm in integrated learning according to the underwater target image area and the confidence degrees corresponding to the underwater target image area and combining the characteristics of the underwater target in the image registration result, and determining the detection and identification result of the underwater target.

According to the underwater target detection and identification device provided by the embodiment of the invention, the detection and identification result of the underwater target is determined through the first detection and identification subunit, so that the detection rate and the identification rate of the underwater target are improved, and the false alarm rate is reduced.

According to any of the above embodiments, the detection identification unit comprises a second detection identification subunit;

the second detection and identification subunit is used for inputting the image registration result into the joint detection and identification model to obtain a detection and identification result of the underwater target output by the joint detection and identification model;

the joint detection and identification model is obtained by training based on the sample image registration result and the detection and identification result of the corresponding sample underwater target.

Specifically, the image registration result is an image obtained by fusing a plurality of single-view sonograms containing the same underwater target after registration. The joint detection and identification model is trained through a large number of samples and has strong image feature extraction capability aiming at underwater targets. And inputting the image registration result into the joint detection and identification model to obtain a detection and identification result of the underwater target output by the joint detection and identification model.

The joint detection recognition model can be obtained by pre-training, and specifically can be obtained by training in the following way: firstly, collecting a large number of sample image registration results and detection and identification results of sample underwater targets corresponding to the sample image registration results, and training an initial joint detection and identification model by using the sample image registration results and the detection and identification results of the sample underwater targets corresponding to the sample image registration results. The initial joint detection recognition model can adopt a convolutional neural network, and the embodiment of the invention does not specifically limit the type and the specific structure of the initial joint detection recognition model.

The trained and optimized joint detection and recognition model can learn the image feature extraction capability aiming at the underwater target, so that the detection rate and the recognition rate of the underwater target are improved, and the false alarm rate is reduced.

According to the underwater target detection and identification device provided by the embodiment of the invention, the detection and identification result of the underwater target is determined through the second detection and identification subunit, so that the detection rate and the identification rate of the underwater target are improved, and the false alarm rate is reduced.

Based on any of the embodiments above, the registration unit is specifically configured to:

and carrying out image registration based on multi-information fusion on single-view sonograms acquired by all single-view side-scan sonars on the single side of the underwater carrier aiming at the same underwater target, wherein the multi-information fusion comprises fusion of at least two of the navigation speed of the underwater carrier, the position of the underwater carrier, the posture of the single-view side-scan sonar and the sonogram.

Specifically, the navigation speed of the underwater carrier, the position of the underwater carrier, the single-view side-scan sonar posture and the sonogram can be obtained, at least two kinds of information are adopted to carry out image registration on all the single-view sonograms on the single side of the underwater carrier, and the detection rate and the recognition rate of the underwater target are improved.

Based on any one of the above embodiments, the method further comprises a feature enhancement unit:

the characteristic enhancement unit is used for performing target characteristic enhancement on each single-view sound image by adopting a super-resolution algorithm and/or an image enhancement algorithm.

Specifically, before each single-view sound image is subjected to target detection processing, the feature enhancement unit is used for performing target feature enhancement on each single-view sound image, so that the extraction capability of underwater target features in the target detection processing is greatly improved.

The Super-Resolution algorithm is adopted to improve the detail information of the underwater target in the single-view sonogram, bilinear interpolation and bicubic interpolation can be adopted, and algorithms such as SRCNN (Super-Resolution coherent network), SRGAN (source code for generating anti-inflammatory network) and the like based on deep learning can also be adopted, and the embodiment of the invention is not limited to this specifically.

The image enhancement algorithm is adopted to realize the edge enhancement of the underwater target image region, and the image enhancement algorithm can be based on Gamma (Gamma) transformation and histogram transformation algorithms, and can also be based on algorithms of deep learning such as DLSR (cognitive left regression), EnlightGAN and the like, and the embodiment of the invention is not limited to the specific method.

The single-view sound image after the target feature enhancement can be used for the first detection and identification subunit to perform target detection processing, and can also be used for the image registration unit to perform image registration.

Based on any one of the above embodiments, each single-view side-scan sonar is provided with a backflow enclosure cover.

Specifically, consider that a plurality of single-view side-scan sonars can receive the influence of rivers resistance when actual underwater motion and lead to the unstability of gesture, consequently the device can dispose refluence outer housing cover respectively to every sonar, perhaps whole plus refluence outer housing cover to guarantee the gesture stability of single-view side-scan sonar when underwater motion.

According to the underwater target detection and identification device provided by the embodiment of the invention, the backflow shell cover is configured for each single-view side-scan sonar, so that the imaging stability of the single-view side-scan sonar and the imaging quality of the sonogram are improved.

Based on any one of the above embodiments, fig. 4 is a schematic flow chart of the underwater target detection and identification method provided by the embodiment of the present invention, as shown in fig. 4, the method includes:

step 401, performing image registration on single-view sonograms collected by all single-view side-scan sonars on a single side of an underwater carrier aiming at the same underwater target to obtain an image registration result;

step 402, determining a detection and identification result of the underwater target based on the image registration result;

wherein, each side of the underwater carrier is provided with at least two single-view side-scan sonars with different scanning directions.

Specifically, the sonograms of the same target at different times, different visual angles and different positions can be acquired through at least two single-view side-scan sonars with different scanning directions, which are arranged on each side of the underwater carrier.

In step 401, since all the single-view side-scan sonars on a single side of the underwater vehicle have severe distortion and strong background noise interference for a single-view sound image acquired by the same underwater target, accurate registration needs to be performed for the sound-view images acquired at multiple different times, different positions and different viewing angles, that is, the same region in the multiple sound-view images is subjected to registration fusion, so that the observation of the same target from multiple different positions and viewing angles is realized, the probability of large noise interference at multiple positions is reduced, and the phenomenon that the target is blocked is avoided.

Before image registration, in order to obtain a high-quality image registration result, an acoustic image preprocessing operation may be performed on each single-view acoustic image. The sonogram pre-processing operation includes at least one of distortion correction, filtering, and enhancement. The preprocessing method may use image processing algorithms such as Gamma (Gamma) transform and histogram transform, which is not specifically limited in this embodiment of the present invention.

The image registration operation may be performed by using a method for extracting stable characteristic corners, such as a corner detection algorithm based on a Harris operator, or may be performed by using a method based on a relevant phase, which is not specifically limited in the embodiment of the present invention.

The image registration result provides a basis for the detection and identification of the underwater target. The image registration result may be directly output in an image form after image fusion, or may be output in other forms, which is not specifically limited in the embodiment of the present invention.

And step 402, performing final joint detection and identification on the area which is judged to be possibly the underwater target in the two or more acoustic images with different visual angles according to the image registration result, and finally determining the detection and identification result of the underwater target.

According to the underwater target detection and identification method provided by the embodiment of the invention, the single-view side scan sonar with different scanning angles is respectively installed at the plurality of different positions of each side of the underwater carrier, so that the acquisition of a plurality of sound images of the same target from different positions and different visual angles is realized, the probability of large noise interference at the plurality of positions is greatly reduced, the phenomenon that the target is shielded is avoided at a high probability, the detection rate and the identification rate of the underwater target are improved, the false alarm rate is reduced, meanwhile, the repeated detection and identification of the underwater target by one navigation path are realized, the energy efficiency ratio of the underwater carrier is improved, and the cost is reduced.

Based on any of the above embodiments, step 402 specifically includes:

carrying out target detection processing on each single-vision sound image to obtain an underwater target image area in each single-vision sound image and a corresponding confidence coefficient thereof;

and determining the detection and identification result of the underwater target based on the image registration result, the underwater target image area in each single-visual-sound image and the corresponding confidence coefficient.

Specifically, when a plurality of single-view side-scan sonars are used to detect a target, the worst case that signal noise is large in the sound images acquired by the plurality of single-view side-scan sonars still needs to be considered, that is, the case that the actual target resolution of each single-view sound image is low and strong noise causes the edge information of the target feature (the bright area and the shadow area) to be blurred or lost needs to be considered. Meanwhile, the randomness of the shape of the target and the situations of being blocked and partially blocked by non-target objects need to be considered.

Therefore, it is necessary to perform target detection processing on each single-view sound image, that is, to perform segmentation and detection on an underwater target image region on each single-view sound image, so as to improve the detection rate and the recognition rate of an underwater target and reduce the false alarm rate. And when the detection of the underwater target image area in each single-view sound image is completed, the confidence coefficient of the underwater target image area is given, and a basic basis is provided for the detection and identification of the underwater target in the subsequent multi-view sound images.

The segmentation of the underwater target image region is performed on each single-view acoustic image, and algorithms such as threshold segmentation, region growing, markov, level set and the like may be adopted, which is not specifically limited in the embodiment of the present invention. The detection of the underwater target image region and the determination of the confidence coefficient may adopt algorithms such as an SVM (Support Vector Machine), AdaBoost, significance test, template matching and the like, or may use algorithms such as YOLOv3, ssd (single Shot multiple boxdetector) and fast R-CNN based on a neural network, which is not specifically limited in the embodiment of the present invention.

And establishing a probability judgment model of joint detection and identification by adopting a random forest algorithm in integrated learning according to the underwater target image area and the confidence degrees corresponding to the underwater target image area and combining the characteristics of the underwater target in the image registration result, and determining the detection and identification result of the underwater target.

Based on any of the above embodiments, step 402 specifically includes:

inputting the image registration result into the joint detection and identification model to obtain a detection and identification result of the underwater target output by the joint detection and identification model;

the joint detection and identification model is obtained by training based on the sample image registration result and the detection and identification result of the corresponding sample underwater target.

Specifically, the image registration result is an image obtained by fusing a plurality of single-view sonograms containing the same underwater target after registration. The joint detection and identification model is trained through a large number of samples and has strong image feature extraction capability aiming at underwater targets. And inputting the image registration result into the joint detection and identification model to obtain a detection and identification result of the underwater target output by the joint detection and identification model.

The joint detection recognition model can be obtained by pre-training, and specifically can be obtained by training in the following way: firstly, collecting a large number of sample image registration results and detection and identification results of sample underwater targets corresponding to the sample image registration results, and training an initial joint detection and identification model by using the sample image registration results and the detection and identification results of the sample underwater targets corresponding to the sample image registration results. The initial joint detection recognition model can adopt a convolutional neural network, and the embodiment of the invention does not specifically limit the type and the specific structure of the initial joint detection recognition model.

The trained and optimized joint detection and recognition model can learn the image feature extraction capability aiming at the underwater target, so that the detection rate and the recognition rate of the underwater target are improved, and the false alarm rate is reduced.

Based on any one of the above-mentioned embodiments, carry out image registration to all single-vision side scan sonars of underwater carrier unilateral to the single-vision sonogram that same underwater target was gathered, include:

and carrying out image registration based on multi-information fusion on single-view sonograms acquired by all single-view side-scan sonars on the single side of the underwater carrier aiming at the same underwater target, wherein the multi-information fusion comprises fusion of at least two of the position of the underwater carrier and the sonograms.

Specifically, the navigation speed of the underwater carrier, the position of the underwater carrier, the single-view side-scan sonar posture and the sonogram can be obtained, at least two kinds of information are adopted to carry out image registration on all the single-view sonograms on the single side of the underwater carrier, and the detection rate and the recognition rate of the underwater target are improved.

Based on any of the above embodiments, the target detection processing is performed on each single-view image to obtain an underwater target image region and a confidence corresponding to the underwater target image region, and the method further includes:

and performing target feature enhancement on each single-view sound image by adopting a super-resolution algorithm and/or an image enhancement algorithm.

Specifically, target feature enhancement is performed on each single-view sound image, and the extraction capability of underwater target features in target detection processing is greatly improved.

The Super-Resolution algorithm is adopted to improve the detail information of the underwater target in the single-view sonogram, bilinear interpolation and bicubic interpolation can be adopted, and algorithms such as SRCNN (Super-Resolution coherent network), SRGAN (source code for generating anti-inflammatory network) and the like based on deep learning can also be adopted, and the embodiment of the invention is not limited to this specifically.

The image enhancement algorithm is used for realizing the edge enhancement of the underwater target image region, and may be based on Gamma (Gamma) transformation and histogram transformation algorithms, or may also be based on algorithms of deep learning (discrete Least square regression), enlightngan and the like, which is not specifically limited in the embodiment of the present invention.

The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.

Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes commands for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments or some parts of the embodiments.

Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.