CN115995032B - Multi-mode feature fusion ship intelligent identification method and system - Google Patents
Multi-mode feature fusion ship intelligent identification method and system Download PDFInfo
- Publication number
- CN115995032B CN115995032B CN202310286289.7A CN202310286289A CN115995032B CN 115995032 B CN115995032 B CN 115995032B CN 202310286289 A CN202310286289 A CN 202310286289A CN 115995032 B CN115995032 B CN 115995032B
- Authority
- CN
- China
- Prior art keywords
- ship
- data
- carrying
- dimensional
- image data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The invention relates to a ship intelligent identification method and a system with multi-mode feature fusion, wherein the method comprises the following steps: collecting ship track data, ship one-dimensional distance image data and ship synthetic aperture radar image data; respectively carrying out multi-mode data characteristic extraction according to the ship track data, the ship one-dimensional distance image data and the ship synthetic aperture radar image data; and carrying out feature fusion on the obtained multi-mode data, and carrying out multi-classification according to a layer of fully connected neural network and a normalized exponential function. Compared with other ship identification methods with single data sources, the method can effectively comprehensively consider multi-mode data characteristics, so that the track data, one-dimensional range profile data and synthetic aperture radar data on the sea of the ship are fully utilized, and stable and reliable automatic ship identification technology can be utilized in a complex sea area environment to realize accurate positioning and real-time detection of illegal ships by maritime departments.
Description
Technical Field
The invention belongs to the field of data processing, and particularly relates to a ship intelligent identification method and system based on multi-mode feature fusion.
Background
In the ship navigation safety problem of the sea area environment, the high-efficiency ship target identification method can help a ship monitoring system to timely master ship dynamics, and meanwhile, the ship navigation risk can be analyzed. Aiming at the problem of remote target identification of the offshore ship, the information of the offshore track, the high-resolution range profile and the synthetic aperture radar image characteristic information of the target ship are important, so that the target identification is an important characteristic for realizing the target identification of the offshore ship.
The track information recorded by the ship Automatic Identification System (AIS) comprises a navigation route, a navigation speed, a navigation course and the like. The high-resolution range profile (HRRP) utilizes the radar to acquire one-dimensional projection data of electromagnetic scattering of a target ship in the radar observation direction. Synthetic Aperture Radar (SAR) images by using synthetic aperture, pulse compression mechanism and signal processing technique, and its reflected signal can reflect the electromagnetic scattering property of the target, and can realize characteristic complementation capability with the optical sensor. In a real sea area environment, as a plurality of marine interference targets and the environment is complex, the single data characteristic is difficult to achieve a very good effect on the ship identification in the sea area environment, so that the marine ship targets are required to be comprehensively identified by combining multi-mode data, and the accuracy and the stability of the ship identification are improved.
At present, most of the schemes adopt a single data source for marine ship identification aiming at ship identification research in the sea area environment, and the accuracy of ship identification is not improved by comprehensively utilizing multi-mode data. The prior art methods all adopt independent feature extraction methods, and lack of an effective method can realize simultaneous feature extraction and fusion of track data, one-dimensional range profile data and synthetic aperture radar images.
Therefore, how to realize comprehensive identification of marine ship targets through multi-mode data is a technical problem to be solved urgently by those skilled in the art.
Disclosure of Invention
In order to solve the problems of multiple offshore interference targets, complex sea area environment and lower accuracy of ship identification by a single data source, the invention provides a ship intelligent identification method and system with multi-mode feature fusion, wherein the method comprises the following steps:
step S101: collecting ship track data (AIS data), ship one-dimensional range profile data (HRRP data) and ship synthetic aperture radar image data (SAR data);
step S102: respectively carrying out multi-mode data characteristic extraction according to the ship track data, the ship one-dimensional distance image data and the ship synthetic aperture radar image data;
step S103: and carrying out feature fusion on the obtained multi-mode data, and carrying out multi-classification according to a layer of fully connected neural network and a normalized exponential function (Softmax function).
In some embodiments, the step S102 includes: extracting the characteristics of the ship track data comprises the following steps: geographic features, behavioral features, and trajectory features of the vessel; wherein the geographic feature comprises location information of the vessel; the behavior characteristics comprise the heading and the navigational speed of the ship; the trajectory features include: average speed, average acceleration, average steering rate, maximum steering rate, minimum steering rate, maximum heading, minimum heading, and heading variance of the vessel.
In some embodiments, the step S102 includes: the ship one-dimensional range profile data is converted through one-dimensional random convolution check, so that the data context semantic structure can be reserved, and the consistency of feature dimensions is ensured.
In some embodiments, the step S102 includes: and carrying out gray preprocessing on the ship synthetic aperture radar image data, inhibiting noise contained in the image data, and extracting features in the data through a convolutional neural network model.
In some specific embodiments, the step S103 includes: splicing the extracted features each time through a feature extraction process of multi-mode data, taking the fused features as the input of a next convolutional neural network, and obtaining one-dimensional vectors from the ship track data through multiple statistics
The ship one-dimensional range profile is subjected to one-dimensional convolution to obtain a one-dimensional vector +.>
The ship synthetic aperture radar is subjected to one-dimensional eigenvector through a layer of convolutional neural network>
Then multi-classification is carried out according to a layer of fully connected neural network and the normalized exponential function,
the expression is:
the normalized exponential function is assisted by the cross entropy loss function to carry out multi-classification, and the expression is as follows: />
Wherein (1)>
Representing the number of ship types>
Representing the corresponding real sample tag, if sample list +.>
Is->
Then
The method comprises the steps of carrying out a first treatment on the surface of the Otherwise 0 +.>
Representing the class of the output.
The invention provides a multi-mode feature fusion based intelligent ship identification system, which comprises:
and a data input module: collecting and inputting ship track data, ship one-dimensional distance image data and ship synthetic aperture radar image data;
and the feature extraction module is used for: respectively carrying out multi-mode data characteristic extraction according to the ship track data, the ship one-dimensional distance image data and the ship synthetic aperture radar image data;
and a fusion classification module: and carrying out feature fusion on the obtained multi-mode data, and carrying out multi-classification according to a layer of fully connected neural network and a normalized exponential function.
In some of these specific embodiments, the feature extraction module includes: extracting the characteristics of the ship track data comprises the following steps: geographic features, behavioral features, and trajectory features of the vessel; wherein the geographic feature comprises location information of the vessel; the behavior characteristics comprise the heading and the navigational speed of the ship; the trajectory features include: average speed, average acceleration, average steering rate, maximum steering rate, minimum steering rate, maximum heading, minimum heading, and heading variance of the vessel.
In some of these specific embodiments, the feature extraction module includes: the ship one-dimensional range profile data is converted through one-dimensional random convolution check, so that the data context semantic structure can be reserved, and the consistency of feature dimensions is ensured.
In some of these specific embodiments, the feature extraction module includes: and carrying out gray preprocessing on the ship synthetic aperture radar image data, inhibiting noise contained in the image data, and extracting features in the data through a convolutional neural network model.
In some embodiments, the fusion classification module includes: splicing the extracted features each time through a feature extraction process of multi-mode data, taking the fused features as the input of a next convolutional neural network, and obtaining one-dimensional vectors from the ship track data through multiple statistics
The ship one-dimensional range profile is subjected to one-dimensional convolution to obtain a one-dimensional vector +.>
The ship synthetic aperture radar is subjected to one-dimensional eigenvector through a layer of convolutional neural network>
Then, multi-classification is carried out according to a layer of fully connected neural network and a normalized exponential function, and the parts are added>
The expression is:
the normalized exponential function is assisted by the cross entropy loss function to carry out multi-classification, and the expression is as follows: />
Wherein (1)>
Representing the number of ship types>
Representing the corresponding real sample tag, if sample list +.>
Is->
Then
The method comprises the steps of carrying out a first treatment on the surface of the Otherwise 0 +.>
Representing the class of the output.
The invention has the beneficial effects that:
compared with the ship identification method of other single data sources, the intelligent ship identification method and system of the multi-mode feature fusion can effectively comprehensively consider multi-mode data features, so that the track data, the one-dimensional range profile data and the synthetic aperture radar data on the sea of the ship are fully utilized, the accurate positioning and real-time detection of illegal ships by maritime departments can be realized by utilizing a stable and reliable automatic ship identification technology in a complex sea area environment, scientific early warning and risk analysis can be carried out on external invasion targets according to the confidence level of the ship identification technology, and the judgment capability and early warning capability of the maritime departments are enhanced. The intelligent ship identification technology not only can reduce economic loss and casualties caused by marine accidents, but also has military strategic significance of accurately tracking ship targets and realizing accurate guidance.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly introduced below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of some embodiments of a multi-modal feature fusion method for intelligent identification of ships according to the present invention;
FIG. 2 is a diagram of a model architecture of a multi-modal feature fusion method for intelligent identification of ships in the present invention;
FIG. 3 is a schematic structural diagram of some embodiments of a multi-modal feature fusion intelligent ship identification system of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments.
Examples of the embodiments are illustrated in the accompanying drawings, wherein like or similar symbols indicate like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative and intended to explain the present invention and should not be construed as limiting the invention.
In the description of the present invention, it should be understood that the terms "top," "bottom," "inner," "outer," "axis," "circumferential," and the like indicate an orientation or a positional relationship based on that shown in the drawings, and are merely for convenience in describing the present invention or simplifying the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention.
Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
In the present invention, unless explicitly specified and limited otherwise, the terms "mounted," "connected," "secured," "engaged," "hinged," and the like are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally formed; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communicated with the inside of two elements or the interaction relationship of the two elements. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances.
Referring to fig. 1 and 2, a method for intelligently identifying a ship by multi-mode feature fusion includes:
step S101: the method comprises the steps of collecting ship track data (AIS data), ship one-dimensional range profile data (HRRP data) and ship synthetic aperture radar image data (SAR data).
Specifically, the ship track data comprises geographic features, behavioral features and track features of the ship; wherein the geographic feature comprises location information of the vessel; the behavior characteristics comprise the heading and the navigational speed of the ship; the trajectory features include: average speed, average acceleration, average steering rate, maximum steering rate, minimum steering rate, maximum heading, minimum heading, and heading variance of the vessel. The ship one-dimensional range profile data not only provides the geometric shape and structural characteristics of the ship, but also contains more relevant information required by target identification. The ship synthetic aperture radar image data utilizes the synthetic aperture principle to carry out coherent processing on echoes received at different positions, thereby realizing self-high-resolution microwave imaging.
Step S102: and respectively carrying out multi-mode data characteristic extraction according to the ship track data, the ship one-dimensional distance image data and the ship synthetic aperture radar image data.
In some embodiments of the present invention, extracting the characteristics of the ship track data includes: geographic features, behavioral features, and trajectory features of the vessel; wherein the geographic feature comprises location information of the vessel; the behavior characteristics comprise the heading and the navigational speed of the ship; the trajectory features include: average speed, average acceleration, average steering rate, maximum steering rate, minimum steering rate, maximum heading, minimum heading, and heading variance of the vessel.
Specifically, the invention provides a statistical feature combination to represent the track feature of the ship, namely, a plurality of statistical features, in particular, average speed, average navigational speed and average acceleration of the ship are utilized; average steering rate, maximum steering rate, minimum steering rate; maximum heading, minimum heading, and heading variance; these three sets of data represent the trajectory characteristics of the vessel.
In some embodiments of the invention, the ship one-dimensional range profile data is converted through one-dimensional random convolution check, so that the data context semantic structure can be reserved, and the consistency of feature dimensions is ensured.
In some embodiments of the present invention, the ship synthetic aperture radar image data is subjected to gray preprocessing, noise contained in the image data is suppressed, and features in the data are extracted through a convolutional neural network model.
Step S103: and carrying out feature fusion on the obtained multi-mode data, and carrying out multi-classification according to a layer of fully connected neural network and a normalized exponential function (Softmax function).
In some embodiments of the present invention, features extracted each time are spliced through a feature extraction process of multi-modal data, the fused features are used as input of a next convolutional neural network, and the ship track data is subjected to multiple statistics to obtain a one-dimensional vector
The ship one-dimensional range profile is subjected to one-dimensional convolution to obtain a one-dimensional vector
The ship synthetic aperture radar is subjected to one-dimensional eigenvector through a layer of convolutional neural network
Then multi-classification is carried out according to a layer of fully connected neural network and the normalized exponential function,
the expression is:
the normalized exponential function is assisted by the cross entropy loss function to carry out multi-classification, and the expression is as follows: />
Wherein (1)>
Representing the number of ship types>
Representing the corresponding real sample tag, if sample list +.>
Is->
Then
The method comprises the steps of carrying out a first treatment on the surface of the Otherwise 0 +.>
Representing the class of the output.
Referring to fig. 3, a multi-modal feature fusion ship intelligent recognition system includes:
data input module 10: collecting and inputting ship track data, ship one-dimensional distance image data and ship synthetic aperture radar image data;
feature extraction module 20: respectively carrying out multi-mode data characteristic extraction according to the ship track data, the ship one-dimensional distance image data and the ship synthetic aperture radar image data;
fusion classification module 30: and carrying out feature fusion on the obtained multi-mode data, and carrying out multi-classification according to a layer of fully connected neural network and a normalized exponential function (Softmax function).
In some embodiments of the present invention, the data input module 10 includes: and sequentially inputting the acquired ship track data, the ship one-dimensional distance image data and the ship synthetic aperture radar image data into the system.
In some embodiments of the present invention, the feature extraction module 20 includes: extracting the characteristics of the ship track data comprises the following steps: geographic features, behavioral features, and trajectory features of the vessel; wherein the geographic feature comprises location information of the vessel; the behavior characteristics comprise the heading and the navigational speed of the ship; the trajectory features include: average speed, average acceleration, average steering rate, maximum steering rate, minimum steering rate, maximum heading, minimum heading, and heading variance of the vessel.
In some embodiments of the present invention, the feature extraction module 20 includes: the ship one-dimensional range profile data is converted through one-dimensional random convolution check, so that the data context semantic structure can be reserved, and the consistency of feature dimensions is ensured.
In some embodiments of the present invention, the feature extraction module 20 includes: and carrying out gray preprocessing on the ship synthetic aperture radar image data, inhibiting noise contained in the image data, and extracting features in the data through a convolutional neural network model.
In some embodiments of the present invention, the fusion classification module 30 includes: splicing the extracted features each time through a feature extraction process of multi-mode data, taking the fused features as the input of a next convolutional neural network, and obtaining one-dimensional vectors from the ship track data through multiple statistics
The ship one-dimensional range profile is subjected to one-dimensional convolution to obtain a one-dimensional vector +.>
The ship synthetic aperture radar is subjected to one-dimensional eigenvector through a layer of convolutional neural network>
Then multi-classification is carried out according to a layer of fully connected neural network and the normalized exponential function,
the expression is:
the normalized exponential function is assisted by the cross entropy loss function to carry out multi-classification, and the expression is as follows: />
Wherein (1)>
Representing the number of ship types>
Representing the corresponding real sample tag, if sample list +.>
Is->
Then
The method comprises the steps of carrying out a first treatment on the surface of the Otherwise 0 +.>
Representing the class of the output.
Aiming at the problem of single data source identification accuracy of ships in the sea area environment, the invention provides an effective multi-mode feature extraction and fusion method to improve the identification accuracy. Aiming at the problem that the lengths of ship track data are not uniform, the invention provides an effective statistical feature for identification, and has good interpretation. And the characteristic extraction is carried out on the ship one-dimensional range profile and the SAR profile by using an efficient neural network model architecture, so that the subsequent characteristic level fusion effect is convenient to improve.
In the description of the present specification, reference to the terms "one embodiment," "some embodiments," "examples," "particular examples," "one particular embodiment," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, schematic representations of terms do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The present invention is not limited to the above preferred embodiments, and any person skilled in the art, within the scope of the present invention, may apply to the present invention, and equivalents and modifications thereof are intended to be included in the scope of the present invention.
Claims (6)
1. The intelligent ship identification method based on multi-mode feature fusion is characterized by comprising the following steps of:
step S101: collecting ship track data, ship one-dimensional distance image data and ship synthetic aperture radar image data;
step S102: respectively carrying out multi-mode data feature extraction according to the ship track data, the ship one-dimensional distance image data and the ship synthetic aperture radar image data, wherein the feature of the ship track data comprises geographic features, behavior features and track features of a ship; the geographic features include location information of the vessel; the behavior characteristics comprise the heading and the navigational speed of the ship; the trajectory features include: average speed, average navigational speed, average acceleration, average steering rate, maximum steering rate, minimum steering rate, maximum heading, minimum heading and heading variance of the vessel;
step S103: the method comprises the steps of carrying out feature fusion on the extracted multi-modal data, carrying out multi-classification according to a layer of fully-connected neural network and a normalized index function, splicing the extracted features each time through a feature extraction process of the multi-modal data, taking the fused features as the input of a next convolutional neural network, carrying out multi-statistics on ship track data to obtain a one-dimensional vector Feat_, carrying out one-dimensional convolution on the ship one-dimensional range profile to obtain a one-dimensional vector Feat_, carrying out one-layer convolution on the ship one-dimensional range profile to obtain a one-dimensional feature vector Feat_ SAR, and carrying out multi-classification according to a layer of fully-connected neural network and the normalized index function, wherein the expression is as follows:
f＝"softmax"("FULL_CONNECT"(Feat_AIS⊕Feat_HRRP⊕Feat_SAR)
and performing multi-classification by using the cross entropy loss function to assist the normalized exponential function, wherein the expression is as follows:
wherein K represents the number of ship types, y represents the corresponding real sample tag, and if the sample list y (i) is i, y (i) =1; otherwise, 0, f (i) represents the class of output.
2. The intelligent recognition method for the ship with the multi-modal feature fusion according to claim 1, wherein the step S102 includes: and converting the ship one-dimensional range profile data through one-dimensional random convolution check, reserving a data context semantic structure, and ensuring consistency of feature dimensions.
3. The intelligent recognition method for the ship with the multi-modal feature fusion according to claim 1, wherein the step S102 includes: and carrying out gray preprocessing on the ship synthetic aperture radar image data, inhibiting noise contained in the image data, and extracting features in the data through a convolutional neural network model.
4. The utility model provides a warship intelligent identification system of multimode characteristic fusion which characterized in that includes:
and a data input module: collecting and inputting ship track data, ship one-dimensional distance image data and ship synthetic aperture radar image data;
and the feature extraction module is used for: respectively carrying out multi-mode data feature extraction according to the ship track data, the ship one-dimensional distance image data and the ship synthetic aperture radar image data, wherein the feature of the ship track data comprises geographic features, behavior features and track features of a ship; the geographic features include location information of the vessel; the behavior characteristics comprise the heading and the navigational speed of the ship; the trajectory features include: average speed, average navigational speed, average acceleration, average steering rate, maximum steering rate, minimum steering rate, maximum heading, minimum heading and heading variance of the vessel;
and a fusion classification module: the method comprises the steps of carrying out feature fusion on the extracted multi-modal data, carrying out multi-classification according to a layer of fully-connected neural network and a normalized index function, splicing the extracted features each time through a feature extraction process of the multi-modal data, taking the fused features as the input of the next convolutional neural network, carrying out multi-statistics on ship track data to obtain a one-dimensional vector Feat_AIS, carrying out one-dimensional convolution on the ship one-dimensional range profile to obtain a one-dimensional vector Feat_HRRP, carrying out multi-classification on the ship synthetic aperture radar according to a layer of fully-connected neural network and the normalized index function, and carrying out multi-classification according to the one-layer of fully-connected neural network and the normalized index function, wherein the expression is as follows:
f＝"softmax"("FULL_CONNECT"(Feat_AIS⊕Feat_HRRP⊕Feat_SAR)
and performing multi-classification by using the cross entropy loss function to assist the normalized exponential function, wherein the expression is as follows:
wherein K represents the number of ship types, y represents the corresponding real sample tag, and if the sample list y (i) is i, y (i) =1; otherwise, 0, f (i) represents the class of output.
5. The multi-modal feature fusion ship intelligent recognition system of claim 4, wherein the feature extraction module is further configured to: the ship one-dimensional range profile data is converted through one-dimensional random convolution check, so that the data context semantic structure can be reserved, and the consistency of feature dimensions is ensured.
6. The multi-modal feature fusion ship intelligent recognition system of claim 4, wherein the feature extraction module is further configured to: and carrying out gray preprocessing on the ship synthetic aperture radar image data, inhibiting noise contained in the image data, and extracting features in the data through a convolutional neural network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310286289.7A CN115995032B (en) | 2023-03-23 | 2023-03-23 | Multi-mode feature fusion ship intelligent identification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310286289.7A CN115995032B (en) | 2023-03-23 | 2023-03-23 | Multi-mode feature fusion ship intelligent identification method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115995032A CN115995032A (en) | 2023-04-21 |
| CN115995032B true CN115995032B (en) | 2023-06-06 |
Family
ID=85993824
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310286289.7A Active CN115995032B (en) | 2023-03-23 | 2023-03-23 | Multi-mode feature fusion ship intelligent identification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115995032B (en) |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109460740B (en) * | 2018-11-15 | 2020-08-11 | 上海埃威航空电子有限公司 | Ship identity recognition method based on AIS and video data fusion |
| CN111025246B (en) * | 2019-11-28 | 2021-09-07 | 北京遥测技术研究所 | Simulation system and method for composite scene imaging of sea surface and ship by using stationary orbit SAR |
| CN111027459A (en) * | 2019-12-06 | 2020-04-17 | 江苏海事职业技术学院 | Ship track prediction method and system |
| CN112163461B (en) * | 2020-09-07 | 2022-07-05 | 中国海洋大学 | Underwater target identification method based on multi-mode fusion |
| CN112132012B (en) * | 2020-09-22 | 2022-04-26 | 中国科学院空天信息创新研究院 | High-resolution SAR ship image generation method based on generation countermeasure network |
| CN112308019B (en) * | 2020-11-19 | 2021-08-17 | 中国人民解放军国防科技大学 | SAR ship target detection method based on network pruning and knowledge distillation |
| CN114594440A (en) * | 2022-01-21 | 2022-06-07 | 中国人民解放军战略支援部队信息工程大学 | Radar high-resolution one-dimensional range profile target identification method and system based on double parallel networks |
| CN114898222B (en) * | 2022-04-21 | 2024-01-02 | 中国人民解放军91977部队 | Ship target track identification method and device |
| CN115542317A (en) * | 2022-09-30 | 2022-12-30 | 中国人民解放军海军航空大学 | Sea surface target comprehensive identification method and system based on multiple information sources |
-
2023
- 2023-03-23 CN CN202310286289.7A patent/CN115995032B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115995032A (en) | 2023-04-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Thombre et al. | Sensors and AI techniques for situational awareness in autonomous ships: A review | |
| US10706285B2 (en) | Automatic ship tracking method and system based on deep learning network and mean shift | |
| CN110175186B (en) | Intelligent ship environment threat target sensing system and method | |
| Neves et al. | Rotated object detection with forward-looking sonar in underwater applications | |
| CN110363115B (en) | AIS (automatic identification system) track data based ship operation abnormity semi-supervised real-time detection method | |
| CN111899568B (en) | Bridge anti-collision early warning system, method and device and storage medium | |
| Zacchini et al. | Forward-looking sonar CNN-based automatic target recognition: an experimental campaign with FeelHippo AUV | |
| Gao et al. | Improved YOLOv4 based on attention mechanism for ship detection in SAR images | |
| CN116027324B (en) | Fall detection method and device based on millimeter wave radar and millimeter wave radar equipment | |
| Zhang et al. | A object detection and tracking method for security in intelligence of unmanned surface vehicles | |
| Molina-Molina et al. | Autonomous marine robot based on AI recognition for permanent surveillance in marine protected areas | |
| CN104648627B (en) | Early warning method for ship | |
| Gong et al. | Small ship detection of SAR images based on optimized feature pyramid and sample augmentation | |
| Reggiannini et al. | Remote sensing for maritime prompt monitoring | |
| Kot | Review of obstacle detection systems for collision avoidance of autonomous underwater vehicles tested in a real environment | |
| Sun et al. | Image target detection algorithm compression and pruning based on neural network | |
| CN114740441A (en) | Low-slow small-target radar echo identification method based on small samples | |
| Wang et al. | Marine vessel detection dataset and benchmark for unmanned surface vehicles | |
| CN117647812A (en) | Navigation radar-based ship travel route detection and early warning method and system | |
| CN115995032B (en) | Multi-mode feature fusion ship intelligent identification method and system | |
| CN110751077B (en) | Optical remote sensing picture ship detection method based on component matching and distance constraint | |
| Balasooriya et al. | In-situ sea ice detection using DeepLabv3 semantic segmentation | |
| Helgesen | Sensor fusion for detection and tracking of maritime vessels | |
| CN114898222B (en) | Ship target track identification method and device | |
| Bhaganagar et al. | A Novel Machine-Learning Framework With a Moving Platform for Maritime Drift Calculations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |