CN120564022A - Underwater target recognition method and system - Google Patents
Underwater target recognition method and systemInfo
- Publication number
- CN120564022A CN120564022A CN202511071807.9A CN202511071807A CN120564022A CN 120564022 A CN120564022 A CN 120564022A CN 202511071807 A CN202511071807 A CN 202511071807A CN 120564022 A CN120564022 A CN 120564022A
- Authority
- CN
- China
- Prior art keywords
- image
- water wave
- target
- underwater
- track
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/05—Underwater scenes
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
- G06V10/14—Optical characteristics of the device performing the acquisition or on the illumination arrangements
- G06V10/143—Sensing or illuminating at different wavelengths
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides an underwater target identification method and system, comprising the steps of acquiring an underwater image to be identified of a target area and an area environment image; the method comprises the steps of identifying an area environment image to obtain area information, dividing the area environment image into a plurality of sub-images, carrying out enhancement processing, obtaining flow field data of a target area, carrying out simulation through fluid mechanics to obtain a predicted moving track, carrying out comparison analysis according to each sub-image after the enhancement processing and an underwater image to be identified to obtain a preliminary comparison result, and correcting the preliminary comparison result according to the predicted moving track and a static water wave feature map to obtain a target identification result. Compared with the prior art, the method and the device consider the technical scheme except the water body, and effectively improve the accuracy of target identification by acquiring the flow field data of the target area, utilizing the hydrodynamic simulation to predict the moving track so as to correct the comparison result, and considering the disturbance factors of the underwater flow field.
Description
Technical Field
The invention relates to the field of image recognition, in particular to an underwater target recognition method and system.
Background
The underwater target identification has important significance for ocean exploration and underwater resource development. At present, the existing underwater target identification method is always worry when facing complex environments, especially in the aspects of adapting to dynamic changes and coping with multiple interference, the problem of insufficient comprehensive consideration of environmental factors generally exists, and therefore the identification precision and stability are difficult to meet the actual demands.
In some prior arts, in order to reduce the influence of environmental factors on the accuracy of recognition, the sensor network is mainly used to directly collect the illumination data in the environment, and then the collected result is used as reference data to process the collected image to be recognized.
Disclosure of Invention
The invention provides an underwater target identification method and an underwater target identification system, which aim to solve the technical problem of how to improve the identification accuracy.
In order to solve the above technical problems, the present invention provides an underwater target recognition method, including:
Acquiring an underwater image to be identified of a target area and an area environment image in real time, wherein the area environment image is acquired through a multispectral camera and a laser radar, and the multispectral camera adopts wave bands including visible light, infrared spectrum and ultraviolet spectrum;
The method comprises the steps of identifying an area environment image to obtain area information, dividing the area environment image into a plurality of sub-images, respectively carrying out enhancement processing on each sub-image according to the area information, acquiring flow field data of a target area in real time, obtaining a predicted moving track of an object to be identified through fluid mechanics simulation, extracting static water wave characteristics based on the area environment image by adopting a target detection algorithm, and generating a static water wave characteristic image;
performing contrast analysis according to each sub-image after enhancement treatment and the underwater image to be identified to obtain a preliminary contrast result;
And correcting the preliminary comparison result by utilizing the water wave type label and the predicted movement track to obtain a target recognition result of the object to be recognized.
As a preferred solution, the performing a comparison analysis according to each sub-image after the enhancement processing and the underwater image to be identified to obtain a preliminary comparison result includes:
dividing the underwater image to be identified according to each sub-image after the enhancement processing and the underwater image to be identified, and separating out a target signal and background noise;
Performing feature recognition on the target signal to obtain a signal feature set;
Updating the signal feature set in real time by using a dynamic mapping framework to obtain target feature description, wherein the dynamic mapping framework is used for describing association mapping between environment parameters and signal features;
and obtaining the preliminary comparison result based on the target feature description.
Preferably, the segmenting the underwater image to be identified, separating out a target signal and background noise, includes:
constructing a gray level co-occurrence matrix of the underwater image to be identified, wherein the gray level co-occurrence matrix is used for reflecting texture characteristics of the underwater image to be identified;
decomposing the gray level co-occurrence matrix into a plurality of signal components by utilizing independent component analysis to obtain a decomposition result;
Reconstructing texture features of the underwater image to be identified according to the energy ratio of each signal component in the decomposition result to obtain a reconstruction result;
the signal intensity of each pixel point in the reconstruction result is determined, and the pixel points with the signal intensity larger than a first preset intensity threshold value are screened out from the reconstruction result to obtain an effective pixel point combination;
and determining a target signal according to the effective pixel point combination, and determining the residual pixel points of the underwater image to be identified as the background noise.
Preferably, before the updating the signal feature set in real time by using the dynamic mapping framework, the method includes:
Acquiring a sample image, carrying out feature recognition on the sample image to obtain sample features, and acquiring environmental parameters of a corresponding area of the sample image, wherein the environmental parameters comprise suspended particle information and illumination change information;
screening out significant features from the sample features according to the signal intensity of the sample image, and constructing a dynamic mapping relation between the significant features and the suspended particle information and the illumination change information by utilizing a data table;
And obtaining the dynamic mapping frame according to the dynamic mapping relation.
The method for constructing dynamic mapping relation between the salient features, the suspended particle information and the illumination change information by utilizing a data table comprises the following steps:
constructing a linear regression model according to the turbidity and illumination change information of the water body;
Obtaining a correlation coefficient based on the linear regression model and the image contrast;
recording model parameters of the linear regression model, the image contrast and the correlation coefficient by using the data table to obtain a recording result;
and recording the recording result as the dynamic mapping relation.
As a preferred solution, the comprehensively analyzing according to the predicted moving track, the preset hydrodynamic condition and the static water ripple feature map to obtain a water ripple type label includes:
acquiring a water wave dynamic video in the same time period according to the time information of the flow field data of the target area, and splitting the water wave dynamic video into a water wave image sequence;
Setting feature conditions based on feature similarity according to the static water wave feature map, and screening a plurality of water wave images from the water wave image sequence according to the feature conditions to obtain screening results;
And deducing based on the screening result, the predicted moving track and preset hydrodynamic conditions to obtain a plurality of water wave class labels.
Preferably, the water wave image of the screening result comprises an interfering object; the deduction is performed based on the screening result, the predicted moving track and the preset hydrodynamic condition to obtain a plurality of water mark class labels, which comprises the following steps:
extracting contour points of each interference object in the water wave image of the screening result;
Cutting off water wave lines in each water wave image containing the interference object in the screening result by taking the outline points of the interference object as references, extracting characteristic points of the water wave lines, and connecting the characteristic points to obtain a water wave movement track;
Invoking a pre-trained water wave prediction model based on the water wave movement track to obtain a water wave prediction track;
and carrying out deduction on the basis of preset hydrodynamic conditions according to the water wave prediction track and the prediction movement track to obtain a plurality of water wave movement modes, and determining the water wave movement modes as water wave type labels.
As a preferred solution, the acquiring flow field data of the target area in real time, and obtaining a predicted movement track of the object to be identified through fluid mechanics simulation includes:
Acquiring flow field data of the target area through an underwater sensor network, and determining an initial track according to the flow field data, wherein the flow field data comprises flow velocity information and flow direction information of a flow field;
analyzing the flow velocity information and the flow direction information in real time to obtain the disturbance frequency and the disturbance amplitude of the flow field;
Decomposing the disturbance frequency and the disturbance amplitude of the flow field through Fourier transformation, and determining disturbance characteristic frequency and amplitude fluctuation range;
determining track adjustment parameters according to the disturbance characteristic frequency and the amplitude fluctuation range;
and predicting the next moment of the initial track by utilizing the track adjustment parameters based on the fluid model to obtain the predicted moving track of the object to be identified.
As a preferred solution, the correcting the preliminary comparison result by using the water wave category label and the predicted movement track to obtain the target recognition result of the object to be recognized includes:
Acquiring a plurality of distinguishing pixel characteristics of each sub-image after enhancement processing and the underwater image to be identified from the preliminary comparison result, wherein each distinguishing pixel characteristic comprises a plurality of pixel points;
Analyzing the predicted moving track to obtain a deviation value between the predicted moving track and the reference track;
selecting track points with deviation values larger than a preset deviation threshold value from the predicted moving track;
And correcting the preliminary comparison result according to the superposition condition of the track points and the distinguishing pixel characteristics to obtain a target identification result.
Correspondingly, the invention also provides an underwater target recognition system which comprises an image acquisition module, an area information processing module, a comparison module and a target recognition module, wherein,
The image acquisition module is used for acquiring an underwater image to be identified of a target area and an area environment image in real time, wherein the area environment image is acquired through a multispectral camera and a laser radar, and wave bands adopted by the multispectral camera comprise visible light, infrared spectrum and ultraviolet spectrum;
The regional information processing module is used for identifying the regional environment image to obtain regional information, dividing the regional environment image into a plurality of sub-images, and respectively carrying out enhancement processing on each sub-image according to the regional information;
The processing module is used for carrying out contrast analysis according to each sub-image after the enhancement processing and the underwater image to be identified to obtain a preliminary contrast result;
the target recognition module is used for comprehensively analyzing according to the predicted moving track, the preset hydrodynamic condition and the static water wave feature map to obtain a water wave type label, and correcting the preliminary comparison result by utilizing the water wave type label and the predicted moving track to obtain a target recognition result of the object to be recognized.
As a preferred solution, the target recognition module performs a comparison analysis according to each sub-image after the enhancement processing and the underwater image to be recognized, to obtain a preliminary comparison result, including:
The target recognition module divides the underwater image to be recognized according to each sub-image after the enhancement processing and the underwater image to be recognized, and separates out a target signal and background noise;
Performing feature recognition on the target signal to obtain a signal feature set;
Updating the signal feature set in real time by using a dynamic mapping framework to obtain target feature description, wherein the dynamic mapping framework is used for describing association mapping between environment parameters and signal features;
and obtaining the preliminary comparison result based on the target feature description.
As a preferred solution, the target recognition module segments the underwater image to be recognized, separates a target signal and background noise, and includes:
The target recognition module constructs a gray level co-occurrence matrix of the underwater image to be recognized, wherein the gray level co-occurrence matrix is used for reflecting texture features of the underwater image to be recognized;
decomposing the gray level co-occurrence matrix into a plurality of signal components by utilizing independent component analysis to obtain a decomposition result;
Reconstructing texture features of the underwater image to be identified according to the energy ratio of each signal component in the decomposition result to obtain a reconstruction result;
the signal intensity of each pixel point in the reconstruction result is determined, and the pixel points with the signal intensity larger than a first preset intensity threshold value are screened out from the reconstruction result to obtain an effective pixel point combination;
and determining a target signal according to the effective pixel point combination, and determining the residual pixel points of the underwater image to be identified as the background noise.
Preferably, the underwater target recognition system further comprises a mapping frame construction module, wherein the mapping frame construction module is used for updating the signal characteristic set in real time before the dynamic mapping frame is utilized:
Acquiring a sample image, carrying out feature recognition on the sample image to obtain sample features, and acquiring environmental parameters of a corresponding area of the sample image, wherein the environmental parameters comprise suspended particle information and illumination change information;
screening out significant features from the sample features according to the signal intensity of the sample image, and constructing a dynamic mapping relation between the significant features and the suspended particle information and the illumination change information by utilizing a data table;
And obtaining the dynamic mapping frame according to the dynamic mapping relation.
The mapping frame construction module utilizes a data table to construct a dynamic mapping relation between the salient features, the suspended particle information and the illumination change information, and the mapping frame construction module comprises:
The mapping frame construction module constructs a linear regression model according to the turbidity and illumination change information of the water body;
Obtaining a correlation coefficient based on the linear regression model and the image contrast;
recording model parameters of the linear regression model, the image contrast and the correlation coefficient by using the data table to obtain a recording result;
and recording the recording result as the dynamic mapping relation.
As a preferred solution, the target recognition module performs comprehensive analysis according to the predicted movement track, a preset hydrodynamic condition and the static water wave feature map to obtain a water wave class label, which includes:
acquiring a water wave dynamic video in the same time period according to the time information of the flow field data of the target area, and splitting the water wave dynamic video into a water wave image sequence;
Setting feature conditions based on feature similarity according to the static water wave feature map, and screening a plurality of water wave images from the water wave image sequence according to the feature conditions to obtain screening results;
And deducing based on the screening result, the predicted moving track and preset hydrodynamic conditions to obtain a plurality of water wave class labels.
The target recognition module derives a plurality of water wave category labels based on the screening result, the predicted moving track and a preset hydrodynamic condition, and comprises:
the target recognition module extracts contour points of each interference object in the water wave image of the screening result;
Cutting off water wave lines in each water wave image containing the interference object in the screening result by taking the outline points of the interference object as references, extracting characteristic points of the water wave lines, and connecting the characteristic points to obtain a water wave movement track;
Invoking a pre-trained water wave prediction model based on the water wave movement track to obtain a water wave prediction track;
and carrying out deduction on the basis of preset hydrodynamic conditions according to the water wave prediction track and the prediction movement track to obtain a plurality of water wave movement modes, and determining the water wave movement modes as water wave type labels.
As a preferred solution, the area information processing module acquires flow field data of the target area in real time, and obtains a predicted moving track of the object to be identified through fluid mechanics simulation, including:
the area information processing module acquires flow field data of the target area through an underwater sensor network and determines an initial track according to the flow field data, wherein the flow field data comprises flow velocity information and flow direction information of a flow field;
analyzing the flow velocity information and the flow direction information in real time to obtain the disturbance frequency and the disturbance amplitude of the flow field;
Decomposing the disturbance frequency and the disturbance amplitude of the flow field through Fourier transformation, and determining disturbance characteristic frequency and amplitude fluctuation range;
determining track adjustment parameters according to the disturbance characteristic frequency and the amplitude fluctuation range;
and predicting the next moment of the initial track by utilizing the track adjustment parameters based on the fluid model to obtain the predicted moving track of the object to be identified.
As a preferred solution, the target recognition module corrects the preliminary comparison result by using a water wave category label and a predicted movement track to obtain a target recognition result of the object to be recognized, and the method includes:
The target recognition module acquires a plurality of distinguishing pixel characteristics of each sub-image after the enhancement processing and the underwater image to be recognized from the preliminary comparison result, wherein each distinguishing pixel characteristic comprises a plurality of pixel points;
Analyzing the predicted moving track to obtain a deviation value between the predicted moving track and the reference track;
selecting track points with deviation values larger than a preset deviation threshold value from the predicted moving track;
And correcting the preliminary comparison result according to the superposition condition of the track points and the distinguishing pixel characteristics to obtain a target identification result.
Compared with the prior art, the invention has the following beneficial effects:
The invention provides an underwater target identification method and system, the method comprises the steps of acquiring an underwater image to be identified of a target area and an area environment image in real time, acquiring the area environment image through a multispectral camera and a laser radar, wherein wave bands adopted by the multispectral camera comprise visible light, infrared spectrum and ultraviolet spectrum, identifying the area environment image to obtain area information, dividing the area environment image into a plurality of sub-images, respectively carrying out enhancement processing on the sub-images according to the area information, acquiring flow field data of the target area in real time, obtaining a predicted moving track of an object to be identified through hydrodynamic simulation, adopting a target detection algorithm, extracting static water wave characteristics based on the area environment image to generate a static water wave characteristic image, carrying out comparison analysis according to the sub-images after enhancement processing and the underwater image to be identified to obtain a preliminary comparison result, carrying out comprehensive analysis according to the predicted moving track, a preset hydrodynamic condition and the static water wave characteristic image to obtain a water wave type label, and correcting the preliminary comparison result to obtain the object to be identified. Compared with the prior art, the method and the device for identifying the target by using the underwater flow field have the advantages that the flow field data of the target area are obtained, the moving track is predicted by using the hydrodynamic simulation so as to correct the comparison result, the disturbance factors of the underwater flow field are considered, the accuracy of target identification is effectively improved, in addition, the regional environment image is divided into a plurality of sub-images, the sub-images are respectively subjected to enhancement processing of different degrees or the same degree according to the regional information, the quality of the regional environment image can be improved, the accuracy of comparison analysis is improved, the preliminary comparison result with higher quality is obtained, the accuracy of the target identification result is further improved, besides, the static water wave characteristics are considered, namely the water wave type label is built for correcting the preliminary comparison result by combining relative microscopic and relative macroscopic motions, and the accuracy of the target identification result is further improved.
Drawings
FIG. 1 is a schematic flow chart of an embodiment of an underwater target recognition method provided by the present application.
FIG. 2 is a schematic flow chart of a preferred implementation of an embodiment of the method for identifying underwater targets according to the present invention.
FIG. 3 is a schematic flow chart of another preferred implementation of an embodiment of the underwater target recognition method provided by the present application.
FIG. 4 is a schematic diagram illustrating an embodiment of an underwater target identification system according to 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. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
Referring to fig. 1, fig. 1 is a schematic diagram of an underwater target recognition method according to the present invention, which includes steps S101 to S104,
Step S101, acquiring an underwater image to be identified of a target area and an area environment image in real time.
In the step, the regional environment image is acquired through a multispectral camera and a laser radar, and the wave bands adopted by the multispectral camera comprise visible light, infrared spectrum and ultraviolet spectrum.
The multispectral camera and the laser radar can acquire different wave band data and accurate depth information of a target area, the different wave band data comprise visible light data, infrared spectrum data and ultraviolet spectrum data, the multispectral image can help to capture water surface details, and the laser radar data are favorable for acquiring accurate water surface profile and height information.
In this embodiment, the underwater image to be identified of the target area may be acquired by one or more image acquisition devices. The type of image acquisition device may be a camera, which may be a monocular, binocular or multi-view camera, etc.
In some embodiments, coordinate data of the target area can be obtained through a high-precision positioning system carried by the underwater robot, for example, longitude and latitude coordinates (116.32, 39.96) of the underwater depth of 50 meters are set, and positioning errors are controlled within 0.5 meter by utilizing a sonar positioning technology in combination with an inertial navigation system, so that accuracy of image acquisition is ensured. Next, when underwater image data is acquired, an underwater high definition camera (e.g., an industrial-grade camera with a resolution of 1920×1080) is used to continuously capture images of a target area at a frequency of 5 frames per second for 10 minutes, and about 3000 pieces of image data are acquired, which can be used as an area environment image or a sample image for some pre-processing.
Where the regional environment image refers to an image of the target region when there is no target to be identified and tracked.
According to the embodiment, the target recognition result is obtained through respectively processing the underwater image to be recognized and the regional environment image and comparing and analyzing the underwater image to be recognized and the regional environment image.
It will be appreciated that when there is no target in the target area that needs to be identified, the result of the comparative analysis of the underwater image to be identified and the regional environment image is "consistent" or "highly similar". There is a case of "high similarity" because the subsequent step may perform a certain degree of processing such as enhancement processing on the area environment image, the processed image may be somewhat different from the original image, and the purpose of the processing is to better recognize and make the image quality of the target recognition result higher.
Step S102, identifying the regional environment image to obtain regional information, dividing the regional environment image into a plurality of sub-images, respectively carrying out enhancement processing on each sub-image according to the regional information, acquiring flow field data of the target region in real time, obtaining a predicted moving track of an object to be identified through fluid mechanics simulation, and extracting static water wave features based on the regional environment image by adopting a target detection algorithm to generate a static water wave feature map.
In some implementations, the region information may include information of region illumination, region depth, and the like. The illumination and depth of different areas may be different. Therefore, in this embodiment, by taking into consideration the difference in factors such as illumination and depth, enhancement processing is performed to different degrees or to the same degree on each sub-image in the subsequent step.
In this embodiment, the area environment image may be uniformly divided into a plurality of sub-images, or may be unevenly divided into a plurality of sub-images as appropriate. For example, in some application scenarios, the illumination of the environment outside the water body may be non-uniform, and at this time, the regional environment image may be non-uniformly segmented according to the distribution of the illumination.
Further, in the embodiment in which the region environment image is uniformly divided into a plurality of sub-images, it is considered that enhancement processing of different degrees is performed on each sub-image according to the region information in this case, whereas in the embodiment in which the region environment image is unevenly divided into a plurality of sub-images, enhancement processing of the same degree is performed on each sub-image according to the region information in this case.
The purpose of the enhancement processing in this embodiment is to optimize the details of the image, enhance the sharpness of the image, and highlight the feature information of the target area, so that the step S104 can be better and more accurately identified. In addition, noise filtering, sharpening, image contrast adjustment and the like can be performed.
In some preferred embodiments, as shown in fig. 2, the step S102 of acquiring flow field data of the target area in real time, and obtaining a predicted moving track of the object to be identified through hydrodynamic simulation includes steps S201 to S205. Wherein the steps are detailed as follows:
Step S201, acquiring flow field data of the target area through an underwater sensor network, and determining an initial track according to the flow field data, wherein the flow field data comprises flow velocity information and flow direction information of a flow field;
Step S202, analyzing the flow velocity information and the flow direction information in real time to obtain the disturbance frequency and the disturbance amplitude of a flow field;
Step S203, decomposing the disturbance frequency and the disturbance amplitude of the flow field through Fourier transformation, and determining disturbance characteristic frequency and amplitude fluctuation range;
Step S204, determining track adjustment parameters according to the disturbance characteristic frequency and the amplitude fluctuation range;
Step S205, carrying out hydrodynamic simulation based on a Navigator equation, constructing a fluid model according to flow field data, and predicting the next moment of the initial track by utilizing the track adjustment parameters based on the fluid model to obtain a predicted moving track of the object to be identified.
For example, the underwater sensor network may collect flow field data of a target area once per second, for example, the flow rate in the area is 0.5 m/s, the flow direction is 45 degrees northeast, etc. And then the flow field data can be uploaded to the cloud server in real time, and the calculation power of the cloud server is utilized for processing, analyzing or calculating.
Preferably, assuming that the parameters of the initial trajectory are 5.2 m/s speed and 30 degrees direction angle, the parameters are input into the system through a built-in trajectory planning algorithm, and the expected initial trajectory is calculated by using a quadratic Bezier curve algorithm.
For example, some existing path planning formulas:
B (t) = (1-t) 2×P0 + 2×(1-t) ×t×P1 + t2 ×p2, where P0 is the start point coordinate, P1 is the control point between the start point and the end point, P2 is the end point coordinate, t is the time parameter (or time coefficient) and can vary from 0 to 1, so that an initial trajectory can be obtained.
Further, the disturbance frequency and the disturbance amplitude of the flow field are decomposed through Fourier transformation, the disturbance characteristic frequency is mainly concentrated at 0.1 Hz, the amplitude fluctuation range is 0.2 m/s, the influence of the ocean current on the target track is shown to have periodic characteristics, and the deviation range can be calculated to be +/-1.5 m in the horizontal direction and +/-0.8 m in the vertical direction (the values are only used as examples).
A fluid model is constructed based on the flow field data by hydrodynamic and Navier-Stokes equations and numerical simulation using finite element analysis software.
And predicting and updating the next moment of the initial track by using the simulated fluid model and the calculated track adjustment parameters to obtain the predicted moving track.
And step S103, performing contrast analysis according to each sub-image after the enhancement processing and the underwater image to be identified, and obtaining a preliminary contrast result.
In some preferred embodiments, as shown in fig. 3, the performing a contrast analysis according to each sub-image after the enhancement processing and the underwater image to be identified to obtain a preliminary contrast result includes steps S301 to S304, where each step is described in detail as follows:
Step S301, dividing the underwater image to be identified according to each sub-image after the enhancement processing and the underwater image to be identified, and separating out a target signal and background noise;
step S302, carrying out feature recognition on the target signal to obtain a signal feature set;
Step S303, updating the signal feature set in real time by using a dynamic mapping framework to obtain target feature description, wherein the dynamic mapping framework is used for describing association mapping between environment parameters and signal features;
and step S304, obtaining the preliminary comparison result based on the target feature description.
In the embodiment, the target signal and the background noise can be roughly determined and separated by carrying out contrast analysis on each sub-image after the enhancement processing and the underwater image to be identified, further, the target signal is subjected to feature recognition and extraction to obtain a signal feature set, then, the environment parameters are simulated in real time by utilizing a dynamic mapping frame, the latest target feature description is obtained by utilizing the association mapping between the environment parameters and the signal features in real time, and further, the preliminary comparison result with updated in real time and high image quality is obtained.
Further, the segmenting the underwater image to be identified, separating out a target signal and background noise, includes:
constructing a gray level co-occurrence matrix of the underwater image to be identified, wherein the gray level co-occurrence matrix is used for reflecting texture characteristics of the underwater image to be identified;
decomposing the gray level co-occurrence matrix into a plurality of signal components by using independent component analysis to obtain decomposition results, for example, the signal components are respectively A1, A2 and A3, and the energy ratio can be 70%, 20%, 10% and the like in sequence);
Reconstructing texture features of the underwater image to be identified according to the energy ratio of each signal component in the decomposition result to obtain a reconstruction result;
Screening out the pixel points with the signal intensity larger than a first preset intensity threshold value from the reconstruction result to obtain effective pixel point combinations (for example, 2500 effective points are screened out from an image with the size of 100x100, the ratio is 25 percent, so that low-intensity noise is effectively removed);
and determining a target signal according to the effective pixel point combination, and determining the residual pixel points of the underwater image to be identified as the background noise.
In some preferred embodiments, before updating the signal feature set in real time by using the dynamic mapping frame in step S303, the method further includes obtaining a sample image, performing feature recognition on the sample image to obtain a sample feature, and obtaining an environmental parameter of a region corresponding to the sample image, where the environmental parameter includes suspended particle information and illumination change information, screening out a significant feature (for example, a significant feature with a signal strength greater than a second preset strength threshold) from the sample feature according to the signal strength of the sample image, and constructing a dynamic mapping relation between the significant feature and the suspended particle information and the illumination change information by using a data table, and obtaining the dynamic mapping frame according to the dynamic mapping relation.
Further, the suspended particle information includes water turbidity (in NTU).
The constructing a dynamic mapping relation between the salient features, suspended particle information and illumination change information by using a data table comprises the following steps:
constructing a linear regression model according to the turbidity and illumination change information of the water body;
obtaining a correlation coefficient (for example, positive correlation may be obtained, the value of the correlation coefficient is 0.85, and the value of the correlation coefficient is generally between 0 and 1) based on the linear regression model and the image contrast;
recording model parameters of the linear regression model, the image contrast and the correlation coefficient by using the data table to obtain a recording result;
and recording the recording result as the dynamic mapping relation.
In some further preferred embodiments, when the dynamic mapping relationship is constructed, factors such as illumination intensity (lux, english), image signal-to-noise ratio, gray information when the image is converted into a gray image, and the like can be further considered, so that the accuracy and the refinement degree of the dynamic mapping relationship are further improved.
And step S104, comprehensively analyzing according to the predicted moving track, the preset hydrodynamic condition and the static water wave feature map to obtain a water wave type label, and correcting the preliminary comparison result by utilizing the water wave type label and the predicted moving track to obtain a target recognition result of the object to be recognized.
In this embodiment, the comprehensive analysis is performed according to the predicted movement track, the preset hydrodynamic condition and the static water wave feature map to obtain a water wave type label, which includes acquiring a water wave dynamic video in the same time period according to time information of flow field data of the target area, splitting the water wave dynamic video into a water wave image sequence, setting feature conditions based on feature similarity according to the static water wave feature map, screening a plurality of water wave images from the water wave image sequence according to the feature conditions to obtain a screening result, and deducting based on the screening result, the predicted movement track and the preset hydrodynamic condition to obtain a plurality of water wave type labels.
According to the preferred embodiment, based on the static water wave feature map and the feature similarity set feature conditions, a plurality of water wave images are screened out from a water wave image sequence obtained by splitting the water wave dynamic video, so that screening results are obtained, screening results meeting the water wave feature requirements can be ensured, the quality of the screened images is ensured, and further in the deduction of the follow-up step, the accurate setting of the water wave type labels can be ensured.
Further, the water wave image of the screening result comprises an interference object; the method comprises the steps of obtaining a plurality of water wave type labels based on screening results, the predicted moving tracks and preset hydrodynamic conditions, extracting contour points of each interference object in water wave images of the screening results, cutting off water wave lines in the water wave images containing the interference objects in the screening results based on the contour points of the interference objects, extracting characteristic points of the water wave lines, connecting the characteristic points to obtain water wave moving tracks, calling a pre-trained water wave prediction model based on the water wave moving tracks to obtain water wave predicting tracks, performing deduction on the basis of the preset hydrodynamic conditions according to the water wave predicting tracks and the predicted moving tracks to obtain a plurality of water wave moving modes, and determining the water wave moving modes as the water wave type labels.
According to the preferred embodiment, the water wave movement track can be analyzed by means of the interference objects, specifically, the contour points of the interference objects are taken as the reference, the water wave lines in each water wave image containing the interference objects in the screening result are cut off, the characteristic points are extracted, the water wave movement track is obtained by connecting the water wave lines, the water wave prediction track is further obtained, then a plurality of water wave movement modes can be deduced, clustered and/or classified, and a reference is provided for determining the reference track in the follow-up step, so that the deviation value between the movement track and the reference track can be accurately obtained.
In some preferred embodiments, the correcting the preliminary comparison result by using the water wave category label and the predicted movement track to obtain the target recognition result of the object to be recognized includes:
Acquiring a plurality of distinguishing pixel characteristics of each sub-image after enhancement processing and the underwater image to be identified from the preliminary comparison result, wherein each distinguishing pixel characteristic comprises a plurality of pixel points;
Analyzing the predicted moving track to obtain a deviation value between the predicted moving track and the reference track;
selecting track points with deviation values larger than a preset deviation threshold value from the predicted moving track;
And correcting the preliminary comparison result according to the superposition condition of the track points and the distinguishing pixel characteristics to obtain a target identification result.
In this embodiment, the underwater image to be identified is identified through one or more embodiments described above, so as to obtain one or more target objects, and obtain a target identification result. The target object may be a fish or a plant or the like under water. According to the method and the device, factors such as disturbance of the underwater flow field are considered, the deviation threshold is further utilized to consider the disturbance factor, and compared with the prior art, the method and the device consider the technical scheme except the water body, and therefore accuracy of target identification is effectively improved.
Accordingly, as shown in fig. 4, the present invention further provides an underwater target recognition system 400, which includes an image acquisition module 401, an area information processing module 402, a comparison module 403, and a target recognition module 404, wherein,
The image acquisition module 401 is configured to acquire an underwater image to be identified of a target area and an area environment image in real time, wherein the area environment image is acquired through a multispectral camera and a laser radar, and a wave band adopted by the multispectral camera comprises visible light, infrared spectrum and ultraviolet spectrum;
The regional information processing module 402 is used for identifying the regional environment image to obtain regional information, dividing the regional environment image into a plurality of sub-images, and respectively carrying out enhancement processing on each sub-image according to the regional information;
The comparison module 403 is configured to perform a comparison analysis according to each sub-image after the enhancement processing and the underwater image to be identified, so as to obtain a preliminary comparison result;
The target recognition module 404 is configured to perform comprehensive analysis according to the predicted movement track, a preset hydrodynamic condition, and the static water wave feature map to obtain a water wave type tag, and correct the preliminary comparison result by using the water wave type tag and the predicted movement track to obtain a target recognition result of the object to be recognized.
Preferably, the target recognition module 404 performs a comparison analysis according to each sub-image after the enhancement processing and the underwater image to be recognized, so as to obtain a preliminary comparison result, which includes:
The target recognition module 404 segments the underwater image to be recognized according to each sub-image after the enhancement processing and the underwater image to be recognized, and separates out a target signal and background noise;
Performing feature recognition on the target signal to obtain a signal feature set;
Updating the signal feature set in real time by using a dynamic mapping framework to obtain target feature description, wherein the dynamic mapping framework is used for describing association mapping between environment parameters and signal features;
and obtaining the preliminary comparison result based on the target feature description.
Preferably, the object recognition module 404 segments the underwater image to be recognized, separates out the object signal and the background noise, and includes:
The target recognition module 404 constructs a gray level co-occurrence matrix of the underwater image to be recognized, wherein the gray level co-occurrence matrix is used for reflecting texture features of the underwater image to be recognized;
decomposing the gray level co-occurrence matrix into a plurality of signal components by utilizing independent component analysis to obtain a decomposition result;
Reconstructing texture features of the underwater image to be identified according to the energy ratio of each signal component in the decomposition result to obtain a reconstruction result;
the signal intensity of each pixel point in the reconstruction result is determined, and the pixel points with the signal intensity larger than a first preset intensity threshold value are screened out from the reconstruction result to obtain an effective pixel point combination;
and determining a target signal according to the effective pixel point combination, and determining the residual pixel points of the underwater image to be identified as the background noise.
Preferably, the underwater target identification system 400 further comprises a mapping frame construction module for, before the updating of the signal feature set in real time using a dynamic mapping frame:
Acquiring a sample image, carrying out feature recognition on the sample image to obtain sample features, and acquiring environmental parameters of a corresponding area of the sample image, wherein the environmental parameters comprise suspended particle information and illumination change information;
screening out significant features from the sample features according to the signal intensity of the sample image, and constructing a dynamic mapping relation between the significant features and the suspended particle information and the illumination change information by utilizing a data table;
And obtaining the dynamic mapping frame according to the dynamic mapping relation.
The mapping frame construction module utilizes a data table to construct a dynamic mapping relation between the salient features, the suspended particle information and the illumination change information, and the mapping frame construction module comprises:
The mapping frame construction module constructs a linear regression model according to the turbidity and illumination change information of the water body;
Obtaining a correlation coefficient based on the linear regression model and the image contrast;
recording model parameters of the linear regression model, the image contrast and the correlation coefficient by using the data table to obtain a recording result;
and recording the recording result as the dynamic mapping relation.
As a preferred solution, the target recognition module performs comprehensive analysis according to the predicted movement track, a preset hydrodynamic condition and the static water wave feature map to obtain a water wave class label, which includes:
acquiring a water wave dynamic video in the same time period according to the time information of the flow field data of the target area, and splitting the water wave dynamic video into a water wave image sequence;
Setting feature conditions based on feature similarity according to the static water wave feature map, and screening a plurality of water wave images from the water wave image sequence according to the feature conditions to obtain screening results;
And deducing based on the screening result, the predicted moving track and preset hydrodynamic conditions to obtain a plurality of water wave class labels.
The target recognition module derives a plurality of water wave category labels based on the screening result, the predicted moving track and a preset hydrodynamic condition, and comprises:
the target recognition module extracts contour points of each interference object in the water wave image of the screening result;
Cutting off water wave lines in each water wave image containing the interference object in the screening result by taking the outline points of the interference object as references, extracting characteristic points of the water wave lines, and connecting the characteristic points to obtain a water wave movement track;
Invoking a pre-trained water wave prediction model based on the water wave movement track to obtain a water wave prediction track;
and carrying out deduction on the basis of preset hydrodynamic conditions according to the water wave prediction track and the prediction movement track to obtain a plurality of water wave movement modes, and determining the water wave movement modes as water wave type labels.
As a preferred solution, the area information processing module 402 obtains flow field data of the target area in real time, and obtains a predicted moving track of the object to be identified through fluid mechanics simulation, including:
The area information processing module 402 acquires flow field data of the target area through an underwater sensor network and determines an initial track according to the flow field data, wherein the flow field data comprises flow velocity information and flow direction information of a flow field;
analyzing the flow velocity information and the flow direction information in real time to obtain the disturbance frequency and the disturbance amplitude of the flow field;
Decomposing the disturbance frequency and the disturbance amplitude of the flow field through Fourier transformation, and determining disturbance characteristic frequency and amplitude fluctuation range;
determining track adjustment parameters according to the disturbance characteristic frequency and the amplitude fluctuation range;
and predicting the next moment of the initial track by utilizing the track adjustment parameters based on the fluid model to obtain the predicted moving track of the object to be identified.
Preferably, the target recognition module 404 corrects the preliminary comparison result by using the water wave category label and the predicted movement track to obtain a target recognition result of the object to be recognized, including:
The target recognition module 404 obtains each sub-image after enhancement processing and a plurality of distinguishing pixel characteristics of the underwater image to be recognized from the preliminary comparison result, wherein each distinguishing pixel characteristic comprises a plurality of pixel points;
Analyzing the predicted moving track to obtain a deviation value between the predicted moving track and the reference track;
selecting track points with deviation values larger than a preset deviation threshold value from the predicted moving track;
And correcting the preliminary comparison result according to the superposition condition of the track points and the distinguishing pixel characteristics to obtain a target identification result.
Compared with the prior art, the invention has the following beneficial effects:
The invention provides an underwater target identification method and system, the method comprises the steps of acquiring an underwater image to be identified of a target area and an area environment image in real time, acquiring the area environment image through a multispectral camera and a laser radar, wherein wave bands adopted by the multispectral camera comprise visible light, infrared spectrum and ultraviolet spectrum, identifying the area environment image to obtain area information, dividing the area environment image into a plurality of sub-images, respectively carrying out enhancement processing on the sub-images according to the area information, acquiring flow field data of the target area in real time, obtaining a predicted moving track of an object to be identified through hydrodynamic simulation, adopting a target detection algorithm, extracting static water wave characteristics based on the area environment image to generate a static water wave characteristic image, carrying out comparison analysis according to the sub-images after enhancement processing and the underwater image to be identified to obtain a preliminary comparison result, carrying out comprehensive analysis according to the predicted moving track, a preset hydrodynamic condition and the static water wave characteristic image to obtain a water wave type label, and correcting the preliminary comparison result to obtain the object to be identified. Compared with the prior art, the method and the device for identifying the target by using the underwater flow field have the advantages that the flow field data of the target area are obtained, the moving track is predicted by using the hydrodynamic simulation so as to correct the comparison result, the disturbance factors of the underwater flow field are considered, the accuracy of target identification is effectively improved, in addition, the regional environment image is divided into a plurality of sub-images, the sub-images are respectively subjected to enhancement processing of different degrees or the same degree according to the regional information, the quality of the regional environment image can be improved, the accuracy of comparison analysis is improved, the preliminary comparison result with higher quality is obtained, the accuracy of the target identification result is further improved, besides, the static water wave characteristics are considered, namely the water wave type label is built for correcting the preliminary comparison result by combining relative microscopic and relative macroscopic motions, and the accuracy of the target identification result is further improved.
The foregoing embodiments have been provided for the purpose of illustrating the general principles of the present invention, and are not to be construed as limiting the scope of the invention. It should be noted that any modifications, equivalent substitutions, improvements, etc. made by those skilled in the art without departing from the spirit and principles of the present invention are intended to be included in the scope of the present invention.
Claims (10)
1. An underwater target identification method, comprising:
Acquiring an underwater image to be identified of a target area and an area environment image in real time, wherein the area environment image is acquired through a multispectral camera and a laser radar, and the multispectral camera adopts wave bands including visible light, infrared spectrum and ultraviolet spectrum;
The method comprises the steps of identifying an area environment image to obtain area information, dividing the area environment image into a plurality of sub-images, respectively carrying out enhancement processing on each sub-image according to the area information, acquiring flow field data of a target area in real time, obtaining a predicted moving track of an object to be identified through fluid mechanics simulation, extracting static water wave characteristics based on the area environment image by adopting a target detection algorithm, and generating a static water wave characteristic image;
performing contrast analysis according to each sub-image after enhancement treatment and the underwater image to be identified to obtain a preliminary contrast result;
And correcting the preliminary comparison result by utilizing the water wave type label and the predicted movement track to obtain a target recognition result of the object to be recognized.
2. The method for identifying an underwater target according to claim 1, wherein the performing a comparison analysis according to each sub-image after the enhancement processing and the underwater image to be identified to obtain a preliminary comparison result comprises:
dividing the underwater image to be identified according to each sub-image after the enhancement processing and the underwater image to be identified, and separating out a target signal and background noise;
Performing feature recognition on the target signal to obtain a signal feature set;
Updating the signal feature set in real time by using a dynamic mapping framework to obtain target feature description, wherein the dynamic mapping framework is used for describing association mapping between environment parameters and signal features;
and obtaining the preliminary comparison result based on the target feature description.
3. An underwater target identification method as claimed in claim 2, further comprising, prior to said updating of said signal feature set in real time using a dynamic mapping framework:
Acquiring a sample image, carrying out feature recognition on the sample image to obtain sample features, and acquiring environmental parameters of a corresponding area of the sample image, wherein the environmental parameters comprise suspended particle information and illumination change information;
screening out significant features from the sample features according to the signal intensity of the sample image, and constructing a dynamic mapping relation between the significant features and the suspended particle information and the illumination change information by utilizing a data table;
And obtaining the dynamic mapping frame according to the dynamic mapping relation.
4. The method of claim 3, wherein the salient features include image contrast, the suspended particle information includes turbidity of the body of water, and the constructing a dynamic mapping relationship between the salient features and the suspended particle information and the illumination variation information using the data table includes:
constructing a linear regression model according to the turbidity and illumination change information of the water body;
Obtaining a correlation coefficient based on the linear regression model and the image contrast;
recording model parameters of the linear regression model, the image contrast and the correlation coefficient by using the data table to obtain a recording result;
and recording the recording result as the dynamic mapping relation.
5. The method for identifying an underwater target according to claim 1, wherein the comprehensively analyzing according to the predicted moving track, the preset hydrodynamic condition and the static water wave feature map to obtain a water wave type tag comprises:
acquiring a water wave dynamic video in the same time period according to the time information of the flow field data of the target area, and splitting the water wave dynamic video into a water wave image sequence;
Setting feature conditions based on feature similarity according to the static water wave feature map, and screening a plurality of water wave images from the water wave image sequence according to the feature conditions to obtain screening results;
And deducing based on the screening result, the predicted moving track and preset hydrodynamic conditions to obtain a plurality of water wave class labels.
6. The method for identifying an underwater target according to claim 5, wherein the water wave image of the screening result includes an interfering object, the deduction is performed based on the screening result, the predicted moving track and a preset hydrodynamic condition to obtain a plurality of water wave class labels, the method comprises:
extracting contour points of each interference object in the water wave image of the screening result;
Cutting off water wave lines in each water wave image containing the interference object in the screening result by taking the outline points of the interference object as references, extracting characteristic points of the water wave lines, and connecting the characteristic points to obtain a water wave movement track;
Invoking a pre-trained water wave prediction model based on the water wave movement track to obtain a water wave prediction track;
and carrying out deduction on the basis of preset hydrodynamic conditions according to the water wave prediction track and the prediction movement track to obtain a plurality of water wave movement modes, and determining the water wave movement modes as water wave type labels.
7. The method for identifying an underwater target according to claim 1, wherein the acquiring flow field data of the target area in real time, obtaining a predicted moving track of an object to be identified through hydrodynamic simulation, comprises:
Acquiring flow field data of the target area through an underwater sensor network, and determining an initial track according to the flow field data, wherein the flow field data comprises flow velocity information and flow direction information of a flow field;
analyzing the flow velocity information and the flow direction information in real time to obtain the disturbance frequency and the disturbance amplitude of the flow field;
Decomposing the disturbance frequency and the disturbance amplitude of the flow field through Fourier transformation, and determining disturbance characteristic frequency and amplitude fluctuation range;
determining track adjustment parameters according to the disturbance characteristic frequency and the amplitude fluctuation range;
and predicting the next moment of the initial track by utilizing the track adjustment parameters based on the fluid model to obtain the predicted moving track of the object to be identified.
8. The method for identifying an underwater target according to claim 1, wherein the correcting the preliminary comparison result by using the water wave type label and the predicted movement track to obtain the target identification result of the object to be identified comprises:
Acquiring a plurality of distinguishing pixel characteristics of each sub-image after enhancement processing and the underwater image to be identified from the preliminary comparison result, wherein each distinguishing pixel characteristic comprises a plurality of pixel points;
Analyzing the predicted moving track to obtain a deviation value between the predicted moving track and the reference track;
selecting track points with deviation values larger than a preset deviation threshold value from the predicted moving track;
And correcting the preliminary comparison result according to the superposition condition of the track points and the distinguishing pixel characteristics to obtain a target identification result.
9. An underwater target recognition system is characterized by comprising an image acquisition module, an area information processing module, a comparison module and a target recognition module, wherein,
The image acquisition module is used for acquiring an underwater image to be identified of a target area and an area environment image in real time, wherein the area environment image is acquired through a multispectral camera and a laser radar, and wave bands adopted by the multispectral camera comprise visible light, infrared spectrum and ultraviolet spectrum;
The regional information processing module is used for identifying the regional environment image to obtain regional information, dividing the regional environment image into a plurality of sub-images, and respectively carrying out enhancement processing on each sub-image according to the regional information;
the contrast module is used for carrying out contrast analysis according to each sub-image after the enhancement treatment and the underwater image to be identified to obtain a preliminary contrast result;
the target recognition module is used for comprehensively analyzing according to the predicted moving track, the preset hydrodynamic condition and the static water wave feature map to obtain a water wave type label, and correcting the preliminary comparison result by utilizing the water wave type label and the predicted moving track to obtain a target recognition result of the object to be recognized.
10. The underwater target recognition system of claim 9, wherein the target recognition module performs a comparison analysis according to each sub-image after the enhancement process and the underwater image to be recognized, to obtain a preliminary comparison result, comprising:
The target recognition module divides the underwater image to be recognized according to each sub-image after the enhancement processing and the underwater image to be recognized, and separates out a target signal and background noise;
Performing feature recognition on the target signal to obtain a signal feature set;
Updating the signal feature set in real time by using a dynamic mapping framework to obtain target feature description, wherein the dynamic mapping framework is used for describing association mapping between environment parameters and signal features;
and obtaining the preliminary comparison result based on the target feature description.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202511071807.9A CN120564022B (en) | 2025-08-01 | 2025-08-01 | Underwater target identification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202511071807.9A CN120564022B (en) | 2025-08-01 | 2025-08-01 | Underwater target identification method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN120564022A true CN120564022A (en) | 2025-08-29 |
| CN120564022B CN120564022B (en) | 2025-09-26 |
Family
ID=96820731
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202511071807.9A Active CN120564022B (en) | 2025-08-01 | 2025-08-01 | Underwater target identification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN120564022B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180112983A1 (en) * | 2016-10-24 | 2018-04-26 | Invensense Inc. | Method and system for global shape matching a trajectory |
| JP2019050007A (en) * | 2018-11-01 | 2019-03-28 | エスゼット ディージェイアイ テクノロジー カンパニー リミテッドＳｚ Ｄｊｉ Ｔｅｃｈｎｏｌｏｇｙ Ｃｏ．，Ｌｔｄ | Method and device for determining position of mobile body and computer readable medium |
| CN118015445A (en) * | 2024-02-06 | 2024-05-10 | 大连理工大学 | An underwater target detection method based on YoloV7 and multimodal data |
| CN119942355A (en) * | 2025-04-09 | 2025-05-06 | 沈阳宸阳信息技术有限公司 | A method and system for intelligent water pattern analysis of water conservancy projects |
-
2025
- 2025-08-01 CN CN202511071807.9A patent/CN120564022B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180112983A1 (en) * | 2016-10-24 | 2018-04-26 | Invensense Inc. | Method and system for global shape matching a trajectory |
| JP2019050007A (en) * | 2018-11-01 | 2019-03-28 | エスゼット ディージェイアイ テクノロジー カンパニー リミテッドＳｚ Ｄｊｉ Ｔｅｃｈｎｏｌｏｇｙ Ｃｏ．，Ｌｔｄ | Method and device for determining position of mobile body and computer readable medium |
| CN118015445A (en) * | 2024-02-06 | 2024-05-10 | 大连理工大学 | An underwater target detection method based on YoloV7 and multimodal data |
| CN119942355A (en) * | 2025-04-09 | 2025-05-06 | 沈阳宸阳信息技术有限公司 | A method and system for intelligent water pattern analysis of water conservancy projects |
Also Published As
| Publication number | Publication date |
|---|---|
| CN120564022B (en) | 2025-09-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20200160061A1 (en) | Automatic ship tracking method and system based on deep learning network and mean shift | |
| CN112257605B (en) | 3D target detection method, system and device based on self-labeled training samples | |
| Periyasamy et al. | How to get the most out of u-net for glacier calving front segmentation | |
| CN119339063B (en) | A tunnel disease identification and detection method and system based on deep learning | |
| Zheng et al. | Tuning a SAM-based model with multicognitive visual adapter to remote sensing instance segmentation | |
| Anilkumar et al. | An adaptive DeepLabv3+ for semantic segmentation of aerial images using improved golden eagle optimization algorithm | |
| CN119417912A (en) | External parameter calibration method, device and electronic equipment | |
| CN120564022B (en) | Underwater target identification method and system | |
| CN115331082B (en) | Path generation method of tracking sound source, training method of model and electronic equipment | |
| CN116310321A (en) | A Method of Object Detection in High Resolution Aerial Imagery | |
| CN116229395A (en) | A target detection model training method, target detection method, device and vehicle | |
| CN113920311B (en) | A remote sensing image segmentation method and system based on edge auxiliary information | |
| CN120708211A (en) | A detection method, device and storage medium for intelligent visual detection of parts | |
| CN120070511A (en) | Target recognition and positioning-based bird sound directional acquisition method, device and storage medium | |
| CN120070837A (en) | Deep sea sampling area positioning method based on deep learning and multi-mode data | |
| CN119896494A (en) | Ultrasonic probe positioning method, device, electronic device and computer readable medium | |
| CN117876422A (en) | A method for calculating sea ice drift velocity vector | |
| CN114485608B (en) | Local point cloud rapid registration method for high-precision map making | |
| CN113658180B (en) | Surface defect region segmentation method and device based on spatial context guidance | |
| CN116543017A (en) | Single-target tracking method based on sparse optical flow motion enhancement | |
| CN113870412B (en) | A method, device and electronic device for processing aerial scene images | |
| CN115222771B (en) | A target tracking method and device | |
| CN119846612B (en) | Multi-beam sonar target detection and imaging method and system | |
| CN104182990A (en) | A method for acquiring a sequence image motion target area in real-time | |
| CN121074619B (en) | Device and method for intelligently investigating fish types and amounts |
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 |