CN108520511A - A kind of underwater fish target detection and identification method based on fish finder - Google Patents
A kind of underwater fish target detection and identification method based on fish finder Download PDFInfo
- Publication number
- CN108520511A CN108520511A CN201810226818.3A CN201810226818A CN108520511A CN 108520511 A CN108520511 A CN 108520511A CN 201810226818 A CN201810226818 A CN 201810226818A CN 108520511 A CN108520511 A CN 108520511A
- Authority
- CN
- China
- Prior art keywords
- fish
- background
- model
- pixel
- underwater
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F18/2155—Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the incorporation of unlabelled data, e.g. multiple instance learning [MIL], semi-supervised techniques using expectation-maximisation [EM] or naïve labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10132—Ultrasound image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30242—Counting objects in image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
Abstract
The underwater fish target acquisition Study of recognition method based on fish finder that the invention discloses a kind of, includes the following steps:Classification based training is carried out firstly for the underwater sonar image of different fish, obtains effective recognition classifier;Sonar image is acquired secondly by fish finder and is pre-processed;Then pass through self organizing neural network background subtraction model initial back-ground model, sonar image is reassembled as to go to hsv color space after triple channel image, and Gauss weights initialisation background model is used, calculate the minimum range of pixel value in the corresponding background model of pixel;Finally judged, if minimum range is less than threshold value, which is divided into background model, carries out model modification;If it is greater than or equal to threshold value, shade judgement is carried out;If not background model or shade, then be judged as foreground, foreground target be sent into trained effective recognition classifier, classification is identified, and then to the quantity of fish, type, size is counted.The present invention can rapidly and accurately detect fish body target, and classification is correctly effectively identified, and can be applied to the monitoring of the underwater shoal of fish.
Description
Technical field
The present invention relates to a kind of locating fish recognition methods, belong to Intelligent Information Processing and object detection and recognition technology neck
Domain.
Background technology
Fishery resources Acoustic assessment gradually develops the living marine resources investigation and assessment improved as nearly more than 30 year
Method, accurate, assessment fish in time resources and quantity have fisheries management, production, development of resources and relevant industries important
Meaning.
Fishery resources Acoustic assessment is aquatic as detecting using more fixed point reflection sounding-Integral Technologies, wherein fish finder
A kind of important means in goods and materials source, is widely used in fishery acoustic survey.The main method of acoustic sounding has remote control to walk boat side
Method and fix-point method realize that full Cultivated water reliable is surveyed, Cultivated water are obtained after the fused processing of juxtaposition regional signal
The statistical information of interior whole fish.But basic research of the China in terms of fish finder and aquatic resources assessment be not deep, independent research
Product is less, and especially analysis software does not have substantially, although external have in-depth study and form tandem product and corresponding
Specification, but external research is the aquatic resources based on ocean and lake, and feature is that water quality is deep and limpid, fish
Type is comparatively fewer.The characteristics of country's cultivation, is different from foreign countries, and not ready-made instrument and system is available, therefore,
The assessment instrument and system development for carrying out the acoustics aquatic resources of independent intellectual property right, can fill up domestic fishery cultivating science and comment
Estimate the blank of means, pushes the fast development of domestic fishery cultivating and insurance.
Fish finder is mounted in lake-bottom, is fishery resources sound to the sonar that underwater fish target is detected and counted
The distinctive assessment equipment of assessment aspect is learned, accurate, assessment fish in time resources and quantity open fisheries management, production, resource
Hair and relevant industries are of great significance.
In the prior art, what fish finder was commonly used is sonar technology, and fish finder usually sends out high frequency in itself when work
Then sound wave judges the orientation and size of the shoal of fish according to echo, shape, the spy of object are determined by sound transmission and radiation
Survey and identify object in water.The prior art usually first obtains and shows the detecting fish school letter of the target shoal of fish in default detection direction
Breath, the location information and size information of the display target shoal of fish;Again by identifying the image of the target shoal of fish, determining and showing detection fish
The shoal of fish information that group's information does not include.But this can only obtain the general profile of the target shoal of fish, and contain much noise information, no
It can obtain accurate target shoal of fish information.
The sonar image that fish finder obtains contains much noise information and shade is serious, using the target of common optical imagery
Detection algorithm cannot meet the requirement of sonar image target detection.
Invention content
In view of the above-mentioned problems, the object of the present invention is to provide a kind of fast and accurately underwater fish mesh based on fish finder
Mark detection and recognition methods carry out foreground target detection using self organizing neural network background subtraction algorithm and shade judge, energy
The position of fish is detected in the sonar image sequence that relatively good slave fish finder obtains and carries out effectively identification statistics, it is existing to make up
The deficiency of technology.
In order to achieve the above objectives, the present invention adopts the following technical scheme that:
A kind of underwater fish target detection and identification method based on fish finder, includes the following steps:
Step 1: collecting the underwater sonar image for the different fish that fish finder obtains, operating limit learning machine ELM
(Extreme Learning Machine) method carries out classification based training to fish sonar image and obtains recognition classifier;
Step 2: using both horizontally and vertically wave beam automatically scanning, using fixed point detection mode, using fish finder to visiting
The fish surveyed in range carry out acoustic sounding, obtain the underwater sonar image of all fishes in search coverage;
Step 3: being acquired to the underwater sonar image obtained above by fish finder, n frames identification sound to be detected is obtained
Image I={ I1,…,Ii,…In, wherein IiIt indicates the i-th frame identification sonar image sequence to be detected, is gone using image filtering
It makes an uproar, contrast enhancing pre-processes identification sonar image sequence to be detectedReduce noise and prominent identification region to be detected;
Step 4: the t frames obtained in step 3 identification sonar image sequence I to be detectedtThe self-organizing god of (m × n)
It is I through network context subduction (SOBS) model initial back-ground modelB(3m×3n);
Step 5: by above-mentioned ItIt is reassembled as going to hsv color space after triple channel image, and initial using Gauss weight
Change background model, i.e. ItThe corresponding background model of a pixel (x, y) in (i, j) (i=n*x, n*x+1 ..., x* (n+1)-
1);J=n*y, n*y+1 ..., y* (n+1) -1)), each value is (H in background modeli,j,Si,j,Vi,j);
Step 6: in It+1In, calculate It+1Pixel p in frametN in corresponding background model2The most narrow spacing of a pixel value
From that is,WhereinFor ptCorresponding n2Background pixel, cmFor ptIt is matched to
Background pixel, d be to illumination variation have stronger robustness HSV Color Hexcone, i.e.,
If Step 8: d (cm,pt) be more than or equal to threshold value, then shade judgement is carried out, judgment rule is as follows,Wherein remove pt, ciOuter is all parameter preset, if it is determined that be shade,
It is then only identified as background, does not update background model;
Step 9: if not background model or shade, then be judged as foreground, s foreground target (B is obtained1(x,y),...,
Bi(x,y),...,Bs(x,y));
Classification is identified Step 10: the foreground target of acquisition is sent into the trained effective recognition classifiers of ELM, into
And to the quantity of fish, type, size is counted.
Beneficial effects of the present invention:The present invention is solved using self organizing neural network background subtraction algorithm and ELM algorithms
The detection for the underwater sonar image of fish class that fixed-point type fish finder obtains and identification problem, compared with common algorithm of target detection,
Target detection effect is more preferable, can quickly detect fish body target, correctly judges target and shade, fish body is clear-cut, after being
Continuous identifying processing offer accurately enters data.The identification classification that sonar image is carried out using ELM algorithms, according to different fish
Sonar reflected intensity it is different, ELM algorithms learn the feature to different fish, can the effective correctly fish sound to detecting
Image carries out precisely quickly identification classification.Algorithm proposed by the present invention can meet requirement of real-time, can be applied to underwater fish
Group's monitoring application.The present invention has merged detection and recognizer, is asked for fish finder sonar image noise is more, shade is serious
It inscribes, the position of detection fish and progress, which effectively identify, in the relatively good sonar image sequence obtained from fish finder of energy counts.
Description of the drawings
Fig. 1 is the overall flow figure of the present invention.
Fig. 2 is the underwater lake sonar figure that fish finder obtains in embodiment 1.
Fig. 3 is the foreground target figure that Fig. 2 separation foregrounds are obtained with background.
Fig. 4 is that the foreground target obtained by Fig. 3 judges result figure via the identification classification results and shade of ELM.
In figure, black box mark is the shade judgement obtained.
Specific implementation mode
To keep the purpose of the present invention, embodiment and advantage relatively sharp, below in conjunction with the accompanying drawings and pass through specific embodiment
It is next that present invention be described in more detail.
Embodiment 1:
Surf-HM fish finders used by the present embodiment are the fixed point detection instruments built in lake, to underwater fish
The sonar that target is detected and counted, the detection sonar are mainly made of green end and dry end, and wherein green end design, which is equipped with, transmits/receives
Energy converter and scan module, green end are connect by sealing circular flange with loading platform, and dry end design is installed in loading platform
Portion, dry end is interior to be equipped with circuit part, and circuit part includes expelling plate, receiver board, Signal acquiring and processing plate and bottom plate, Yi Jiyu
The network of external connection, 232 serial ports and power interface.
The present embodiment data collecting location:The Jiangsu Province Wuxi fish ponds Yao Wan (about 3200 square metres of area, the depth of water about 2.3
Rice).
The particular flow sheet of the present embodiment is as shown in Figure 1.
In the present embodiment specifically using one section of sonar sequence obtained from above-mentioned place by fish finder be used as it is to be detected and
Identification sonar sequence, as shown in Figure 2.
Following steps should be described in detail in conjunction with attached drawing and concrete outcome, and should be general in invention content
The step of condition.
Step 1: collecting a large amount of underwater sonar figures for the different fish (crucian, bream, black carp, silver carp) that fish finder obtains
Picture carries out classification based training to fish sonar image using ELM methods, obtains an effective recognition classifier;
Step 2: using both horizontally and vertically wave beam automatically scanning, using fixed point detection mode, realize in investigative range
The acoustic sounding of fish, degree of precision obtain the underwater sonar image of all fishes in culturing area;
Step 3: being acquired to the underwater sonar image obtained by fish finder, n frames identification sonar figure to be detected is obtained
As I={ I1,…,Ii,…In, wherein IiIt indicates the i-th frame identification sonar image sequence to be detected, utilizes image filtering denoising, right
Identification sonar image sequence to be detected is pre-processed than degree enhancingReduce noise and prominent identification region to be detected interested;
Step 4: the t frames obtained in step 3 identification sonar image sequence I to be detectedtThe self-organizing god of (m × n)
It is I through network context subduction (SOBS) model initial back-ground modelB(3m×3n);
Step 5:;By above-mentioned ItIt is reassembled as going to hsv color space after triple channel image, and using at the beginning of Gauss weight
Beginningization background model, i.e. ItA pixel (x, y) answer in background model (i, j) (i=n*x, n*x+1 ..., x* (n+1)-
1);J=n*y, n*y+1 ..., y* (n+1) -1)), each value is (H in background modeli,j,Si,j,Vi,j);
Step 6: in It+1In, calculate It+1Pixel p in frametN in corresponding background model2The most narrow spacing of a pixel value
From that is,WhereinFor ptCorresponding n2Background pixel, cmFor ptIt is matched to
Background pixel, d be to illumination variation have stronger robustness HSV Color Hexcone, i.e.,
If Step 7: d (cm,pt) it is less than threshold value, the then pixel ptIt is divided into background model (such as Fig. 3 black portions), carries out
Model modification.Assuming that matched pixel cmPosition is (x', y'), then background template pixel cm neighboring pixels are updated, update rule
It is then as follows,Wherein, α (t) is constant, ωi,jIt is background
When modeling, the corresponding weight of each pixel, is initially Gauss weight, A in background modeltTo update rear backdrop template, At-1For it
Preceding background template;
If Step 8: d (cm,pt) it is more than or equal to threshold value, shade judgement is carried out, judgment rule is as follows,Wherein remove pt, ciOuter is all parameter preset, if it is determined that be shade,
It is then only identified as background, does not update background model (such as Fig. 4 black box identification division);
Step 9: if not background model or shade, then be judged as foreground, s foreground target (B is obtained1(x,y),...,
Bi(x,y),...,Bs(x, y)) (such as Fig. 3 white indicias region is the foreground target obtained);
Classification is identified Step 10: the foreground target of acquisition is sent into the trained effective recognition classifiers of ELM, into
And to the quantity of fish, type, size is counted.
Detection, identification and statistics result are shown in Fig. 4, are the fish identified in rectangle frame, and to its size by rectangle frame
Long and wide mark is carried out, the upper left corner is the display to the programming count result of the value volume and range of product of fish, verified detection
Recognition result is roughly the same with legitimate reading, and accuracy rate is up to 95%.
Claims (3)
1. a kind of underwater fish target detection and identification method based on fish finder, it is characterised in that include the following steps:
Step 1: collecting the underwater sonar image for the different fish that fish finder obtains, operating limit learning machine ELM methods are to fish
Sonar image carries out classification based training and obtains recognition classifier;
Step 2: using both horizontally and vertically wave beam automatically scanning, using fixed point detection mode, using fish finder to detecting model
Fish in enclosing carry out acoustic sounding, obtain the underwater sonar image of all fishes in search coverage;
Step 3: being acquired to the underwater sonar image obtained above by fish finder, n frames identification sonar figure to be detected is obtained
As I={ I1,…,Ii,…In, wherein IiIt indicates the i-th frame identification sonar image sequence to be detected, utilizes image filtering denoising, right
Identification sonar image sequence to be detected is pre-processed than degree enhancingReduce noise and prominent identification region to be detected;
Step 4: the t frames obtained in step 3 identification sonar image sequence I to be detectedtThe self organizing neural network of (m × n)
Background subtraction model initial back-ground model is IB(3m×3n);
Step 5: by above-mentioned ItIt is reassembled as going to hsv color space after triple channel image, and is carried on the back using Gauss weights initialisation
Scape model, i.e. ItThe corresponding background model of a pixel (x, y) in (i, j) (i=n*x, n*x+1 ..., x* (n+1) -1);j
=n*y, n*y+1 ..., y* (n+1) -1)), each value is (H in background modeli,j,Si,j,Vi,j);
Step 6: in It+1In, calculate It+1Pixel p in frametN in corresponding background model2The minimum range of a pixel value, i.e.,Wherein C=(c1,c2,...,cn2) it is ptCorresponding n2Background pixel, cmFor ptIt is matched to
Background pixel, d are the HSV Color Hexcone for having stronger robustness to illumination variation, i.e.,
If Step 7: d (cm,pt) it is less than threshold value, the then pixel ptIt is divided into background model, carries out model modification, it is assumed that matching picture
Plain cmPosition is (x', y'), then background template pixel cmNeighboring pixel is updated;
If Step 8: d (cm,pt) be more than or equal to threshold value, then shade judgement is carried out, if it is determined that being shade, is then only identified as background,
Do not update background model;
Step 9: if not background model or shade, then be judged as foreground, s foreground target (B is obtained1(x,y),...,Bi(x,
y),...,Bs(x,y));
Classification is identified Step 10: the foreground target of acquisition is sent into the trained effective recognition classifiers of ELM, and then right
The quantity of fish, type, size are counted.
2. the underwater fish target detection and identification method according to claim 1 based on fish finder, it is characterised in that step
Background template pixel c described in rapid sevenmThe update rule that neighboring pixel is updated is as follows,
Wherein, α (t) is constant, ωi,jWhen being background modeling, the corresponding weight of each pixel, is initially Gauss weight in background model,
AtTo update rear backdrop template, At-1For background template before.
3. the underwater fish target detection and identification method according to claim 1 based on fish finder, it is characterised in that step
The judgment rule that shade judgement is carried out described in rapid eight is as follows,
Wherein remove pt, ciOuter is all parameter preset.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810226818.3A CN108520511A (en) | 2018-03-19 | 2018-03-19 | A kind of underwater fish target detection and identification method based on fish finder |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810226818.3A CN108520511A (en) | 2018-03-19 | 2018-03-19 | A kind of underwater fish target detection and identification method based on fish finder |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108520511A true CN108520511A (en) | 2018-09-11 |
Family
ID=63432882
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810226818.3A Pending CN108520511A (en) | 2018-03-19 | 2018-03-19 | A kind of underwater fish target detection and identification method based on fish finder |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108520511A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109471116A (en) * | 2018-11-13 | 2019-03-15 | 苏州热工研究院有限公司 | A kind of underwater sound detection regulation device |
CN110046619A (en) * | 2019-04-18 | 2019-07-23 | 马杰 | The full-automatic shoal of fish detection method of unmanned fish finding ship and system, unmanned fish finding ship and storage medium |
CN110751606A (en) * | 2019-10-17 | 2020-02-04 | 广州愿托科技有限公司 | Foam image processing method and system based on neural network algorithm |
CN110764093A (en) * | 2019-09-30 | 2020-02-07 | 苏州佳世达电通有限公司 | Underwater biological identification system and method thereof |
CN112465778A (en) * | 2020-11-26 | 2021-03-09 | 江苏国和智能科技有限公司 | Underwater fish shoal observation device and method |
CN112489108A (en) * | 2020-12-24 | 2021-03-12 | 中国科学院南海海洋研究所 | Method and device for inverting and reconstructing underwater surface area of open-sea coral reef |
CN112883773A (en) * | 2020-12-31 | 2021-06-01 | 中国水产科学研究院东海水产研究所 | Species discrimination method based on acoustic image data evaluation |
CN113569971A (en) * | 2021-08-02 | 2021-10-29 | 浙江索思科技有限公司 | Image recognition-based catch target classification detection method and system |
CN113642650A (en) * | 2021-08-16 | 2021-11-12 | 上海大学 | Multi-scale template matching and self-adaptive color screening based multi-beam sonar sunken ship detection method |
CN116977929A (en) * | 2023-07-31 | 2023-10-31 | 广西大学 | Population identification method and system based on fish water-taking behavior monitoring |
CN117572438A (en) * | 2024-01-12 | 2024-02-20 | 中国水产科学研究院南海水产研究所 | Navigation type fish shoal detection method and system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106682574A (en) * | 2016-11-18 | 2017-05-17 | 哈尔滨工程大学 | One-dimensional deep convolution network underwater multi-target recognition method |
CN106772331A (en) * | 2016-12-13 | 2017-05-31 | 中国电子科技集团公司第三研究所 | Target identification method and Target Identification Unit |
CN106845396A (en) * | 2017-01-18 | 2017-06-13 | 南京理工大学 | Illegal fishing Activity recognition method based on automated graphics identification |
CN107561545A (en) * | 2017-08-24 | 2018-01-09 | 北京臻迪科技股份有限公司 | A kind of method of fish finding, device and detector |
-
2018
- 2018-03-19 CN CN201810226818.3A patent/CN108520511A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106682574A (en) * | 2016-11-18 | 2017-05-17 | 哈尔滨工程大学 | One-dimensional deep convolution network underwater multi-target recognition method |
CN106772331A (en) * | 2016-12-13 | 2017-05-31 | 中国电子科技集团公司第三研究所 | Target identification method and Target Identification Unit |
CN106845396A (en) * | 2017-01-18 | 2017-06-13 | 南京理工大学 | Illegal fishing Activity recognition method based on automated graphics identification |
CN107561545A (en) * | 2017-08-24 | 2018-01-09 | 北京臻迪科技股份有限公司 | A kind of method of fish finding, device and detector |
Non-Patent Citations (2)
Title |
---|
LUCIA MADDALENA等: "A Self-Organizing Approach to Background Subtraction for Visual Surveillance Applications", 《IEEE TRANSACTIONS ON IMAGE PROCESSING》 * |
张训华等: "《海洋地质调查技术》", 31 December 2017 * |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109471116B (en) * | 2018-11-13 | 2020-11-13 | 苏州热工研究院有限公司 | Underwater acoustic detection regulation and control device |
CN109471116A (en) * | 2018-11-13 | 2019-03-15 | 苏州热工研究院有限公司 | A kind of underwater sound detection regulation device |
CN110046619B (en) * | 2019-04-18 | 2022-11-29 | 马杰 | Full-automatic fish school detection method and system for unmanned fish finding boat, unmanned fish finding boat and storage medium |
CN110046619A (en) * | 2019-04-18 | 2019-07-23 | 马杰 | The full-automatic shoal of fish detection method of unmanned fish finding ship and system, unmanned fish finding ship and storage medium |
CN110764093A (en) * | 2019-09-30 | 2020-02-07 | 苏州佳世达电通有限公司 | Underwater biological identification system and method thereof |
CN110751606A (en) * | 2019-10-17 | 2020-02-04 | 广州愿托科技有限公司 | Foam image processing method and system based on neural network algorithm |
CN110751606B (en) * | 2019-10-17 | 2023-09-19 | 广州愿托科技有限公司 | Foam image processing method and system based on neural network algorithm |
CN112465778A (en) * | 2020-11-26 | 2021-03-09 | 江苏国和智能科技有限公司 | Underwater fish shoal observation device and method |
CN112489108B (en) * | 2020-12-24 | 2023-08-01 | 中国科学院南海海洋研究所 | Method and device for reconstructing underwater surface area inversion of open sea coral reef |
CN112489108A (en) * | 2020-12-24 | 2021-03-12 | 中国科学院南海海洋研究所 | Method and device for inverting and reconstructing underwater surface area of open-sea coral reef |
CN112883773A (en) * | 2020-12-31 | 2021-06-01 | 中国水产科学研究院东海水产研究所 | Species discrimination method based on acoustic image data evaluation |
CN113569971B (en) * | 2021-08-02 | 2022-03-25 | 浙江索思科技有限公司 | Image recognition-based catch target classification detection method and system |
CN113569971A (en) * | 2021-08-02 | 2021-10-29 | 浙江索思科技有限公司 | Image recognition-based catch target classification detection method and system |
CN113642650A (en) * | 2021-08-16 | 2021-11-12 | 上海大学 | Multi-scale template matching and self-adaptive color screening based multi-beam sonar sunken ship detection method |
CN113642650B (en) * | 2021-08-16 | 2024-02-20 | 上海大学 | Multi-beam sonar sunken ship detection method based on multi-scale template matching and adaptive color screening |
CN116977929A (en) * | 2023-07-31 | 2023-10-31 | 广西大学 | Population identification method and system based on fish water-taking behavior monitoring |
CN116977929B (en) * | 2023-07-31 | 2024-06-11 | 广西大学 | Population identification method and system based on fish water-taking behavior monitoring |
CN117572438A (en) * | 2024-01-12 | 2024-02-20 | 中国水产科学研究院南海水产研究所 | Navigation type fish shoal detection method and system |
CN117572438B (en) * | 2024-01-12 | 2024-05-03 | 中国水产科学研究院南海水产研究所 | Navigation type fish shoal detection method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108520511A (en) | A kind of underwater fish target detection and identification method based on fish finder | |
CN109460754B (en) | A kind of water surface foreign matter detecting method, device, equipment and storage medium | |
CN105427342B (en) | A kind of underwater Small object sonar image target detection tracking method and system | |
CN108734111A (en) | SAR image surface vessel recognition methods | |
CN115761550A (en) | Water surface target detection method based on laser radar point cloud and camera image fusion | |
CN104021553B (en) | A kind of sonar image object detection method based on pixel layering | |
CN113484867B (en) | Method for detecting density of fish shoal in closed space based on imaging sonar | |
CN113591592B (en) | Overwater target identification method and device, terminal equipment and storage medium | |
CN114724022A (en) | Culture fish school detection method, system and medium fusing SKNet and YOLOv5 | |
CN106097320A (en) | Stichopus japonicus image automatic segmentation method and device under water | |
CN115187666A (en) | Deep learning and image processing combined side-scan sonar seabed elevation detection method | |
CN116740758A (en) | Bird image recognition method and system for preventing misjudgment | |
CN114066887B (en) | Rice chalkiness area detection method, device, equipment and storage medium | |
CN115512215A (en) | Underwater biological monitoring method and device and storage medium | |
CN114612454A (en) | Fish feeding state detection method | |
Shen et al. | Identification and counting of fish targets using adaptive resolution imaging sonar | |
CN113205139A (en) | Unmanned ship water sphere detection method based on density clustering | |
Sravanthi et al. | Efficient image-based object detection for floating weed collection with low cost unmanned floating vehicles | |
CN116883483A (en) | Fish body measuring method based on laser camera system | |
CN116243289A (en) | Unmanned ship underwater target intelligent identification method based on imaging sonar | |
CN114782368A (en) | Shrimp larva automatic counting method and system based on visual detection and fusion of multiple dynamic thresholds | |
CN115248436A (en) | Imaging sonar-based fish resource assessment method | |
CN113822105B (en) | Artificial intelligence water level monitoring system based on online two classifiers of SVM water scale | |
CN114140428A (en) | Method and system for detecting and identifying larch caterpillars based on YOLOv5 | |
CN114037737A (en) | Neural network-based offshore submarine fish detection and tracking statistical method |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180911 |
|
RJ01 | Rejection of invention patent application after publication |