CN116596163A - Image detection-based white croaker resource distribution prediction system and method - Google Patents
Image detection-based white croaker resource distribution prediction system and method Download PDFInfo
- Publication number
- CN116596163A CN116596163A CN202310815700.5A CN202310815700A CN116596163A CN 116596163 A CN116596163 A CN 116596163A CN 202310815700 A CN202310815700 A CN 202310815700A CN 116596163 A CN116596163 A CN 116596163A
- Authority
- CN
- China
- Prior art keywords
- data
- white
- maigre
- image
- fish
- 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
- 238000000034 method Methods 0.000 title claims abstract description 96
- 238000001514 detection method Methods 0.000 title claims abstract description 53
- 241001604174 Genyonemus lineatus Species 0.000 title claims abstract description 48
- 241000347391 Umbrina cirrosa Species 0.000 claims abstract description 240
- 241000251468 Actinopterygii Species 0.000 claims abstract description 145
- 230000000007 visual effect Effects 0.000 claims abstract description 85
- 238000012544 monitoring process Methods 0.000 claims abstract description 57
- 238000013135 deep learning Methods 0.000 claims abstract description 25
- 238000012549 training Methods 0.000 claims description 75
- 238000012360 testing method Methods 0.000 claims description 32
- 238000013528 artificial neural network Methods 0.000 claims description 24
- 239000003016 pheromone Substances 0.000 claims description 24
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 11
- 238000003384 imaging method Methods 0.000 claims description 10
- 238000010586 diagram Methods 0.000 claims description 8
- 150000003839 salts Chemical class 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 7
- 238000000354 decomposition reaction Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 2
- 230000009189 diving Effects 0.000 claims description 2
- 239000000523 sample Substances 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 3
- 241000257303 Hymenoptera Species 0.000 description 12
- 238000003672 processing method Methods 0.000 description 6
- 241000442132 Lactarius lactarius Species 0.000 description 5
- 230000008901 benefit Effects 0.000 description 5
- 230000003044 adaptive effect Effects 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000019637 foraging behavior Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000000050 nutritive effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S15/00—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems
- G01S15/86—Combinations of sonar systems with lidar systems; Combinations of sonar systems with systems not using wave reflection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S15/00—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems
- G01S15/88—Sonar systems specially adapted for specific applications
- G01S15/96—Sonar systems specially adapted for specific applications for locating fish
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Mining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/05—Underwater scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/80—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in fisheries management
- Y02A40/81—Aquaculture, e.g. of fish
Abstract
The invention discloses a method and a system for predicting the distribution of white croaker resources based on image detection, wherein the method comprises the following steps: the method comprises the steps that a detection ship is carried with a fish finder, an acoustic image is obtained through a navigation acoustic detection technology in the fish finder, a white maigre shoal acoustic image and a sonar database are built, a white maigre visual monitoring system is used for obtaining white maigre images, data are summarized to the center of a detection ship body, the white maigre is identified and classified in a deep learning algorithm by combining the acoustic image, the images and the gray level images, multiple scanning comparison is carried out, resource distribution prediction data are obtained by combining marine data and target identification data through an ant colony algorithm, and finally the distribution prediction data are visually displayed in a visual map mode. The invention can improve the positioning tracking and judgment of the operators on the white croaker resources and provide effective data support for fishery resource prediction.
Description
Technical Field
The invention relates to the technical field of fishery resource investigation and prediction, in particular to a white croaker resource distribution prediction system and method suitable for north south China sea.
Background
White maigre is a fish with higher economic benefit, and often appears on dining tables of people due to the advantages of high nutritive value, good taste in mouth and the like. The traditional method for fishing the white maigre consumes time and human resources, and the risk of misjudgment of the distribution position of the white maigre resources is obtained empirically, so that economic benefit is lost. Under the condition of bad weather, the wrong distribution position of the white croaker resources is resolved, so that fishers do idle work, the salvaging risk is greatly increased, and even serious economic loss is likely to happen. Therefore, the development of the rapid, efficient and accurate white croaker resource distribution prediction system has important practical application value, can greatly reduce the salvage burden of fishers, reduce the waste of human resources, improve the economic benefit, and can shorten the offshore time when directly going to the white croaker distribution dense area, thereby protecting the personal safety and the property safety.
Disclosure of Invention
The invention overcomes the defects of the prior art and provides a white croaker resource distribution prediction method and system based on image detection.
The technical scheme adopted by the invention for achieving the purpose is as follows:
the invention discloses a white croaker resource distribution prediction method based on image detection, which comprises the following steps of:
Using a fish finder to perform zigzag navigation, establishing a three-dimensional imaging sonar system to acquire an acoustic image, denoising the image, and storing the image into a sonar database;
the underwater robot carries a visual monitoring system and works together with the fish finder, the underwater robot shoots images of fish shoals under water, and collected data are summarized to an image data center of the visual monitoring system of the detection vessel;
based on a deep learning algorithm, combining sonar system data, visual monitoring system image data and gray level images, identifying the fish shoals of the white maigre and constructing a white maigre target identification model;
obtaining the flow prediction data of the fish shoal of the white maigre based on an ant colony algorithm and ocean data, and obtaining resource distribution data by training the flow prediction data of the fish shoal of the white maigre through a neural network;
and converting the data of the spotted maigre resource distribution prediction into a visual view of the spotted maigre resource distribution prediction by using data visualization software.
In this scheme, use the fish finder to carry out zigzag navigation, establish three-dimensional imaging sonar system and acquire acoustic image, store to sonar database after carrying out denoising processing with the image, specifically do:
the quantitative scientific fish finder with the acoustic data export interface and the GPS data interface is arranged on the detection vessel, geographic information data can be integrated in the acoustic data, and a frequency transducer is used and is arranged in the guide cover, so that bubbles generated by ocean currents are prevented from generating more noise in the data;
When the fish finder works, the energy converter is fixed on the starboard outside of the ship body and is not contacted with the ship body, so that the interference of mechanical noise generated by a ship running motor on the energy converter is avoided;
the navigation mode of each detection ship adopts zigzag navigation, pulse signals with the frequency of 120kHZ and pulse signals with the frequency of 1.35MHZ are respectively transmitted near the transducer position of the detection ship fish finder by means of a sonar head, a sector scanning area with the area of 45 degrees and the area of 1 degree is formed nearby, the number of single pulse acoustic beams is 265, the single pulse acoustic beams are uniformly arranged in the vertical direction according to the same interval, finally 0.178 degrees is formed, all reflected signals of a target fish swarm are synthesized into a 2D plan, and a computer enables a cloud deck to rotate for 360 degrees based on a control function to collect data, so that omnibearing details of the target fish swarm are obtained, and a 3D image is generated;
and denoising speckle noise in the sonar image by adopting a method based on the combination of the rough set and wavelet transformation, performing wavelet decomposition on the image, reserving low-frequency wavelet coefficients, performing wavelet inverse transformation on the unprocessed low-frequency wavelet coefficients and the processed high-frequency wavelet coefficients to obtain a denoised subgraph, and storing the denoised subgraph in a sonar database.
In this scheme, the visual monitoring system is carried to the underwater robot, follows the sonde and works together, and the underwater robot carries out image shooting to the shoal under water to gather the data that gathers to survey ship data storage center, specifically does:
The underwater robot high-definition camera is provided with a visual monitoring system, and the underwater robot follows the fish finder to shoot fish shoal images, wherein the shooting diving depth is 100-200 meters;
the fish swarm image is processed by an underwater robot acquisition unit to acquire image data, and the method comprises the following steps: image memory size, image shooting longitude and latitude, and image resolution size;
the fish shoal image is obtained by the obtaining unit, and then a superior changing unit of similar data is found out from the computer relational database according to the image data;
and according to the superior change unit, modifying and updating the data in the acquisition unit, and summarizing the data to the image data center of the visual monitoring system of the probe ship.
In this scheme, based on the deep learning algorithm, combining sonar system data, visual monitoring system image data and gray level image, the white maigre shoals are identified and white maigre target identification models are constructed, specifically:
comparing whether the fish area is high-yield or not according to the unit fishing effort fishing yield historical data, and calculating a unit fishing effort fishing yield formula as follows:
；
wherein CUPE is used as a unit to Catch the fishing yield of the strength, catch is the total fishing yield in one fishing zone, and Effort is the total operation days of all ships in one fishing zone;
Constructing a white-spotted maigre target recognition model based on a deep learning algorithm, dividing the sonar data into a training data book and a test data book after combining the sonar data with the image data, inputting the training data book into the white-spotted maigre target recognition model, performing reverse training on the white-spotted maigre target recognition model training data book through a cross entropy loss function, and storing training parameters when an error value is smaller;
testing the training parameters of the white maigre target recognition model through a test data book, taking an average value of test results, and taking the parameters meeting preset requirements as final parameters of the white maigre resource recognition model if the data meet the preset requirements;
the method comprises the steps of carrying out graying treatment on image information summarized in a visual monitoring system and a three-dimensional sonar data diagram in a three-dimensional sonar database, wherein the graying treatment method adopts a weighted average method to enable three components of RGB to be averaged to obtain a reasonable gray image, and the graying treatment formula is as follows:
；
wherein Gray is Gray, R represents red, G represents green, and B represents blue;
the granularity of the image is reduced and the definition is increased through gray image processing, and a complete gray image is obtained by adopting a self-adaptive gray compensation method;
Comparing the gray level image, the acoustic image data and the visual monitoring system image to judge the type of the fish shoal;
and combining CUPE parameters, gray level images and final parameters of the white maigre resource identification model to construct a white maigre target identification model.
In this scheme, the ant colony algorithm and ocean data based on the prediction data of the fish shoal of white maigre, and then the prediction data of the fish shoal of white maigre is trained through neural network to obtain resource distribution data, specifically:
taking ocean data as a fish swarm trend judging factor, and adopting an ant colony algorithm to conduct position pre-judging on Bai Gu fish swarms, wherein the ocean data are water temperature and salt content data;
the method comprises the steps of using ocean data as an pheromone of an ant colony, comparing a white maigre colony with the ant colony, making a decision by the white maigre colony through the pheromone, generating information of the most dense white maigre resource, and obtaining flow prediction data of the white maigre colony after training by a neural network training set;
inputting the data of the flow prediction of the fish shoal of the white maigre into a white maigre target recognition model, recognizing the fish shoal of the white maigre and the flow range in the ocean according to the data of the flow prediction of the fish shoal of the white maigre, and obtaining complete data of the distribution prediction of the white maigre resources.
In this scheme, the data visualization software is used to transform the data of the resource distribution prediction of the white maigre into the visual view of the resource distribution prediction of the white maigre, specifically:
and importing the data for predicting the distribution of the white croaker resources into the data visualization software through the data visualization software, and taking a data average value after the computer data visualization software is identified and confirmed for a plurality of times to generate a visual view for predicting the distribution of the white croaker resources for operators to use.
The second aspect of the invention also provides a system for predicting the distribution of white croaker resources based on image detection, which comprises: the system comprises a memory and a processor, wherein the memory stores a white maigre resource distribution prediction method program, and when the white maigre resource distribution prediction method program is executed by the processor, the following steps are realized:
using a fish finder to perform zigzag navigation, establishing a three-dimensional imaging sonar system to acquire an acoustic image, denoising the image, and storing the image into a sonar database;
the underwater robot carries a visual monitoring system and works together with the fish finder, the underwater robot shoots images of fish shoals under water, and collected data are summarized to an image data center of the visual monitoring system of the detection vessel;
Based on a deep learning algorithm, combining sonar system data, visual monitoring system image data and gray level images, identifying the fish shoals of the white maigre and constructing a white maigre target identification model;
obtaining the flow prediction data of the fish shoal of the white maigre based on an ant colony algorithm and ocean data, and obtaining resource distribution data by training the flow prediction data of the fish shoal of the white maigre through a neural network;
and converting the data of the spotted maigre resource distribution prediction into a visual view of the spotted maigre resource distribution prediction by using data visualization software.
In this scheme, based on the deep learning algorithm, combining sonar system data, visual monitoring system image data and gray level image, the white maigre shoals are identified and white maigre target identification models are constructed, specifically:
comparing whether the fish area is high-yield or not according to the unit fishing effort fishing yield historical data, and calculating a unit fishing effort fishing yield formula as follows:
；
wherein CUPE is used as a unit to Catch the fishing yield of the strength, catch is the total fishing yield in one fishing zone, and Effort is the total operation days of all ships in one fishing zone;
constructing a white-spotted maigre target recognition model based on a deep learning algorithm, dividing the sonar data into a training data book and a test data book after combining the sonar data with the image data, inputting the training data book into the white-spotted maigre target recognition model, performing reverse training on the white-spotted maigre target recognition model training data book through a cross entropy loss function, and storing training parameters when an error value is smaller;
Testing the training parameters of the white maigre target recognition model through a test data book, taking an average value of test results, and taking the parameters meeting preset requirements as final parameters of the white maigre resource recognition model if the data meet the preset requirements;
the method comprises the steps of carrying out graying treatment on image information summarized in a visual monitoring system and a three-dimensional sonar data diagram in a three-dimensional sonar database, wherein the graying treatment method adopts a weighted average method to enable three components of RGB to be averaged to obtain a reasonable gray image, and the graying treatment formula is as follows:
；
wherein Gray is Gray, R represents red, G represents green, and B represents blue;
the granularity of the image is reduced and the definition is increased through gray image processing, and a complete gray image is obtained by adopting a self-adaptive gray compensation method;
comparing the gray level image, the acoustic image data and the visual monitoring system image to judge the type of the fish shoal;
and combining CUPE parameters, gray level images and final parameters of the white maigre resource identification model to construct a white maigre target identification model.
In this scheme, the ant colony algorithm and ocean data based on the prediction data of the fish shoal of white maigre, and then the prediction data of the fish shoal of white maigre is trained through neural network to obtain resource distribution data, specifically:
Taking ocean data as a fish swarm trend judging factor, and adopting an ant colony algorithm to conduct position pre-judging on Bai Gu fish swarms, wherein the ocean data are water temperature and salt content data;
the method comprises the steps of using ocean data as an pheromone of an ant colony, comparing a white maigre colony with the ant colony, making a decision by the white maigre colony through the pheromone, generating information of the most dense white maigre resource, and obtaining flow prediction data of the white maigre colony after training by a neural network training set;
inputting the data of the flow prediction of the fish shoal of the white maigre into a white maigre target recognition model, recognizing the fish shoal of the white maigre and the flow range in the ocean according to the data of the flow prediction of the fish shoal of the white maigre, and obtaining complete data of the distribution prediction of the white maigre resources.
In this scheme, the data visualization software is used to transform the data of the resource distribution prediction of the white maigre into the visual view of the resource distribution prediction of the white maigre, specifically:
and importing the data for predicting the distribution of the white croaker resources into the data visualization software through the data visualization software, and taking a data average value after the computer data visualization software is identified and confirmed for a plurality of times to generate a visual view for predicting the distribution of the white croaker resources for operators to use.
The invention discloses a method and a system for predicting the distribution of white croaker resources based on image detection, wherein the method comprises the following steps: the method comprises the steps that a detection ship is carried with a fish finder, an acoustic image is obtained through a navigation acoustic detection technology in the fish finder, a white maigre shoal acoustic image and a sonar database are built, a white maigre visual monitoring system is used for obtaining white maigre images, data are summarized to the center of a detection ship body, the white maigre is identified and classified in a deep learning algorithm by combining the acoustic image, the images and the gray level images, multiple scanning comparison is carried out, resource distribution prediction data are obtained by combining marine data and target identification data through an ant colony algorithm, and finally the distribution prediction data are visually displayed in a visual map mode. The invention can improve the positioning tracking and judgment of the operators on the white croaker resources and provide effective data support for fishery resource prediction.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other embodiments of the drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 shows a flow chart of a method for predicting the distribution of white croaker resources based on image detection;
FIG. 2 shows a flow chart of the recognition model construction of white croaker in accordance with the present application;
FIG. 3 shows a flowchart for obtaining the resource distribution forecast data of the white croaker in accordance with the present application;
FIG. 4 shows a block diagram of a system for predicting the distribution of white croaker resources based on image detection according to the present application.
Detailed Description
In order that the above-recited objects, features and advantages of the present application will be more clearly understood, a more particular description of the application will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, without conflict, the embodiments of the present application and features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application, however, the present application may be practiced in other ways than those described herein, and therefore the scope of the present application is not limited to the specific embodiments disclosed below.
FIG. 1 shows a flow chart of a method for predicting the distribution of white croaker resources based on image detection.
As shown in fig. 1, the first aspect of the present application provides a method for predicting distribution of white croaker resources based on image detection, including:
S102, performing zigzag navigation by using a fish finder, establishing a three-dimensional imaging sonar system to obtain an acoustic image, denoising the image, and storing the image into a sonar database;
s104, carrying a visual monitoring system by the underwater robot, working together with the fish finder, shooting images of fish shoals under water by the underwater robot, and gathering acquired data to an image data center of the visual monitoring system of the detection vessel;
s106, based on a deep learning algorithm, combining sonar system data, visual monitoring system image data and gray level images, identifying the fish shoals of the white maigre and constructing a white maigre target identification model;
s108, obtaining the flow prediction data of the fish shoal of the white maigre based on an ant colony algorithm and marine data, and obtaining resource distribution data by training the flow prediction data of the fish shoal of the white maigre through a neural network;
s110, utilizing data visualization software to convert the data of the resource distribution prediction of the white maigre into a visual view of the resource distribution prediction of the white maigre.
According to the embodiment of the invention, a fish finder is used for zigzag navigation, a three-dimensional imaging sonar system is established to acquire an acoustic image, and the image is stored in a sonar database after denoising, specifically:
The quantitative scientific fish finder with the acoustic data export interface and the GPS data interface is arranged on the detection vessel, geographic information data can be integrated in the acoustic data, and a frequency transducer is used and is arranged in the guide cover, so that bubbles generated by ocean currents are prevented from generating more noise in the data;
when the fish finder works, the energy converter is fixed on the starboard outside of the ship body and is not contacted with the ship body, so that the interference of mechanical noise generated by a ship running motor on the energy converter is avoided;
the navigation mode of each detection ship adopts zigzag navigation, pulse signals with the frequency of 120kHZ and pulse signals with the frequency of 1.35MHZ are respectively transmitted near the transducer position of the detection ship fish finder by means of a sonar head, a sector scanning area with the area of 45 degrees and the area of 1 degree is formed nearby, the number of single pulse acoustic beams is 265, the single pulse acoustic beams are uniformly arranged in the vertical direction according to the same interval, finally 0.178 degrees is formed, all reflected signals of a target fish swarm are synthesized into a 2D plan, and a computer enables a cloud deck to rotate for 360 degrees based on a control function to collect data, so that omnibearing details of the target fish swarm are obtained, and a 3D image is generated;
and denoising speckle noise in the sonar image by adopting a method based on the combination of the rough set and wavelet transformation, performing wavelet decomposition on the image, reserving low-frequency wavelet coefficients, performing wavelet inverse transformation on the unprocessed low-frequency wavelet coefficients and the processed high-frequency wavelet coefficients to obtain a denoised subgraph, and storing the denoised subgraph in a sonar database.
It should be noted that, the wavelet decomposition is to perform logarithmic transformation, and then perform noise variance estimation; compared with other noise reduction modes such as median filtering and mean filtering, the noise reduction edge is kept better, speckle noise can be better filtered, the 1.35MHz frequency emitted by the sonar head can be used for short-distance high-resolution detection, the 120kHZ frequency can be used for medium-distance detection, and the method can be used for locking a tracking target, so that the noise reduction processing effect of a sonar image is better, and the image is clearer.
According to the embodiment of the invention, based on a deep learning algorithm, by combining sonar system data, visual monitoring system image data and gray level images, the white maigre fish shoal is identified and a white maigre fish target identification model is constructed, as shown in fig. 2, fig. 2 shows a white maigre fish identification model construction flow chart, specifically:
s202, judging whether the white croaker in the area is high in yield or not by utilizing a CUPE formula;
s204, after a white croaker target recognition model is constructed based on a deep learning algorithm, sonar data and image data are tested and trained, and training parameters are saved;
s206, debugging the test data, and taking an average value to obtain the character eye recognition model parameters of the white maigre;
S208, gray processing is carried out on the image information collected in the visual monitoring system and the three-dimensional sonar image in the three-dimensional sonar database;
s210, combining CUPE parameters, gray level images and final parameters of the white maigre resource identification model to construct a white maigre target identification model.
Comparing whether the fish area is high-yield or not according to the unit fishing effort fishing yield historical data, and calculating a unit fishing effort fishing yield formula as follows:
；
wherein CUPE is used as a unit to Catch the fishing yield of the strength, catch is the total fishing yield in one fishing zone, and Effort is the total operation days of all ships in one fishing zone;
the method can calculate the area with the maximum unit fishing effort and fishing yield by using a formula, and judge whether the area is high in yield or not.
The gray scale formula for gray scale processing of the collected image information in the visual monitoring system and the three-dimensional sonar image in the three-dimensional sonar database is as follows:
；
where Gray represents Gray, R represents red, G represents green, and B represents blue.
The method can use a weighted average method formula to enable the image to carry out gray processing, and compared with a color image gray image, the gray image of the gray image occupies smaller memory and has higher running speed; the gray scale image may be followed by visually increasing contrast, highlighting the target area.
According to the embodiment of the invention, after the deep learning algorithm is based on the white croaker target recognition model, sonar data and image data are tested and trained, and training parameters are saved, specifically:
constructing a white-spotted maigre target recognition model based on a deep learning algorithm, dividing the sonar data into a training data book and a test data book after combining the sonar data with the image data, inputting the training data book into the white-spotted maigre target recognition model, performing reverse training on the white-spotted maigre target recognition model training data book through a cross entropy loss function, and storing training parameters when an error value is smaller;
it should be noted that, after the sonar data and the image data are combined, the sonar data are divided into a training data book and a test data book, and as the sonar data and the image data are continuously updated, the training data book and the test data book are also required to be continuously updated; the training data book of the white maigre target recognition model is reversely trained through the cross entropy loss function, when the error value is smaller, the training parameters are stored, when the error value is larger, the training needs to be carried out again, and when the error value is smaller, the training parameters are stored. The method can enable the white maigre target recognition model to be trained all the time, and errors can be reduced slowly along with the white maigre target recognition model, so that the white maigre target recognition model data are more accurate.
According to the embodiment of the invention, the test data are debugged, and an average value is taken to obtain the character eye recognition model parameters of the white croaker, which are specifically as follows:
testing the training parameters of the white maigre target recognition model through a test data book, taking an average value of test results, and taking the parameters meeting preset requirements as final parameters of the white maigre resource recognition model if the data meet the preset requirements;
it should be noted that, taking the average value requires attention to the magnitude of the average value, and if the data has a larger difference, the training parameters of the target recognition model of the white maigre are tested again. The error can be further reduced by the method of taking the average value, so that the data of the white maigre target recognition model is more accurate.
According to the embodiment of the invention, the gray processing is performed on the summarized image information in the visual monitoring system and the three-dimensional sonar image in the three-dimensional sonar database, specifically:
carrying out gray scale treatment on the summarized image information in the visual monitoring system and the three-dimensional sonar data graph in the three-dimensional sonar database, wherein the gray scale treatment method adopts a weighted average method to ensure that the RGB three components can be averaged to obtain a more reasonable gray scale image;
the granularity of the image is reduced and the definition is increased through gray image processing, and a complete gray image is obtained by adopting a self-adaptive gray compensation method;
Comparing the gray level image, the acoustic image data and the visual monitoring system image to judge the type of the fish shoal;
it should be noted that, the adaptive gray level compensation method has the condition of insufficient gray level compensation, and the gray level compensation needs to be manually performed when the condition of insufficient gray level compensation occurs, so as to improve the gray level degree of the picture; the gray scale processing method also comprises a maximum value method and an average value method, and three gray scale processing methods are used according to actual conditions; judging the type of the fish shoal needs to be obtained by combining a gray level image, an acoustic image and a visual monitoring system image through comparison, and the type of the fish shoal cannot be judged by only one image. The method can make the image gray higher, the image clearer, the memory smaller and the target area highlighted.
According to the embodiment of the invention, the flow prediction data of the fish shoal is obtained based on the ant colony algorithm and the ocean data, and then the flow prediction data of the fish shoal is trained by a neural network to obtain the resource distribution data, as shown in fig. 3, fig. 3 shows a flow chart for obtaining the flow of the flow prediction data of the fish shoal, specifically:
s302, inputting two kinds of ocean data of water temperature and salt content into a computer;
s304, performing white maigre pheromone training by using an ant colony algorithm in combination with ocean data to obtain white maigre shoal flow prediction data;
S306, combining the flow prediction data of the fish shoal of the white maigre and the prediction model of the white maigre to obtain the resource distribution prediction data of the white maigre.
According to the embodiment of the invention, the whitefish pheromone training is carried out by combining the ocean data and using an ant colony algorithm to obtain the whitefish shoal flow prediction data, which specifically comprises the following steps:
the method comprises the steps of comparing the white maigre colony with the ant colony by taking ocean data as an pheromone of the ant colony, making a decision by the white maigre colony through the pheromone, generating information of the most dense white maigre resource, and obtaining the flow prediction data of the white maigre colony after training by a neural network training set.
The principle of the ant colony algorithm is to simulate foraging behavior of ants in real life, the ants can leave pheromones on paths through which the ants pass in the movement process, and the ants can also sense the existence concentration of the pheromones so as to guide the moving direction of the ants, and each ant tends to move towards the direction with high concentration of the pheromones. The method forms a positive feedback phenomenon, almost all ants will select the same path to move over time (because the pheromone concentration of the path is far greater than that of other paths), so that the optimal solution of the problem is that the ocean data update frequency is high, the flow possibility of the white maigre is high, and the neural network needs to be trained continuously to obtain new flow prediction data of the white maigre. The accuracy of the flow prediction data of the fish shoal of the white maigre can be improved through the method.
According to the embodiment of the invention, the whitefish resource distribution prediction data is obtained by combining the whitefish shoal flow prediction data and the whitefish prediction model, and specifically comprises the following steps:
inputting the data of the flow prediction of the fish shoal of the white maigre into a white maigre target recognition model, distinguishing the fish shoal of the white maigre and the flow range in the ocean according to the data of the flow prediction of the fish shoal of the white maigre, and obtaining the complete data of the distribution prediction of the white maigre resources.
FIG. 4 shows a block diagram of a system for predicting the distribution of white croaker resources based on image detection according to the present invention.
Using a fish finder to perform zigzag navigation, establishing a three-dimensional imaging sonar system to acquire an acoustic image, denoising the image, and storing the image into a sonar database;
the underwater robot carries a visual monitoring system and works together with the fish finder, the underwater robot shoots images of fish shoals under water, and collected data are summarized to an image data center of the visual monitoring system of the detection vessel;
based on a deep learning algorithm, combining sonar system data, visual monitoring system image data and gray level images, identifying the fish shoals of the white maigre and constructing a white maigre target identification model;
obtaining the flow prediction data of the fish shoal of the white maigre based on an ant colony algorithm and ocean data, and obtaining resource distribution data by training the flow prediction data of the fish shoal of the white maigre through a neural network;
And converting the data of the spotted maigre resource distribution prediction into a visual view of the spotted maigre resource distribution prediction by using data visualization software.
According to the embodiment of the invention, based on a deep learning algorithm, the white maigre shoals are identified and a white maigre target identification model is constructed by combining sonar system data, visual monitoring system image data and gray level images, and the method specifically comprises the following steps:
comparing whether the fish area is high-yield or not according to the unit fishing effort fishing yield historical data, and calculating a unit fishing effort fishing yield formula as follows:
；
wherein CUPE is used as a unit to Catch the fishing yield of the strength, catch is the total fishing yield in one fishing zone, and Effort is the total operation days of all ships in one fishing zone;
constructing a white-spotted maigre target recognition model based on a deep learning algorithm, dividing the sonar data into a training data book and a test data book after combining the sonar data with the image data, inputting the training data book into the white-spotted maigre target recognition model, performing reverse training on the white-spotted maigre target recognition model training data book through a cross entropy loss function, and storing training parameters when an error value is smaller;
testing the training parameters of the white maigre target recognition model through a test data book, taking an average value of test results, and taking the parameters meeting preset requirements as final parameters of the white maigre resource recognition model if the data meet the preset requirements;
The method comprises the steps of carrying out graying treatment on image information summarized in a visual monitoring system and a three-dimensional sonar data diagram in a three-dimensional sonar database, wherein the graying treatment method adopts a weighted average method to enable three components of RGB to be averaged to obtain a reasonable gray image, and the graying treatment formula is as follows:
；
wherein Gray is Gray, R represents red, G represents green, and B represents blue;
the granularity of the image is reduced and the definition is increased through gray image processing, and a complete gray image is obtained by adopting a self-adaptive gray compensation method;
comparing the gray level image, the acoustic image data and the visual monitoring system image to judge the type of the fish shoal;
and combining CUPE parameters, gray level images and final parameters of the white maigre resource identification model to construct a white maigre target identification model.
It should be noted that, the adaptive gray level compensation method has the condition of insufficient gray level compensation, and the gray level compensation needs to be manually performed when the condition of insufficient gray level compensation occurs, so as to improve the gray level degree of the picture; the gray scale processing method also comprises a maximum value method and an average value method, and three gray scale processing methods are used according to actual conditions; judging the type of the fish shoal by combining a gray level image, an acoustic image and a visual monitoring system image, and judging the type of the fish shoal without singly relying on one image; the method can make the image gray higher, the image clearer, the memory smaller and the target area highlighted.
The cure was used as a unit of catch effort fishing gain, and the english language was called Catch Per Unit Effort, and the unit was 103kg/d.
According to the embodiment of the invention, the ant colony algorithm and the ocean data are based to obtain the flow prediction data of the fish shoal of the white maigre, and the flow prediction data of the fish shoal of the white maigre is trained by a neural network to obtain the resource distribution data, specifically:
taking ocean data as a fish swarm trend judging factor, and adopting an ant colony algorithm to conduct position pre-judging on Bai Gu fish swarms, wherein the ocean data are water temperature and salt content data;
the method comprises the steps of using ocean data as an pheromone of an ant colony, comparing a white maigre colony with the ant colony, making a decision by the white maigre colony through the pheromone, generating information of the most dense white maigre resource, and obtaining flow prediction data of the white maigre colony after training by a neural network training set;
inputting the data of the flow prediction of the fish shoal of the white maigre into a white maigre target recognition model, recognizing the fish shoal of the white maigre and the flow range in the ocean according to the data of the flow prediction of the fish shoal of the white maigre, and obtaining complete data of the distribution prediction of the white maigre resources.
It should be noted that, the ocean data update frequency is larger, the flow probability of the white maigre is larger, and the neural network needs to be continuously trained to obtain new flow prediction data of the white maigre. The accuracy of the flow prediction data of the fish shoal of the white maigre can be improved through the method.
According to the embodiment of the invention, the utilization of the data visualization software converts the data of the resource distribution prediction of the white maigre into the visual view of the resource distribution prediction of the white maigre, which is specifically as follows:
and importing the data for predicting the distribution of the white croaker resources into the data visualization software through the data visualization software, and taking a data average value after the computer data visualization software is identified and confirmed for a plurality of times to generate a visual view for predicting the distribution of the white croaker resources for operators to use.
The data visualization software is software for converting data into visual images such as visual bar charts, distribution charts, sector charts and the like, so that the white croaker resource distribution prediction data are displayed in front of operators in the form of visual images. By the method, operators can conveniently specify reasonable fishery policies and fishing plans.
The second aspect of the present invention also provides a system 4 for predicting the distribution of white croaker resources based on image detection, the system comprising: the system comprises a memory 41 and a processor 42, wherein the memory stores a white maigre resource distribution prediction method program, and when the white maigre resource distribution prediction method program is executed by the processor, the following steps are realized:
Using a fish finder to perform zigzag navigation, establishing a three-dimensional imaging sonar system to acquire an acoustic image, denoising the image, and storing the image into a sonar database;
the underwater robot carries a visual monitoring system and works together with the fish finder, the underwater robot shoots images of fish shoals under water, and collected data are summarized to an image data center of the visual monitoring system of the detection vessel;
based on a deep learning algorithm, combining sonar system data, visual monitoring system image data and gray level images, identifying the fish shoals of the white maigre and constructing a white maigre target identification model;
obtaining the flow prediction data of the fish shoal of the white maigre based on an ant colony algorithm and ocean data, and obtaining resource distribution data by training the flow prediction data of the fish shoal of the white maigre through a neural network;
and converting the data of the spotted maigre resource distribution prediction into a visual view of the spotted maigre resource distribution prediction by using data visualization software.
According to the embodiment of the invention, based on a deep learning algorithm, the white maigre shoals are identified and a white maigre target identification model is constructed by combining sonar system data, visual monitoring system image data and gray level images, and the method specifically comprises the following steps:
comparing whether the fish area is high-yield or not according to the unit fishing effort fishing yield historical data, and calculating a unit fishing effort fishing yield formula as follows:
；
Wherein CUPE is used as a unit to Catch the fishing yield of the strength, catch is the total fishing yield in one fishing zone, and Effort is the total operation days of all ships in one fishing zone;
constructing a white-spotted maigre target recognition model based on a deep learning algorithm, dividing the sonar data into a training data book and a test data book after combining the sonar data with the image data, inputting the training data book into the white-spotted maigre target recognition model, performing reverse training on the white-spotted maigre target recognition model training data book through a cross entropy loss function, and storing training parameters when an error value is smaller;
testing the training parameters of the white maigre target recognition model through a test data book, taking an average value of test results, and taking the parameters meeting preset requirements as final parameters of the white maigre resource recognition model if the data meet the preset requirements;
the method comprises the steps of carrying out graying treatment on image information summarized in a visual monitoring system and a three-dimensional sonar data diagram in a three-dimensional sonar database, wherein the graying treatment method adopts a weighted average method to enable three components of RGB to be averaged to obtain a reasonable gray image, and the graying treatment formula is as follows:
；
wherein Gray is Gray, R represents red, G represents green, and B represents blue;
The granularity of the image is reduced and the definition is increased through gray image processing, and a complete gray image is obtained by adopting a self-adaptive gray compensation method;
comparing the gray level image, the acoustic image data and the visual monitoring system image to judge the type of the fish shoal;
and combining CUPE parameters, gray level images and final parameters of the white maigre resource identification model to construct a white maigre target identification model.
It should be noted that, the adaptive gray level compensation method has the condition of insufficient gray level compensation, and the gray level compensation needs to be manually performed when the condition of insufficient gray level compensation occurs, so as to improve the gray level degree of the picture; the gray scale processing method also comprises a maximum value method and an average value method, and three gray scale processing methods are used according to actual conditions; judging the type of the fish shoal by combining a gray level image, an acoustic image and a visual monitoring system image, and judging the type of the fish shoal without singly relying on one image; the method can make the image gray higher, the image clearer, the memory smaller and the target area highlighted.
The cure was used as a unit of catch effort fishing gain, and the english language was called Catch Per Unit Effort, and the unit was 103kg/d.
According to the embodiment of the invention, the ant colony algorithm and the ocean data are based to obtain the flow prediction data of the fish shoal of the white maigre, and the flow prediction data of the fish shoal of the white maigre is trained by a neural network to obtain the resource distribution data, specifically:
Taking ocean data as a fish swarm trend judging factor, and adopting an ant colony algorithm to conduct position pre-judging on Bai Gu fish swarms, wherein the ocean data are water temperature and salt content data;
the method comprises the steps of using ocean data as an pheromone of an ant colony, comparing a white maigre colony with the ant colony, making a decision by the white maigre colony through the pheromone, generating information of the most dense white maigre resource, and obtaining flow prediction data of the white maigre colony after training by a neural network training set;
inputting the data of the flow prediction of the fish shoal of the white maigre into a white maigre target recognition model, recognizing the fish shoal of the white maigre and the flow range in the ocean according to the data of the flow prediction of the fish shoal of the white maigre, and obtaining complete data of the distribution prediction of the white maigre resources.
The principle of the ant colony algorithm is to simulate foraging behavior of ants in real life, the ants can leave pheromones on paths through which the ants pass in the movement process, and the ants can also sense the existence concentration of the pheromones so as to guide the moving direction of the ants, and each ant tends to move towards the direction with high concentration of the pheromones. The method forms a positive feedback phenomenon, almost all ants will select the same path to move over time (because the pheromone concentration of the path is far greater than that of other paths), so that the optimal solution of the problem is that the ocean data update frequency is high, the flow possibility of the white maigre is high, and the neural network needs to be trained continuously to obtain new flow prediction data of the white maigre. The accuracy of the flow prediction data of the fish shoal of the white maigre can be improved through the method.
According to the embodiment of the invention, the utilization of the data visualization software converts the data of the resource distribution prediction of the white maigre into the visual view of the resource distribution prediction of the white maigre, which is specifically as follows:
and importing the data for predicting the distribution of the white croaker resources into the data visualization software through the data visualization software, and taking a data average value after the computer data visualization software is identified and confirmed for a plurality of times to generate a visual view for predicting the distribution of the white croaker resources for operators to use.
The data visualization software is software for converting data into visual images such as visual bar charts, distribution charts, sector charts and the like, so that the white croaker resource distribution prediction data are displayed in front of operators in the form of visual images. By the method, operators can conveniently specify reasonable fishery policies and fishing plans.
The invention discloses a method and a system for predicting the distribution of white croaker resources based on image detection, wherein the method comprises the following steps: the method comprises the steps that a detection ship is carried with a fish finder, an acoustic image is obtained through a navigation acoustic detection technology in the fish finder, a white maigre shoal acoustic image and a sonar database are built, a white maigre visual monitoring system is used for obtaining white maigre images, data are summarized to the center of a detection ship body, the white maigre is identified and classified in a deep learning algorithm by combining the acoustic image, the images and the gray level images, multiple scanning comparison is carried out, resource distribution prediction data are obtained by combining marine data and target identification data through an ant colony algorithm, and finally the distribution prediction data are visually displayed in a visual map mode. The invention can improve the positioning tracking and judgment of the operators on the white croaker resources and provide effective data support for fishery resource prediction.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. The utility model provides a white croaker resource distribution prediction method based on image detection, which is characterized by comprising the following steps:
using a fish finder to perform zigzag navigation, establishing a three-dimensional imaging sonar system to acquire an acoustic image, denoising the image, and storing the image into a sonar database;
the underwater robot carries a visual monitoring system and works together with the fish finder, the underwater robot shoots images of fish shoals under water, and collected data are summarized to an image data center of the visual monitoring system of the detection vessel;
based on a deep learning algorithm, combining sonar system data, visual monitoring system image data and gray level images, identifying the fish shoals of the white maigre and constructing a white maigre target identification model;
obtaining the flow prediction data of the fish shoal of the white maigre based on an ant colony algorithm and ocean data, and obtaining resource distribution data by training the flow prediction data of the fish shoal of the white maigre through a neural network;
And converting the data of the spotted maigre resource distribution prediction into a visual view of the spotted maigre resource distribution prediction by using data visualization software.
2. The method for predicting the distribution of the white croaker resources based on image detection according to claim 1, wherein the zigzag navigation is performed by using a fish finder, a three-dimensional imaging sonar system is established to obtain an acoustic image, the image is denoised and stored in a sonar database, and the method specifically comprises the steps of:
the quantitative scientific fish finder with the acoustic data export interface and the GPS data interface is arranged on the detection vessel, geographic information data can be integrated in the acoustic data, and a frequency transducer is used and is arranged in the guide cover, so that bubbles generated by ocean currents are prevented from generating more noise in the data;
when the fish finder works, the energy converter is fixed on the starboard outside of the ship body and is not contacted with the ship body, so that the interference of mechanical noise generated by a ship running motor on the energy converter is avoided;
the navigation mode of each detection ship adopts zigzag navigation, pulse signals with the frequency of 120kHZ and pulse signals with the frequency of 1.35MHZ are respectively transmitted near the transducer position of the detection ship fish finder by means of a sonar head, a sector scanning area with the area of 45 degrees and the area of 1 degree is formed nearby, the number of single pulse acoustic beams is 265, the single pulse acoustic beams are uniformly arranged in the vertical direction according to the same interval, finally 0.178 degrees is formed, all reflected signals of a target fish swarm are synthesized into a 2D plan, and a computer enables a cloud deck to rotate for 360 degrees based on a control function to collect data, so that omnibearing details of the target fish swarm are obtained, and a 3D image is generated;
And denoising speckle noise in the sonar image by adopting a method based on the combination of the rough set and wavelet transformation, performing wavelet decomposition on the image, reserving low-frequency wavelet coefficients, performing wavelet inverse transformation on the unprocessed low-frequency wavelet coefficients and the processed high-frequency wavelet coefficients to obtain a denoised subgraph, and storing the denoised subgraph in a sonar database.
3. The image detection-based white croaker resource distribution prediction method according to claim 1, wherein the underwater robot carries a visual monitoring system and works together with a fish finder, and the underwater robot shoots images of fish shoals under water and gathers acquired data to a data storage center of a detection boat, specifically comprising:
the underwater robot high-definition camera is provided with a visual monitoring system, and the underwater robot follows the fish finder to shoot fish shoal images, wherein the shooting diving depth is 100-200 meters;
the fish swarm image is processed by an underwater robot acquisition unit to acquire image data, and the method comprises the following steps: image memory size, image shooting longitude and latitude, and image resolution size;
the fish shoal image is obtained by the obtaining unit, and then a superior changing unit of similar data is found out from the computer relational database according to the image data;
And according to the superior change unit, modifying and updating the data in the acquisition unit, and summarizing the data to the image data center of the visual monitoring system of the probe ship.
4. The method for predicting the distribution of the white maigre resources based on the image detection according to claim 1, wherein the method for predicting the distribution of the white maigre resources based on the image detection is characterized by combining sonar system data, visual monitoring system image data and gray level images to identify the white maigre fish shoals and construct a white maigre fish target identification model, and specifically comprises the following steps:
comparing whether the fish area is high-yield or not according to the unit fishing effort fishing yield historical data, and calculating a unit fishing effort fishing yield formula as follows:
；
wherein CUPE is the unit of Catch effort fishing gain, and Catch is the total fishing gain in a fishing zone; effort is the total operation days of all ships in a fishing zone;
constructing a white-spotted maigre target recognition model based on a deep learning algorithm, dividing the sonar data into a training data book and a test data book after combining the sonar data with the image data, inputting the training data book into the white-spotted maigre target recognition model, performing reverse training on the white-spotted maigre target recognition model training data book through a cross entropy loss function, and storing training parameters when an error value is smaller;
Testing the training parameters of the white maigre target recognition model through a test data book, taking an average value of test results, and taking the parameters meeting preset requirements as final parameters of the white maigre resource recognition model if the data meet the preset requirements;
the method comprises the steps of carrying out graying treatment on image information summarized in a visual monitoring system and a three-dimensional sonar data diagram in a three-dimensional sonar database, wherein the graying treatment method adopts a weighted average method to enable three components of RGB to be averaged to obtain a reasonable gray image, and the graying treatment formula is as follows:
；
wherein Gray is Gray, R represents red, G represents green, and B represents blue;
the granularity of the image is reduced and the definition is increased through gray image processing, and a complete gray image is obtained by adopting a self-adaptive gray compensation method;
comparing the gray level image, the acoustic image data and the visual monitoring system image to judge the type of the fish shoal;
and combining CUPE parameters, gray level images and final parameters of the white maigre resource identification model to construct a white maigre target identification model.
5. The method for predicting the distribution of the white maigre resources based on the image detection of claim 1, wherein the method for predicting the flow of the white maigre fish shoal based on the ant colony algorithm and the ocean data is characterized in that the method for predicting the flow of the white maigre fish shoal based on the ant colony algorithm and the ocean data is used for obtaining the distribution data of the white maigre fish shoal through training of a neural network, and specifically comprises the following steps:
Taking ocean data as a fish swarm trend judging factor, and adopting an ant colony algorithm to conduct position pre-judging on Bai Gu fish swarms, wherein the ocean data are water temperature and salt content data;
the method comprises the steps of using ocean data as an pheromone of an ant colony, comparing a white maigre colony with the ant colony, making a decision by the white maigre colony through the pheromone, generating information of the most dense white maigre resource, and obtaining flow prediction data of the white maigre colony after training by a neural network training set;
inputting the data of the flow prediction of the fish shoal of the white maigre into a white maigre target recognition model, recognizing the fish shoal of the white maigre and the flow range in the ocean according to the data of the flow prediction of the fish shoal of the white maigre, and obtaining complete data of the distribution prediction of the white maigre resources.
6. The method for predicting the distribution of the white maigre resources based on the image detection according to claim 1, wherein the data visualization software is used for converting the prediction data of the distribution of the white maigre resources into a prediction visual view of the distribution of the white maigre resources, specifically:
and importing the data for predicting the distribution of the white croaker resources into the data visualization software through the data visualization software, and taking a data average value after the computer data visualization software is identified and confirmed for a plurality of times to generate a visual view for predicting the distribution of the white croaker resources for operators to use.
7. The utility model provides a white croaker resource distribution prediction system based on image detection which characterized in that, this system includes: the system comprises a memory and a processor, wherein the memory stores a white maigre resource distribution prediction method program, and when the white maigre resource distribution prediction method program is executed by the processor, the following steps are realized:
using a fish finder to perform zigzag navigation, establishing a three-dimensional imaging sonar system to acquire an acoustic image, denoising the image, and storing the image into a sonar database;
the underwater robot carries a visual monitoring system and works together with the fish finder, the underwater robot shoots images of fish shoals under water, and collected data are summarized to an image data center of the visual monitoring system of the detection vessel;
based on a deep learning algorithm, combining sonar system data, visual monitoring system image data and gray level images, identifying the fish shoals of the white maigre and constructing a white maigre target identification model;
obtaining the flow prediction data of the fish shoal of the white maigre based on an ant colony algorithm and ocean data, and obtaining resource distribution data by training the flow prediction data of the fish shoal of the white maigre through a neural network;
and converting the data of the spotted maigre resource distribution prediction into a visual view of the spotted maigre resource distribution prediction by using data visualization software.
8. The system for predicting the distribution of the white croaker resources based on image detection according to claim 7, wherein the system for identifying the white croaker shoals and constructing a white croaker target identification model is based on a deep learning algorithm by combining sonar system data, visual monitoring system image data and gray level images, and specifically comprises the following steps:
comparing whether the fish area is high-yield or not according to the unit fishing effort fishing yield historical data, and calculating a unit fishing effort fishing yield formula as follows:
；
wherein CUPE is used as a unit to Catch the fishing yield of the strength, catch is the total fishing yield in one fishing zone, and Effort is the total operation days of all ships in one fishing zone;
constructing a white-spotted maigre target recognition model based on a deep learning algorithm, dividing the sonar data into a training data book and a test data book after combining the sonar data with the image data, inputting the training data book into the white-spotted maigre target recognition model, performing reverse training on the white-spotted maigre target recognition model training data book through a cross entropy loss function, and storing training parameters when an error value is smaller;
testing the training parameters of the white maigre target recognition model through a test data book, taking an average value of test results, and taking the parameters meeting preset requirements as final parameters of the white maigre resource recognition model if the data meet the preset requirements;
The method comprises the steps of carrying out graying treatment on image information summarized in a visual monitoring system and a three-dimensional sonar data diagram in a three-dimensional sonar database, wherein the graying treatment method adopts a weighted average method to enable three components of RGB to be averaged to obtain a reasonable gray image, and the graying treatment formula is as follows:
；
wherein Gray is Gray, R represents red, G represents green, and B represents blue;
the granularity of the image is reduced and the definition is increased through gray image processing, and a complete gray image is obtained by adopting a self-adaptive gray compensation method;
comparing the gray level image, the acoustic image data and the visual monitoring system image to judge the type of the fish shoal;
and combining CUPE parameters, gray level images and final parameters of the white maigre resource identification model to construct a white maigre target identification model.
9. The system for predicting the distribution of the white maigre resources based on the image detection of claim 7, wherein the flow prediction data of the white maigre is obtained based on an ant colony algorithm and ocean data, and the flow prediction data of the white maigre is trained by a neural network to obtain the distribution data of the resources, specifically:
taking ocean data as a fish swarm trend judging factor, and adopting an ant colony algorithm to conduct position pre-judging on Bai Gu fish swarms, wherein the ocean data are water temperature and salt content data;
The method comprises the steps of using ocean data as an pheromone of an ant colony, comparing a white maigre colony with the ant colony, making a decision by the white maigre colony through the pheromone, generating information of the most dense white maigre resource, and obtaining flow prediction data of the white maigre colony after training by a neural network training set;
inputting the data of the flow prediction of the fish shoal of the white maigre into a white maigre target recognition model, recognizing the fish shoal of the white maigre and the flow range in the ocean according to the data of the flow prediction of the fish shoal of the white maigre, and obtaining complete data of the distribution prediction of the white maigre resources.
10. The image detection-based white maigre resource distribution prediction system of claim 7, wherein the data visualization software is used for converting white maigre resource distribution prediction data into white maigre resource distribution prediction visual view, and specifically comprises:
and importing the data for predicting the distribution of the white croaker resources into the data visualization software through the data visualization software, and taking a data average value after the computer data visualization software is identified and confirmed for a plurality of times to generate a visual view for predicting the distribution of the white croaker resources for operators to use.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310815700.5A CN116596163B (en) | 2023-07-05 | 2023-07-05 | Image detection-based white croaker resource distribution prediction system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310815700.5A CN116596163B (en) | 2023-07-05 | 2023-07-05 | Image detection-based white croaker resource distribution prediction system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116596163A true CN116596163A (en) | 2023-08-15 |
| CN116596163B CN116596163B (en) | 2024-02-20 |
Family
ID=87604686
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310815700.5A Active CN116596163B (en) | 2023-07-05 | 2023-07-05 | Image detection-based white croaker resource distribution prediction system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116596163B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116821807A (en) * | 2023-08-30 | 2023-09-29 | 中国水产科学研究院南海水产研究所 | Machine vision-based fishery object identification and automatic recording method and system |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111406693A (en) * | 2020-04-23 | 2020-07-14 | 上海海洋大学 | Marine ranch fishery resource maintenance effect evaluation method based on bionic sea eels |
| KR102333428B1 (en) * | 2021-01-25 | 2021-12-02 | 주식회사 시리오 | Method, apparatus and computer program for detecting fish school using artificial intelligence |
| CN114494837A (en) * | 2022-01-06 | 2022-05-13 | 中国水产科学研究院南海水产研究所 | Intelligent density identification method and system for fishery resources |
| CN114994691A (en) * | 2022-05-31 | 2022-09-02 | 中国水产科学研究院东海水产研究所 | Visualization-based fishery resource cluster distribution analysis method |
| CN115100512A (en) * | 2022-05-16 | 2022-09-23 | 中国水产科学研究院南海水产研究所 | Monitoring, identifying and catching method and system for marine economic species and storage medium |
| CN115293658A (en) * | 2022-10-08 | 2022-11-04 | 中国水产科学研究院南海水产研究所 | Fishery resource planning method and system based on big data |
| CN115293402A (en) * | 2022-06-30 | 2022-11-04 | 闽南理工学院 | Fish condition forecasting method and related equipment |
-
2023
- 2023-07-05 CN CN202310815700.5A patent/CN116596163B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111406693A (en) * | 2020-04-23 | 2020-07-14 | 上海海洋大学 | Marine ranch fishery resource maintenance effect evaluation method based on bionic sea eels |
| KR102333428B1 (en) * | 2021-01-25 | 2021-12-02 | 주식회사 시리오 | Method, apparatus and computer program for detecting fish school using artificial intelligence |
| CN114494837A (en) * | 2022-01-06 | 2022-05-13 | 中国水产科学研究院南海水产研究所 | Intelligent density identification method and system for fishery resources |
| CN115100512A (en) * | 2022-05-16 | 2022-09-23 | 中国水产科学研究院南海水产研究所 | Monitoring, identifying and catching method and system for marine economic species and storage medium |
| CN114994691A (en) * | 2022-05-31 | 2022-09-02 | 中国水产科学研究院东海水产研究所 | Visualization-based fishery resource cluster distribution analysis method |
| CN115293402A (en) * | 2022-06-30 | 2022-11-04 | 闽南理工学院 | Fish condition forecasting method and related equipment |
| CN115293658A (en) * | 2022-10-08 | 2022-11-04 | 中国水产科学研究院南海水产研究所 | Fishery resource planning method and system based on big data |
Non-Patent Citations (1)
| Title |
|---|
| 于增知等: "松花江流域冷水鱼潜在分布预测模型研究", 水资源与水工程学报, vol. 28, no. 3, pages 48 - 54 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116821807A (en) * | 2023-08-30 | 2023-09-29 | 中国水产科学研究院南海水产研究所 | Machine vision-based fishery object identification and automatic recording method and system |
| CN116821807B (en) * | 2023-08-30 | 2024-01-09 | 中国水产科学研究院南海水产研究所 | Machine vision-based fishery object identification and automatic recording method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116596163B (en) | 2024-02-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Korneliussen | Acoustic target classification | |
| CN116596163B (en) | Image detection-based white croaker resource distribution prediction system and method | |
| CN110414396A (en) | A kind of unmanned boat perception blending algorithm based on deep learning | |
| CN110046619B (en) | Full-automatic fish school detection method and system for unmanned fish finding boat, unmanned fish finding boat and storage medium | |
| CN112766221A (en) | Ship direction and position multitask-based SAR image ship target detection method | |
| Moreno et al. | Echotrace classification and spatial distribution of pelagic fish aggregations around drifting fish aggregating devices (DFAD) | |
| CN110675410A (en) | Side-scan sonar sunken ship target unsupervised detection method based on selective search algorithm | |
| Al-Rawi et al. | Intensity normalization of sidescan sonar imagery | |
| CN115995041A (en) | Attention mechanism-based SAR image multi-scale ship target detection method and device | |
| Yin et al. | Sonar image target detection based on deep learning | |
| CN112859056B (en) | Remote early warning system and method for large marine organisms | |
| CN113837924A (en) | Water bank line detection method based on unmanned ship sensing system | |
| Zhou et al. | Deep denoising method for side scan sonar images without high-quality reference data | |
| CN110988888B (en) | Method and device for acquiring seabed information | |
| Zhao et al. | Detecting moving targets in active sonar echograph of harbor environment using high-order time lacunarity | |
| CN116562467A (en) | Marine fish target species identification and distribution prediction method and system | |
| Kristmundsson et al. | Fish monitoring in aquaculture using multibeam echosounders and machine learning. | |
| CN115248436A (en) | Imaging sonar-based fish resource assessment method | |
| Kracker et al. | Integration of fisheries acoustics surveys and bathymetric mapping to characterize midwater-seafloor habitats of US Virgin Islands and Puerto Rico (2008-2010) | |
| CN115018285A (en) | Storm surge and sea wave fine early warning system and early warning method | |
| CN112816956A (en) | Method and device for acquiring radar target information | |
| CN116542405B (en) | Image processing-based blue-round-clad resource distribution prediction method and system | |
| CN117572438B (en) | Navigation type fish shoal detection method and system | |
| TWI736446B (en) | Seabed edge detection method capable of filtering horizontal noise | |
| CN116562472B (en) | Method and system for identifying and predicting target species of middle-upper marine organisms |
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 |