CN111523612B - Deep-sea fish image classification and identification method - Google Patents
Deep-sea fish image classification and identification method Download PDFInfo
- Publication number
- CN111523612B CN111523612B CN202010379432.3A CN202010379432A CN111523612B CN 111523612 B CN111523612 B CN 111523612B CN 202010379432 A CN202010379432 A CN 202010379432A CN 111523612 B CN111523612 B CN 111523612B
- Authority
- CN
- China
- Prior art keywords
- fish
- deep
- similarity
- sea
- degree
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a deep-sea fish image classification and identification method which is characterized by comprising the following steps: step one, establishing a fish state multi-source information database according to the fin position information quantity, the fish size information quantity, the fish tail shape information quantity and the body color information quantity of deep-sea fishes, and dividing the state similarity of the fishes into 0-1; step two, establishing a three-layer neural network based on a BP algorithm; and step three, setting parameters of all items on the basis of the neural network tool, and training and verifying the neural network. The invention respectively makes four types of information scoring rules for fish category state evaluation, combines the scoring of the four types of information, adopts a neural network as an evaluation algorithm to obtain a comprehensive score which is used as a basis for fish classification, integrates various data and parameters, comprehensively and accurately evaluates the category of deep-sea fish, provides a scientific basis for deep-sea fish classification, can effectively develop and utilize deep-sea fish resources, and achieves the strategic significance of long-term development.
Description
Technical Field
The invention relates to the technical field of power grids, in particular to a deep-sea fish image classification and identification method.
Background
China is one of the world oceans, has a sea area of more than 300 million square kilometers, breeds extremely rich biological resources in the wide sea area of China, and according to statistics, the marine biological resources of China are as high as 20278 species, wherein about three thousand species of fishes account for about 20% of the fish species of the world, but the current marine resource development and utilization degree of China is low, and the overall level of marine economic development is not high;
the fish identification is the first step of marine fish resource detection and also an important basis for developing and utilizing marine resources, but the deep-sea fishes are different in shape and size, so that the task complexity is much higher compared with other identification tasks, and different varieties of the same fish usually have similar appearance, size and texture color, so that the identification difficulty of the fish is further increased, and therefore, deep-sea high-tech research, especially the deep-sea fish related technical research, has great strategic significance for effective development and utilization of marine biological resources and long-term development in China.
Disclosure of Invention
The invention aims to provide a deep-sea fish image classification and identification method to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: a deep-sea fish image classification and identification method comprises the following steps:
step one, establishing a fish state multi-source information database according to the fin position information quantity, the fish size information quantity, the fish tail shape information quantity and the body color information quantity of deep-sea fishes, and dividing the state similarity of the fishes into 0-1;
establishing a three-layer neural network based on a BP algorithm, wherein the three-layer neural network comprises an input layer, a hidden layer and an output layer; wherein the input layer: receiving external signals and data, and taking indexes of fin position information quantity, fish size information quantity, fish tail shape information quantity and body color information quantity as nodes of an input layer;
an output layer: a BP network model in an ANN network is adopted, qualitative output is converted into quantitative output through a BP network neural model, and a set output result is comprehensively evaluated;
hidden layer: constructing a network according to a neuron number algorithm, and establishing a simple and efficient model;
setting parameters of various items on the basis of a neural network tool, and training and verifying the neural network;
step four, establishing an image acquisition module; the system is used for acquiring image data of deep-sea fishes and uploading the shot data image data to a multi-source information database for further processing;
introducing data of fin position information quantity, fish size information quantity, fish tail shape information quantity and body color information quantity into the trained neural network to obtain relevant numerical values of P1, P2, P3 and P4;
sixthly, calculating according to the obtained numerical values of P1, P2, P3 and P4 to obtain specific values of each state quantity of the fish evaluation at the time;
step seven, inputting specific values of the state quantities of the state evaluation, and outputting a comprehensive evaluation result P;
and step eight, inputting the obtained result P into a multi-source information database to perform classification and identification on the deep-sea fishes.
Preferably, the state similarity can be divided into five states: degree I similarity is divided into 0-0.2 degree I similarity, degree II similarity is divided into 0.2-0.4 degree II similarity, degree III similarity is divided into 0.4-0.6 degree III similarity, degree V similarity is divided into 0.6-0.8 degree V similarity, and degree VI similarity is divided into 0.8-1 degree VI similarity.
Preferably, in step three, the parameters are correction rate, training times and training target error.
Preferably, in the fourth step, the image needs to be preprocessed, and after the acquired fish image is preprocessed, the image reading, the size adjustment and the data conversion are performed.
Preferably, in step four, the image acquisition module includes: the device comprises an underwater camera equipment, an instrument interface, an underwater connection module, a data caching and forwarding module and a decomposition framing preprocessing module.
Preferably, in step eight, the multi-source information database obtains the result and displays the result, for example: fish A is 0.5, fish B is 0.63, fish C is 0.72, and fish D is 0.82.
The invention provides a deep-sea fish image classification and identification method, which has the beneficial effects that: by adopting a multi-index scoring method, a standard of comprehensive score classification is formulated by giving a standard of an evaluation system of multi-index scoring of fishes, four information scoring rules of fish category state evaluation are formulated respectively according to the characteristics of four information of deep-sea fishes, a neural network is adopted as an evaluation algorithm by combining the scoring of the four information to obtain comprehensive scores, the comprehensive scores are used as a basis of fish classification, various data and parameters are integrated, the categories of the deep-sea fishes are comprehensively and accurately evaluated, scientific bases are provided for deep-sea fish classification, deep-sea fish resources can be effectively developed and utilized, and the strategic significance of long-term development is achieved.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
The invention provides a technical scheme that: a deep-sea fish image classification and identification method comprises the following steps: step one, establishing a fish state multi-source information database according to the fin position information quantity, the fish size information quantity, the fish tail shape information quantity and the body color information quantity of deep-sea fishes, dividing the state similarity of the fishes into 0-1, and dividing the state similarity into five states: 0-0.2 is similar for degree I, 0.2-0.4 is similar for degree II, 0.4-0.6 is similar for degree III, 0.6-0.8 is similar for degree V, and 0.8-1 is similar for degree VI; establishing a three-layer neural network based on a BP algorithm, wherein the three-layer neural network comprises an input layer, a hidden layer and an output layer; wherein the input layer: receiving external signals and data, and taking indexes of fin position information amount, fish size information amount, fish tail shape information amount and body color information amount as nodes of an input layer; an output layer: a BP network model in an ANN network is adopted, qualitative output is converted into quantitative output through a BP network neural model, and a set output result is comprehensively evaluated; hidden layer: constructing a network according to a neuron number algorithm, and establishing a simple and efficient model; setting parameters of each item on the basis of a neural network tool, and training and verifying the neural network, wherein the parameters are correction rate, training times and training target error; step four, establishing an image acquisition module; the image acquisition module is used for acquiring deep-sea fish image data, uploading the shot data image data to a multi-source information database for further processing, preprocessing the image, reading the image, adjusting the size and converting the data after preprocessing the collected fish image, and comprises: the system comprises an underwater camera device, an instrument interface, an underwater connection module, a data caching and forwarding module and a decomposition framing preprocessing module; introducing data of fin position information quantity, fish size information quantity, fish tail shape information quantity and body color information quantity into the trained neural network to obtain relevant numerical values of P1, P2, P3 and P4; sixthly, calculating according to the obtained numerical values of P1, P2, P3 and P4 to obtain specific values of the state quantities evaluated by the fishes at this time; step seven, inputting specific values of the state quantities of the state evaluation, and outputting a comprehensive evaluation result P; step eight, inputting the obtained result P into a multi-source information database for classification and identification of the deep-sea fishes, and displaying the result in the multi-source information database, wherein the steps are as follows: fish A is 0.5, fish B is 0.63, fish C is 0.72, and fish D is 0.82.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (6)
1. A deep-sea fish image classification and identification method is characterized by comprising the following steps:
step one, establishing a fish state multi-source information database according to the fin position information quantity, the fish size information quantity, the fish tail shape information quantity and the body color information quantity of deep-sea fishes, and dividing the state similarity of the fishes into 0-1;
establishing a three-layer neural network based on a BP algorithm, wherein the three-layer neural network comprises an input layer, a hidden layer and an output layer; wherein
An input layer: receiving external signals and data, and taking indexes of fin position information amount, fish size information amount, fish tail shape information amount and body color information amount as nodes of an input layer;
an output layer: a BP network model in an ANN network is adopted, qualitative output is converted into quantitative output through a BP network neural model, and a set output result is comprehensively evaluated;
hidden layer: constructing a network according to a neuron number algorithm, and establishing a simple and efficient model;
setting parameters of various items on the basis of a neural network tool, and training and verifying the neural network;
step four, establishing an image acquisition module; the system is used for acquiring image data of deep-sea fishes and uploading the shot data image data to a multi-source information database for further processing;
introducing data of fin position information quantity, fish size information quantity, fish tail shape information quantity and body color information quantity into the trained neural network to obtain relevant numerical values of P1, P2, P3 and P4;
sixthly, calculating according to the obtained numerical values of P1, P2, P3 and P4 to obtain specific values of each state quantity of the fish evaluation at the time;
step seven, inputting specific values of the state quantities of the state evaluation, and outputting a comprehensive evaluation result P;
and step eight, inputting the obtained result P into a multi-source information database to perform classification and identification on the deep-sea fishes.
2. The deep-sea fish image classification and identification method according to claim 1, characterized in that: in the first step, the state similarity can be divided into five states: degree I similarity is divided into 0-0.2 degree I similarity, degree II similarity is divided into 0.2-0.4 degree II similarity, degree III similarity is divided into 0.4-0.6 degree III similarity, degree V similarity is divided into 0.6-0.8 degree V similarity, and degree VI similarity is divided into 0.8-1 degree VI similarity.
3. The deep-sea fish image classification and identification method according to claim 1, characterized in that: in the third step, the parameters are correction rate, training times and training target error.
4. The deep-sea fish image classification and identification method according to claim 1, characterized in that: and in the fourth step, the image needs to be preprocessed, and after the acquired fish image is preprocessed, the image is read, the size is adjusted and the data is converted.
5. The deep-sea fish image classification and identification method according to claim 1, characterized in that: in step four, the image acquisition module comprises: the device comprises an underwater camera equipment, an instrument interface, an underwater connection module, a data caching and forwarding module and a decomposition framing preprocessing module.
6. The deep-sea fish image classification and identification method according to claim 1, characterized in that: in the eighth step, the multi-source information database obtains the result and displays the result, for example: fish A is 0.5, fish B is 0.63, fish C is 0.72, and fish D is 0.82.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010379432.3A CN111523612B (en) | 2020-05-07 | 2020-05-07 | Deep-sea fish image classification and identification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010379432.3A CN111523612B (en) | 2020-05-07 | 2020-05-07 | Deep-sea fish image classification and identification method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111523612A CN111523612A (en) | 2020-08-11 |
CN111523612B true CN111523612B (en) | 2022-06-10 |
Family
ID=71905457
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010379432.3A Active CN111523612B (en) | 2020-05-07 | 2020-05-07 | Deep-sea fish image classification and identification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111523612B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113269078B (en) * | 2021-05-20 | 2022-04-19 | 中国水利水电科学研究院 | Unmanned aerial vehicle-based river fish resource amount statistical method and system |
CN113758577B (en) * | 2021-09-09 | 2022-04-22 | 易科捷(武汉)生态科技有限公司成都分公司 | Accurate type fish identification technology system based on big data |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102162861A (en) * | 2010-12-07 | 2011-08-24 | 桂林电子科技大学 | Method and device for detecting underwater object based on terahertz imaging |
CN103400147A (en) * | 2013-07-25 | 2013-11-20 | 中国农业大学 | Image processing-based target fish recognition method and system |
CN103778500A (en) * | 2012-10-22 | 2014-05-07 | 波音公司 | Water area management system |
CN109902761A (en) * | 2019-03-18 | 2019-06-18 | 上海海洋大学 | A kind of fishing feelings prediction technique based on marine environment Factor Fusion and deep learning |
CN110084285A (en) * | 2019-04-08 | 2019-08-02 | 安徽艾睿思智能科技有限公司 | Fish fine grit classification method based on deep learning |
CN110414554A (en) * | 2019-06-18 | 2019-11-05 | 浙江大学 | One kind being based on the improved Stacking integrated study fish identification method of multi-model |
CN110766013A (en) * | 2019-09-25 | 2020-02-07 | 浙江农林大学 | Fish identification method and device based on convolutional neural network |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10599922B2 (en) * | 2018-01-25 | 2020-03-24 | X Development Llc | Fish biomass, shape, and size determination |
US10534967B2 (en) * | 2018-05-03 | 2020-01-14 | X Development Llc | Fish measurement station keeping |
-
2020
- 2020-05-07 CN CN202010379432.3A patent/CN111523612B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102162861A (en) * | 2010-12-07 | 2011-08-24 | 桂林电子科技大学 | Method and device for detecting underwater object based on terahertz imaging |
CN103778500A (en) * | 2012-10-22 | 2014-05-07 | 波音公司 | Water area management system |
CN103400147A (en) * | 2013-07-25 | 2013-11-20 | 中国农业大学 | Image processing-based target fish recognition method and system |
CN109902761A (en) * | 2019-03-18 | 2019-06-18 | 上海海洋大学 | A kind of fishing feelings prediction technique based on marine environment Factor Fusion and deep learning |
CN110084285A (en) * | 2019-04-08 | 2019-08-02 | 安徽艾睿思智能科技有限公司 | Fish fine grit classification method based on deep learning |
CN110414554A (en) * | 2019-06-18 | 2019-11-05 | 浙江大学 | One kind being based on the improved Stacking integrated study fish identification method of multi-model |
CN110766013A (en) * | 2019-09-25 | 2020-02-07 | 浙江农林大学 | Fish identification method and device based on convolutional neural network |
Non-Patent Citations (4)
Title |
---|
《Fish Recognition Based on Robust Features Extraction from Size and Shape Measurements Using Neural Network 》;Mutasem Khalil Alsmadi等;《Journal of Computer Science》;20101231;第6卷(第10期);第1088-1094页 * |
《基于对抗双线性的细粒度鱼类图像分类方法》;刘建明 等;《桂林电子科技大学学报》;20200825;第40卷(第04期);第316-320页 * |
《基于机器视觉的鱼类行为特征提取与分析》;贾贝贝 等;《生态毒理学报》;20171015;第12卷(第05期);第193-203页 * |
《基于深度学习的海底观测视频中鱼类的识别方法》;张俊龙 等;《计算机应用》;20190210;第39卷(第02期);第376-381页 * |
Also Published As
Publication number | Publication date |
---|---|
CN111523612A (en) | 2020-08-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Liu et al. | Tomato diseases and pests detection based on improved Yolo V3 convolutional neural network | |
CN111523612B (en) | Deep-sea fish image classification and identification method | |
CN111400469A (en) | Intelligent generation system and method for voice question answering | |
CN111127423B (en) | Rice pest and disease identification method based on CNN-BP neural network algorithm | |
CN111723239B (en) | Video annotation method based on multiple modes | |
CN109800795A (en) | A kind of fruit and vegetable recognition method and system | |
CN113887517B (en) | Crop remote sensing image semantic segmentation method based on parallel attention mechanism | |
CN113470076B (en) | Multi-target tracking method for yellow feather chickens in flat raising chicken house | |
Han et al. | Research on multiple jellyfish classification and detection based on deep learning | |
CN114627467A (en) | Rice growth period identification method and system based on improved neural network | |
Sun et al. | YOLO-P: An efficient method for pear fast detection in complex orchard picking environment | |
CN106611016B (en) | A kind of image search method based on decomposable word packet model | |
Wang et al. | DualSeg: Fusing transformer and CNN structure for image segmentation in complex vineyard environment | |
Yuan et al. | Sensitivity examination of YOLOv4 regarding test image distortion and training dataset attribute for apple flower bud classification | |
Zhang et al. | Knowledge amalgamation for object detection with transformers | |
Jia et al. | Polar-Net: Green fruit instance segmentation in complex orchard environment | |
Zhang et al. | Multi-class detection of cherry tomatoes using improved Yolov4-tiny model | |
CN115828758B (en) | Seawater three-dimensional prediction method and system based on improved firework algorithm optimization network | |
Zhao et al. | A target detection algorithm for remote sensing images based on a combination of feature fusion and improved anchor | |
Zhang et al. | Remote sensing image retrieval based on DenseNet model and CBAM | |
CN112364160A (en) | Patent text classification method combining ALBERT and BiGRU | |
Liu et al. | Generating underwater images by GANs and similarity measurement | |
An | Xception network for weather image recognition based on transfer learning | |
Wang et al. | Identification of growing points of cotton main stem based on convolutional neural network | |
CN114937153A (en) | Neural network-based visual feature processing system and method under weak texture environment |
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 |