Fish behavior image recognition method based on improved AlexNet

An image recognition and fish technology, applied in the field of marine biological behavior image recognition and processing, can solve the problems of poor timeliness and low judgment accuracy, and achieve the effect of improving accuracy and good commercial application prospects.

Pending Publication Date: 2021-07-27
QINGDAO JIERUI IND CONTROL TECH CO LTD
View PDF2 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

At present, fish disease prevention and control mostly rely on indirect means such as water quality data monitoring and manual experience judgment, which are time-sensitive and the judgment accuracy is not high

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Fish behavior image recognition method based on improved AlexNet
  • Fish behavior image recognition method based on improved AlexNet
  • Fish behavior image recognition method based on improved AlexNet

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] Attached below Figure 1-3 The technology will be further described with specific examples to help understand the content of the present invention.

[0033] Such as figure 1 Shown, a kind of fish behavior image recognition method based on improved AlexNet, comprises the following steps:

[0034] S1. Establish an item-based flexible attention algorithm, which is used for the convolution function of the main part of the algorithm and the multi-layer perceptron, decoder and sampling layer of the branch part of the algorithm;

[0035] S2. Build a fish behavior image recognition sample library, use the assignment method to convert the fish behavior image recognition sample library into a picture text library R, ​​use the improved AlexNet model, put the item-based flexible attention module in the fully connected layer and output between layers, image 3 shown;

[0036] S3. Using the model migration method, the trained fish image recognition knowledge map is transplanted i...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a fish behavior image recognition method based on improved AlexNet, and the method comprises the following steps: S1, building an item-based flexible attention algorithm which is used for the convolution function of a trunk part of the algorithm and the building of a multi-layer sensor, a decoder and a sampling layer of a branch part of the algorithm; s2, constructing a fish behavior image recognition sample library, and placing an item-based flexible attention module between a full connection layer and an output layer by using an improved AlexNet model; and S3, transplanting the trained fish image recognition knowledge graph into the optimized AlexNet model by adopting a model migration method to obtain a fish behavior recognition result. The method has the advantages that a traditional fish behavior image recognition algorithm is improved, under the same recognition speed, the recognition precision is higher, the convergence speed is higher, meanwhile, over-convergence is avoided, and the method has actual engineering application value.

Description

technical field [0001] The invention relates to the field of marine biological behavior image recognition processing, in particular to a fish behavior image recognition method based on improved AlexNet. Background technique [0002] How to realize smart fish farming and promote the transformation of my country's traditional marine fishery model is an urgent problem to be solved in China. The internal and external environment of fishery development is undergoing profound changes: the dual constraints of resources and the environment are tightening, fishery resources are increasingly depleted, waters are seriously polluted, and endangered species increase; fishery development methods are extensive, facilities and equipment are backward, production costs are rising, and benefits continue to decline; aquatic biological diseases are increasing, and there are hidden dangers in quality and safety. Some long-term production and ecological contradictions have not been effectively reso...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/62G06F16/36G06N3/04
CPCG06F16/367G06V40/20G06N3/045G06F18/214
Inventor 赵奎贺保卫崔海朋韩兵兵沈炜皓赵巍
Owner QINGDAO JIERUI IND CONTROL TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap