Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Shellfish classification and identification method and device based on computer vision

A computer vision, classification and recognition technology, applied in the field of image recognition, can solve the problems of large amount of calculation, few species of shellfish recognized, poor generalization ability of neural network, etc., to eliminate useless features, improve calculation speed, and reduce the number of parameters Effect

Active Publication Date: 2021-04-09
MARINE BIOLOGY INST OF SHANDONG PROVINCE +1
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] The automatic recognition method of shellfish based on machine vision can identify few types of shellfish, and cannot realize the recognition of shellfish with similar morphological characteristics.
The generalization ability of neural network is poor, and the amount of calculation is large, so it is difficult to deploy on portable devices

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
  • Shellfish classification and identification method and device based on computer vision
  • Shellfish classification and identification method and device based on computer vision
  • Shellfish classification and identification method and device based on computer vision

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0039] A shellfish classification and recognition method based on computer vision. Taking the existing shellfish pictures as the research object, a step-by-step recognition method is proposed to identify its subjects, and then specifically identify its category. Finally, based on edge computing technology, the obtained model On-site identification of shellfish species deployed on our invented device. The size of the pictures collected by the equipment we use is 256*256. Technical scheme of the present invention will be divided into following four parts and set forth:

[0040] Shellfish subject recognition function based on DNN algorithm:

[0041] Considering that there are large differences in characteristics between the subjects of shellfish, it is easier to distinguish between categories. We decided to train a relatively simple deep neural network on samples of known subjects. The simple deep neural network proposed by the present invention achieves better compression and...

Embodiment 2

[0062] The identification device described in Example 1 is used to identify and classify 80 kinds of shellfish. The size of the pictures collected by the device is 256*256. The pre-training is completed on the server, the configuration is: CPU: i9 9820X, memory: 128G, GPU: Nvidia RTX 2080 Ti*2. 80,000 images are used for training. After the model is trained, after being processed by the pruning technology of the present invention, it is deployed on the embedded device, and the configuration is: CPU: RK3399 (CortexA72*2, CortexA54*4, Mali-T860GPU), memory: 4G. After testing, the recognition accuracy rate for 80 kinds of shellfish is 91%.

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 provides a shellfish classification and recognition method and device based on computer vision, belongs to the field of image recognition, and provides a step-by-step recognition method which is developed on the basis of python language, takes an existing shellfish picture as a research object, firstly recognizes a subject to which the shellfish picture belongs, and then specifically recognizes the category of the shellfish picture. In order to ensure the precision and the efficiency of a shellfish identification algorithm, the shellfish identification algorithm is deployed to portable equipment to carry out field identification on shellfish. According to the pruning technology taking entropy information and similarity as standards, a large number of useless and similar characteristics in a DNN (Deep Neural Network) are eliminated, so that the parameter quantity of a DNN model is reduced, the calculation speed of the DNN is increased, and the shellfish identification accuracy is improved. A device comprising the camera and the edge computing equipment supporting AI processing is designed in the invention, wherein the algorithm model obtained after pruning is operated to recognize shellfish categories on site, and data transmission time and cost are saved.

Description

technical field [0001] The invention belongs to the field of image recognition, in particular to a computer vision-based shellfish classification and recognition method and device. Background technique [0002] my country's marine resources are rich in shellfish resources, and it is one of the countries with the most abundant shellfish diversity in the world. It is difficult for ordinary people and even professionals to accurately distinguish shellfish shells only by naked eye observation. At present, the classification methods and automatic sorting systems for shellfish are still in their infancy, mainly reflected in the traditional shellfish classification methods. It is based on molecular biotechnology to distinguish from the genetic point of view, or to design a specific method based on Gabor transformation to collect features for distinction, and the identification of shellfish species is few and the accuracy is not high. At present, Deep Neural Networks (DNN for short) ...

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): G06K9/62G06K9/46G06N3/04G06N3/08
CPCG06N3/08G06V10/44G06N3/045G06F18/2415Y02A40/81
Inventor 吴莹莹宋爱环王英俊宋娴丽刘童贾世祥岳峻
Owner MARINE BIOLOGY INST OF SHANDONG PROVINCE
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products