Forestry pest identification method and system based on feature selection module

A feature selection and identification method technology, applied in the field of computer vision, can solve the problems of large differences in size and angle, low identification accuracy, and large changes in pest background information, so as to expand pest species, improve identification accuracy, and improve intelligent management. horizontal effect

Pending Publication Date: 2021-11-02
QILU UNIV OF TECH
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The inventors of the present disclosure have found that due to large changes in environmental factors, large changes in the background information of pests, and large differences in the size and angle of the pests in the image, it brings great challenges to the identification research of forestry pests; and the existing methods based on In the forestry pest identification method based on computer vision technology, the influence of factors such as pest background information on the identification results is not considered, and there is a problem of low identification accuracy.

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
  • Forestry pest identification method and system based on feature selection module
  • Forestry pest identification method and system based on feature selection module
  • Forestry pest identification method and system based on feature selection module

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0038] The present disclosure provides a method for identifying forestry pests based on a feature selection module, including:

[0039] Obtain images of forestry pests;

[0040] According to the obtained forestry pest image and the preset forestry pest identification network model, the forestry pest classification result is obtained, and the forestry pest identification is carried out;

[0041] Wherein, in the present embodiment, the ResNet feature extraction network selects the ResNet50 network, and the forestry pest recognition network model is obtained by expanding the forestry pest image after processing and the ResNet50 network training embedded with the feature selection module; the feature selection module is obtained by feature selection Units are combined with residual maps.

[0042] The ResNet residual network is a popular feature extraction network in recent years. The network creatively proposes a residual mapping structure, which effectively solves the problem of...

Embodiment 2

[0055] Such as figure 1 As shown, the present disclosure provides a method for identifying forestry pests based on a feature selection module, including:

[0056] 1. Obtain pictures of forestry pests and classify them according to the types of pests. Specifically, divide the collected pictures into a training set and a test set at a ratio of 4:1. The training set is used to train model parameters, and the test set is used to test the model. Calculation of parameter accuracy;

[0057] 2. Expand the data images in the divided training set;

[0058] In this embodiment, operations such as random flipping, increasing contrast, random grayscale processing, vertical mirror transformation, and horizontal mirror transformation are respectively used for data expansion; preferably, the original picture is randomly flipped from -180° to 180°, and the original The contrast of the image, the contrast floating ratio is 0.1, each image generates a grayscale image with a probability of 0.025...

Embodiment 3

[0071] The present embodiment provides a forestry pest identification system based on a feature selection module, including a data acquisition module and a species identification module;

[0072] The data acquisition module is configured to: acquire images of forestry pests;

[0073] The type identification module is configured to: obtain the forestry pest classification result according to the obtained forestry pest image and the preset forestry pest identification network model, and perform forestry pest identification;

[0074] Wherein, the forestry pest identification network model is obtained by expanding the processed forestry pest image training set and training the ResNet50 network embedded with the feature selection module; the feature selection module is obtained by combining the feature selection unit with the residual mapping.

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 forestry pest identification method and system based on a feature selection module. The method comprises the steps of obtaining a forestry pest image; according to the obtained forestry pest image and a preset forestry pest recognition network model, obtaining a forestry pest classification result, and performing forestry pest recognition; wherein the forestry pest recognition network model is obtained by training a ResNet network embedded with a feature selection module through an expanded forestry pest image training set; obtaining the feature selection module by combining a feature selection unit and residual mapping. According to the invention, a classic ResNet feature extraction network is combined with a designed feature selection module, so that the network can perform feature extraction on purpose, select pest feature information in input data and enhance the pest feature information, the problems of high complexity of a pest picture background and high similarity of part of pests can be effectively solved, and the identification precision is effectively improved.

Description

technical field [0001] The disclosure belongs to the technical field of computer vision, and in particular relates to a method and system for identifying forestry pests based on a feature selection module. Background technique [0002] The destruction of forest by forestry pests is obvious, and how to improve the level of forestry pest control has always been a research hotspot in society. [0003] The inventors of the present disclosure have found that due to large changes in environmental factors, large changes in the background information of pests, and large differences in the size and angle of the pests in the image, it brings great challenges to the identification research of forestry pests; and the existing methods based on In the forestry pest identification method carried out by computer vision technology, the influence of factors such as pest background information on the identification results is not considered, and there is a problem of low identification accurac...

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): G06K9/00G06K9/46G06K9/62G06N3/04G06N3/08
CPCG06N3/04G06N3/084G06F18/24G06F18/214
Inventor 张友梅冯富祥张瑜
Owner QILU UNIV OF TECH
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