Pest detection method and system based on DiverseC2f and LDConv feature extraction

By embedding DiverseC2f and LDConv feature extraction modules in the YOLOv8 network, the existing pest detection methods are solved, and more efficient and accurate pest detection is achieved.

CN119741730APending Publication Date: 2025-04-01HUAIYIN INSTITUTE OF TECHNOLOGY
0 Cites 2 Cited by

Patent Information

Application Number
CN202411596605.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-11
Publication Date
2025-04-01

AI Technical Summary

Technical Problem

Existing pest detection methods rely on individual subjective cognition and experience, have low detection efficiency, and in complex backgrounds and mobile states, machine learning methods require high-quality images and manual feature extraction, affecting detection accuracy and speed.

Method used

Using pest detection method based on DiverseC2f and LDConv feature extraction, by improving the YOLOv8 network, replacing the C2f module with the DiverseC2f module and the CBS module with the LDConv feature extraction module, the feature expression and irregular convolution feature extraction capabilities of the convolution layer are enhanced.

Benefits of technology

It improves the efficiency and accuracy of pest detection, is suitable for complex and variable practical application scenarios, maintains calculation efficiency and generalization capabilities, and reduces the number of parameters and calculation overhead.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a pest detection method and system based on DiverseC2f and LDConv feature extraction, and the method comprises the steps: collecting a picture containing a pest target, making a data set, and dividing the data set into a training set, a test set and a verification set; improvement is carried out based on a YOLOv8 network, and a pest detection model is established; the improvement comprises the steps that a C2f module in the YOLOv8 network is replaced with a DiverseC2f module, and the DiverseC2f module is used for enhancing feature expression of a convolutional layer by fusing various branch structures; a CBS module of the YOLOv8 network is replaced by an LDConv feature extraction module, and the LDConv feature extraction module is used for completing an irregular convolution feature extraction process; training the pest detection model by using a training set and a test set, and detecting a pest target by using the trained pest detection model; according to the method, feature extraction can be efficiently carried out while the feature expression ability is enhanced, so that the model adapts to different forms and postures of pests, the detection robustness is improved, and the recognition accuracy and precision of the pests are improved.
Need to check novelty before this filing date? Find Prior Art