Lightweight kiwi fruit detection method based on improved YOLOv8 model

By improving the YOLOv8 model, replacing key modules and operators, and building a lightweight kiwifruit detection method, the problem of insufficient accuracy and speed of kiwifruit detection in the existing technology is solved, and the effect of quickly and accurately identifying and positioning kiwifruit in complex environments is achieved, and automated picking is supported.

CN119723563AActive Publication Date: 2025-03-28WENZHOU UNIV
5 Cites 0 Cited by

Patent Information

Application Number
CN202411783747.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-06
Publication Date
2025-03-28
Estimated Expiration
2044-12-06

AI Technical Summary

Technical Problem

The existing kiwifruit detection methods are insufficient in the accuracy and speed of identifying and locating kiwifruit in complex environments, making it difficult to operate efficiently on mobile devices with limited computing resources, affecting the picking efficiency.

Method used

Using the improved YOLOv8 model, by replacing the C2f module as the EMBC module, the nearest neighbor upsampling module as the Dysample upsampling operator, and the probe head as the TADDH probe head, a lightweight kiwi detection method is built to improve the robustness and detection accuracy of the model.

Benefits of technology

It significantly improves the accuracy and efficiency of kiwifruit detection, can quickly and accurately identify target fruits in complex environments, and operate efficiently on mobile devices with limited computing resources, supporting automated picking.

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

Abstract

The invention discloses a lightweight kiwi fruit detection method based on an improved YOLOv8 model, and relates to the technical field of image recognition. Comprising the steps of obtaining an image of unpicked kiwi fruits, expanding and marking the image to form a kiwi fruit data set, replacing a C2f module in a YOLOv8 model with an EMBC module, replacing a nearest neighbor upsampling module in the YOLOv8 model with a Dysample upsampling operator, replacing an original edition detection head in a head network of the YOLOv8 model with a TADDH detection head, constructing an improved YOLOv8 model, inputting the kiwi fruit data set into the improved YOLOv8 model, and obtaining the kiwi fruit data set. And continuously iterating to obtain an optimal weight, obtaining a trained improved YOLOv8 model, and detecting kiwi fruits in a complex environment through the trained improved YOLOv8 model. Technical support can be effectively provided for follow-up kiwi fruit positioning and automatic picking.
Need to check novelty before this filing date? Find Prior Art