A method for detecting plums

By acquiring, segmenting, and color-enhancing image data of sycamore fruits, and training a Dinov3 network model, the problems of high data calibration cost and insufficient robustness in sycamore fruit detection are solved, and efficient detection on embedded devices is achieved.

CN122244667APending Publication Date: 2026-06-19NANJING JITU NETWORK TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
NANJING JITU NETWORK TECH CO LTD
Filing Date
2026-02-12
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

The detection of sycamore fruits faces challenges such as high data calibration costs, insufficient robustness of models to seasonal changes, and difficulty in efficient deployment on embedded devices.

Method used

By collecting images of sycamore fruits under different environments and shooting conditions, instance segmentation and edge detection are performed to generate a fruit target detection dataset. Color enhancement processing is performed using color statistics, and a Dinov3 network model of ViT-S size is trained to achieve fruit detection.

Benefits of technology

It reduces data annotation costs, improves the model's robustness and generalization ability to seasonal changes, and enables accurate detection and recognition on embedded devices, meeting the real-time requirements of practical application scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122244667A_ABST
    Figure CN122244667A_ABST
Patent Text Reader

Abstract

This invention discloses a method for detecting sycamore fruits, belonging to the field of computer vision target detection technology. The method includes: acquiring sycamore fruit images under different environmental conditions and shooting conditions; performing instance segmentation on the sycamore fruit images to obtain fruit segmentation results; extracting fruit target contours based on the fruit segmentation results through edge detection to obtain a fruit target detection dataset; calculating color statistics based on the fruit target detection dataset; performing color enhancement processing on the fruit target detection dataset based on the color statistics to obtain an enhanced fruit target detection dataset; training a visual detection model using the enhanced fruit target detection dataset to obtain a fruit detection model; and inputting the image to be detected into the fruit detection model to obtain the sycamore fruit detection results. This invention can reduce data calibration costs, improve model generalization ability, and achieve efficient detection and recognition of fruits on embedded devices.
Need to check novelty before this filing date? Find Prior Art