Radar-unmanned aerial vehicle dual-source data detection model adaptation method based on YOLOv8 visualized training and electronic device

CN122368828APending Publication Date: 2026-07-10DAQING ANRUIDA TECH DEV CO LTD
View PDF 0 Cites -1 Cited by

Patent Information

Application Number
CN202610486985.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-14
Publication Date
2026-07-10

AI Technical Summary

Technical Problem

Existing radar-UAV dual-source data detection models suffer from differences in data format and feature dimensions, lack efficient collaborative processing mechanisms, are complex to operate, have uncontrollable training processes, and lack hardware adaptation flexibility, resulting in poor fusion detection performance and low training efficiency.

Method used

A visualization training method based on YOLOv8 is adopted, which configures dual-source data through a graphical interface to achieve preprocessing, spatiotemporal alignment and hardware adaptive training of radar-UAV data. It supports breakpoint resume training and real-time monitoring, and dynamically adjusts training strategies to adapt to different hardware environments.

Benefits of technology

It improves the accuracy of dual-source data fusion, lowers the operational threshold, enhances the reliability of the training process and the versatility of the hardware environment, improves detection performance and training efficiency, and solves the integrated problem in traditional model training.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122368828A_ABST
    Figure CN122368828A_ABST
Patent Text Reader

Abstract

This invention proposes a radar-UAV dual-source data detection model adaptation method and electronic device based on YOLOv8 visualization training, solving the problems of poor dual-source data adaptability, complex training operations, uncontrollable processes, and insufficient hardware adaptation in existing technologies. The invention includes: starting the visualization system and configuring dual-source data and training parameters; standardizing radar data and enhancing image data, performing millisecond-level time alignment and spatial mapping, adaptively allocating fusion weights, and loading the YOLOv8 model to adapt to the hardware environment; starting training and monitoring logs and curves in real time, supporting breakpoint resume training and optimal model saving; loading the model to infer detection results from the dual-source data; evaluating performance and archiving the entire process data. This invention achieves efficient dual-source data collaboration and zero-code visualization training, applicable to UAV inspection, intelligent security, and low-altitude target control scenarios.
Need to check novelty before this filing date? Find Prior Art