A method and system for identifying a specific radiation source in a complex environment

By embedding Lite-SEAM and Det-LIA modules into the YOLOv8 model and introducing SupCon loss during the training phase, a CIOD-YOLO model was constructed. This solved the problems of insufficient feature extraction and false detection in radiation source identification under complex environments, and achieved higher identification accuracy and stability.

CN122244594APending Publication Date: 2026-06-19HANGZHOU DIANZI UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HANGZHOU DIANZI UNIV
Filing Date
2026-04-09
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing SEI technology struggles to effectively identify specific radiation sources in complex environments, especially when target signals and interference signals overlap, resulting in insufficient feature extraction, significant background interference, and the potential for unknown interference to be misdetected as known categories.

Method used

The CIOD-YOLO model is constructed by embedding the Lite-SEAM module into the YOLOv8 model to enhance discriminative feature extraction, introducing the Det-LIA module to suppress background interference, and introducing SupCon loss to optimize class boundaries during the training phase.

Benefits of technology

It improves the accuracy and stability of radiation source identification in complex environments, reduces the false detection rate of unknown interference, and enhances the model's detection capability in complex backgrounds.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122244594A_ABST
    Figure CN122244594A_ABST
Patent Text Reader

Abstract

This invention discloses a method and system for identifying specific radiation sources in complex environments. The method is as follows: Step 1: Collect mobile phone WiFi signals, construct an open set detection dataset for WiFi radiation sources in complex environments, and preprocess the dataset by dividing it into a training set, a validation set, and a test set; Step 2: Construct a dataset configuration file corresponding to the radiation source dataset, and write the data paths of the training set, validation set, and test set, as well as the known radiation source target category information, into the configuration file; then, configure the training parameters of the model according to the training requirements of the digital spectrum afterglow map detection task in complex environments; and then build a model training environment; Step 3: Construct a CIOD-YOLO model; Step 4: Load the constructed CIOD-YOLO model into the training environment of Step 2, and train and validate the model based on the training set and validation set divided in Step 1; Step 5: Use the trained CIOD-YOLO model to identify the radiation source to be detected.
Need to check novelty before this filing date? Find Prior Art