A method, system, and storage medium for infrared small target detection based on frequency domain attention.

By using a method that fuses frequency domain attention with spatial domain feature maps across channels, the problems of unclear target features and severe background interference in infrared small target detection are solved, achieving efficient and accurate infrared small target detection.

CN119762761BActive Publication Date: 2025-10-28NANJING UNIV OF INFORMATION SCI & TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202510039933.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-01-10
Publication Date
2025-10-28
Estimated Expiration
2045-01-10

AI Technical Summary

Technical Problem

In infrared small target detection suffers from problems such as unclear target features, severe background interference, and limited shape features, resulting in low detection accuracy and high computational cost.

Method used

A method of cross-channel fusion of frequency domain attention and spatial domain feature maps is adopted to reconstruct infrared small target images through a frequency domain attention decoder, thereby reducing computational complexity and improving detection accuracy.

Benefits of technology

It achieves efficient and accurate infrared small target detection, improves the detection accuracy of the detection model and reduces computational complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119762761B_ABST
    Figure CN119762761B_ABST
Patent Text Reader

Abstract

This invention discloses an infrared small target detection method, system, and storage medium based on frequency domain attention, belonging to the field of infrared small target detection technology. The method includes: acquiring infrared image data to be tested and preprocessing it to obtain a standard infrared image; inputting the standard infrared image into a pre-trained infrared small target detection model to obtain detected infrared small targets; wherein, the method for acquiring the pre-trained infrared small target detection model includes: acquiring a historical infrared image sample set; preprocessing each infrared image data in the historical infrared sample set to obtain a standard historical sample set; inputting the standard historical sample set into a pre-constructed detection model to obtain a trained infrared small target detection model, wherein the decoding module of the detection model reconstructs the target image by fusing frequency domain attention-weighted feature maps and spatial domain feature maps, reducing computational complexity and improving the accuracy of infrared small target detection.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Infrared image weak and small target detection algorithm

    CN118196369A

  • Deep image watermarking method based on mixed frequency-domain channel attention

    WO2024032158A1