Foggy day image target detection method based on Radon transformation and feature space decision

By using Radon transformation and feature space decision-making methods in foggy day image processing, the problem of low accuracy of target detection in foggy day is solved, efficient and accurate target detection is achieved, and traffic safety in foggy day environments is significantly improved.

CN119169311AActive Publication Date: 2024-12-20DALI UNIV
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202410820587.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-24
Publication Date
2024-12-20
Estimated Expiration
2044-06-24

AI Technical Summary

Technical Problem

The existing foggy-day image object detection technology has the problem of low accuracy and high error detection rate, and it is difficult to effectively detect targets in images under severe weather conditions.

Method used

Using Radon transformation and feature space decision-making methods, the target pixel points are accurately positioned through preprocessing, edge information extraction, normalization of grayscale Hough transform, Radon transformation, inverse Radon transformation and wavelet transform, and the target is achieved through decision functions. Test.

Benefits of technology

It significantly improves the clarity and contrast of the target trail, reduces false detection and missed detection, improves the accuracy and credibility of detection, and can retain more than 95% of the detailed information in foggy environments, with a peak signal-to-noise ratio of more than 30dB. Low noise ensures reliability of detection results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119169311A_ABST
    Figure CN119169311A_ABST
Patent Text Reader

Abstract

The invention discloses a foggy day image target detection method based on Radon transformation and feature space decision, and belongs to the technical field of image processing. According to the invention, the problem of low accuracy of target detection in the existing foggy day image is solved. According to the method, firstly, through preprocessing and normalized gray Hough transform, the definition and contrast of a target wake are remarkably enhanced, so that a target image in a foggy environment is clearer and easier to distinguish; and secondly, the image is reconstructed by adopting Radon transformation and inverse Radon transformation, a target peak point is extracted, and the extracted target peak point is matched with a peak value of wavelet transformation, so that a detection target can be accurately positioned, and false detection and missing detection are reduced. And finally, the detection error is effectively reduced through feature space decision, and the detection accuracy is improved. The method can be applied to the field of image target detection.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Foggy day image salient target detection method

    CN110210407A