Image processing method and vehicle

By generating a snowflake mask and reconstructing the image using a physical scattering model, the problem of poor image quality in streaming rearview mirrors under snowy conditions was solved, achieving efficient and robust image restoration and improving driving safety in snowy weather.

CN122156379APending Publication Date: 2026-06-05GREAT WALL MOTOR CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
GREAT WALL MOTOR CO LTD
Filing Date
2026-03-05
Publication Date
2026-06-05

AI Technical Summary

Technical Problem

In snowy conditions, the streaming rearview mirror image is interfered with by snowflakes, resulting in a blurred and washed-out image, making it difficult to identify the outline of vehicles behind and affecting driving safety. The physical scattering model in the existing technology fails under high-density snow conditions, and the deep learning method has high computational complexity and limited generalization ability, which cannot meet the requirements of real-time performance and environmental adaptability.

Method used

By generating a snowflake mask, determining the ambient light parameters and target transmittance distribution map, and using a physical scattering model for image reconstruction, combined with a lightweight mask prediction model and motion estimation technology, the spatiotemporal consistency of the snowflake mask is optimized to achieve accurate localization and removal of snowflake interference.

Benefits of technology

It improves image quality in snowy conditions, ensures the naturalness and reliability of image restoration, meets the comprehensive requirements of automotive-grade systems for processing speed, environmental adaptability and computing resources, and enhances driving safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122156379A_ABST
    Figure CN122156379A_ABST
Patent Text Reader

Abstract

The application discloses an image processing method and a vehicle, and belongs to the technical field of vehicle safety. By introducing a snowflake mask as a key link connecting snowflake recognition and physical model reconstruction, accurate positioning and effective removal of snowflake interference are realized. This technical concept not only solves the failure problem of traditional physical models in high-density snow conditions, but also avoids the limitations of deep learning methods in real-time performance, generalization ability and computing resources. The overall scheme meets the comprehensive requirements of the vehicle-level system on processing speed, environmental adaptability and computing resources while ensuring the naturalness and reliability of image restoration, providing an efficient and robust solution for the practical enhancement of streaming rearview mirrors in snowy environments.
Need to check novelty before this filing date? Find Prior Art