On-Board Camera Image Enhancement Using GAN and Lens Cleaning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The image quality captured by onboard cameras in vehicles, especially under poor-lighting conditions, affects the reliability and safety of automatic driving, as it is influenced by camera parameters and driving environments such as rain and fog.
Innovation Solution
A method and system utilizing generative adversarial networks (GAN) to enhance image quality by training a generative model with images from sampling vehicles and integrating a cleaning device with the camera to remove dust, fog, and water droplets from the lens surface, improving image quality in real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image enhancement using GAN is applied, then image quality is improved, but computational complexity and processing time increase
Solution Approach 1:
The GAN model is trained in advance offline to learn the mapping between degraded and high-quality images. During actual driving operation, the pre-trained model performs rapid inference to enhance images in real-time, separating the computationally intensive training phase from the time-critical deployment phase.
Solution Approach 2:
Traditional image enhancement methods using conventional filtering or optimization algorithms are replaced with a deep learning-based GAN approach. The generative model learns complex image transformations through training, substituting iterative mechanical optimization with a trained neural network that performs enhancement through forward propagation.
2Measurement precision
If cleaning device is integrated with camera, then lens cleanliness and image quality are improved, but device complexity increases
Solution Approach 1:
The cleaning device and camera are integrated into a single unified module, where the cleaning mechanism (such as a blowing device or wiping element) is combined with the camera housing. This merging allows the cleaning function to be performed in close proximity to the lens without requiring separate mounting and control systems.
Solution Approach 2:
The cleaning device is designed to automatically clean the camera lens without requiring manual intervention. The system self-monitors lens conditions and activates the cleaning mechanism as needed, performing maintenance functions autonomously to maintain image quality during operation.
Data Source
AI summary
Disclosed are a method and a system for enhancing an image captured by an on-board camera, and a computing device. The method includes: acquiring a plurality of first original image captured by a camera of the sampling vehicle; when a discriminative model in a generative adversarial networks (GAN) model determines that a quality of an output image of a generative model is poorer than an expected quality, training the generative model by using the first original images to obtain an image enhancement model; and enhancing a second original image captured by an on-board camera of a running vehicle by using the image enhancement model to acquire an enhanced image.