Raindrop Removal in Autonomous Vehicle Video Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicle driving systems face reduced accuracy in road condition recognition due to raindrops in video images captured by vehicle cameras, affecting decision-making in adverse weather conditions.
Innovation Solution
A computation apparatus and method that trains a raindrop-image-recognition model using labeled rainy scene images, recognizes raindrops in video sequences, and replaces specific raindrop images with corresponding regions from previous scene images based on overlapping criteria to generate output scene images, improving image clarity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If raindrop removal processing is performed on video images, then image clarity is improved, but processing time and computational complexity increase
Solution Approach 1:
The system performs preliminary actions by training the raindrop-image-recognition model in advance using labeled rainy scene images. This pre-trained model can then quickly identify raindrops in real-time video sequences without requiring extensive processing during actual operation, thus improving image clarity while minimizing processing time loss.
Solution Approach 2:
The patent extracts and isolates raindrop images from video sequences using the trained recognition model. By specifically targeting and extracting only the raindrop elements rather than processing the entire image, the system achieves effective raindrop removal while reducing overall processing time and computational burden.
2Measurement precision
If a raindrop-image-recognition model is trained using labeled rainy scene images, then raindrop recognition accuracy is improved, but training time and computational resources increase
Solution Approach 1:
The system performs the time-consuming model training action in advance during a preliminary phase. The trained model is then stored and reused for multiple real-time raindrop detection tasks, spreading the training time cost over many applications and achieving high recognition accuracy without sacrificing real-time processing speed.
Solution Approach 2:
The patent creates a copied version of the raindrop recognition capability through the trained model. Instead of retraining the model for each new video sequence, the system uses the copied knowledge from the trained model to rapidly identify raindrops in multiple sequences, reducing repeated training time while maintaining high recognition accuracy.
Data Source
AI summary
A method of removing raindrops from video images is provided. The method includes the steps of: training a raindrop image recognition model using a plurality raindrop training images labeled in a plurality of rainy-scene images; recognizing a plurality of raindrop images from a plurality of scene images in a video sequence using the raindrop image recognition model; and in response to a specific raindrop image in a current scene image satisfying a predetermined condition, replacing the specific raindrop image in the current scene image with an image region corresponding to the specific raindrop image in a specific scene image prior to the current scene image to generate an output scene image.