Windshield Raindrop Detection Scaling Ambient Light
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Raindrop detection on vehicle windshields is not well adapted to changing environmental lighting conditions, such as night or tunnel crossings, affecting the accuracy of automatic windscreen wiper control.
Innovation Solution
A method and device that captures images of the windshield, performs edge detection, and scales pixel intensity based on ambient light levels using a target mean and standard deviation, allowing edge detection thresholds to remain fixed and enhancing image contrast.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If raindrop detection is performed using captured images without scaling, then the detection process is simple, but the detection accuracy deteriorates under changing lighting conditions
Solution Approach 1:
The patent applies preliminary action by scaling the captured images to normalize pixel intensity values before performing edge detection. This preprocessing step adjusts the image data to account for varying lighting conditions in advance, ensuring that subsequent detection algorithms operate on standardized data regardless of ambient light levels, thereby maintaining detection accuracy without requiring complex adaptive thresholding
Solution Approach 2:
The patent changes the parameter of pixel intensity values by applying scaling based on ambient light level estimates. This parameter transformation converts raw pixel values into normalized values that are independent of lighting conditions, allowing the same detection thresholds to be used across different environmental conditions without sacrificing detection precision
2Measurement precision
If edge detection thresholds are adjusted for different lighting conditions, then detection accuracy is maintained, but the operational complexity increases
Solution Approach 1:
Instead of adjusting thresholds during operation based on lighting conditions, the patent performs preliminary scaling of the captured images to normalize the data beforehand. This allows fixed thresholds to be used throughout operation, eliminating the need for dynamic threshold adjustment while maintaining detection accuracy across varying light levels
Solution Approach 2:
The patent inverts the conventional approach by not adapting the detection thresholds to lighting conditions, but rather adapting the input images to match fixed thresholds. This inversion simplifies the operational process by eliminating the need for continuous threshold adjustment while maintaining detection precision
Data Source
Figure 1~3
AI summary
The invention relates to a method of raindrop detection on a windscreen by capturing images on said windscreen, including a step of edge detection (102) in an area of interest of said captured images, charactet?ed in that said method further comprises a step of scaling (101) the captured images pixels in function of the ambient light level passing through said windscreen before detecting edges. The invention also relates to an associated device for raindrop detection.