De-blurring Vehicle Images Using Motion and Steering Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing visual information augmentation technologies face challenges in accurately detecting objects due to image blurriness caused by vehicle motion, especially under low illuminance conditions or when lane markings are obscured, making it difficult to distinguish between global and local blurs.
Innovation Solution
A method and apparatus that determine the type of blur in an image based on vehicle control information, such as steering angle and speed, and select appropriate de-blurring schemes using filters like motion filters or de-focusing filters to clarify the image, allowing for accurate object recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If auto exposure sensor is used to capture image at night, then image capture is enabled under low illuminance, but image blur occurs due to dynamically moving vehicle or object
Solution Approach 1:
The system dynamically adjusts the de-blurring approach based on real-time vehicle control information (steering angle, speed, acceleration). When vehicle motion is detected, motion-based de-blurring is applied; when stationary, static de-blurring is used. This dynamic adaptation resolves the contradiction by optimizing image clarity for object detection while maintaining night capture capability.
Solution Approach 2:
The system changes the de-blurring parameters and filter selection based on vehicle state parameters (speed, steering angle, acceleration). By adjusting the de-blurring strength and type according to vehicle motion parameters, the system recovers image quality after exposure, enabling accurate object detection even in low light conditions.
2Measurement precision
If image de-blurring is applied to remove blur, then image clarity is improved, but different blur types require different de-blurring schemes increasing system complexity
Solution Approach 1:
The system segments the blur problem into distinct types (motion blur vs. static blur) based on vehicle control information. By categorizing blur sources according to vehicle state (moving vs. stationary, steering angle thresholds), the system applies targeted de-blurring schemes for each segment, simplifying the overall decision process while maintaining effectiveness.
Solution Approach 2:
The system dynamically selects the appropriate de-blurring scheme based on real-time vehicle state. Rather than maintaining all possible de-blurring algorithms ready, the system activates only the relevant scheme based on current vehicle motion parameters, reducing computational complexity while maintaining image quality.
3Measurement precision
If vehicle motion information is used to determine blur type, then accurate de-blurring selection is achieved, but additional sensors and data processing are required
Solution Approach 1:
The system uses existing vehicle control information (steering angle, speed, acceleration) that is already collected for vehicle operation. By making this multi-functional data serve dual purposes (vehicle control and image de-blurring decision-making), the system avoids adding dedicated sensors while achieving accurate blur type identification.
Solution Approach 2:
The vehicle's own control systems and sensors serve the additional function of determining image blur characteristics. The vehicle's motion data, already being processed for driving control, is reused to inform the de-blurring process, eliminating the need for separate detection systems.
Data Source
AI summary
Disclosed is a method and apparatus for recognizing an object, the method including determining whether an image comprises a blur, determining a blur type of the blur based on control information of a vehicle, in response to the image comprising the blur, selecting a de-blurring scheme corresponding to the determined blur type, de-blurring the image using the selected de-blurring scheme, and recognizing an object in the image based the de-blurred image.