Roadside Image Stabilization via Offset Observation Lists
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image stabilization methods for roadside perception systems, which rely on feature point matching, are inadequate when signal lights or cameras move slightly, leading to position offset changes and reduced recognition accuracy.
Innovation Solution
A method and apparatus that calculate and apply offsets between image stabilization boxes in a current image and a template image, adding offsets to an observation list until a preset condition is met, allowing for accurate determination of signal light positions based on image similarity, using strategies like homography matrix transformation, nearest image stabilization, and mean value signal light image stabilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature point matching method is used for image stabilization, then image position offset can be corrected, but computational complexity increases and processing time extends
Solution Approach 1:
The patent extracts only the necessary image stabilization boxes from the full image, rather than performing feature point matching across the entire image. This extraction approach reduces the computational scope while maintaining the essential function of detecting position offsets, thereby reducing processing time without sacrificing correction accuracy.
Solution Approach 2:
The patent performs preliminary identification of image stabilization boxes in advance, creating a set of pre-identified regions that can be quickly processed later. This preliminary action eliminates the need for time-consuming feature extraction and matching during the actual stabilization process, significantly reducing processing time while preserving accuracy.
2Measurement precision
If feature extraction is performed for each image, then position offset can be calculated, but computational resource consumption increases
Solution Approach 1:
The patent extracts only the essential image stabilization boxes rather than performing comprehensive feature extraction across the entire image. This selective extraction reduces computational resource consumption by focusing only on the relevant regions needed for position offset calculation, while maintaining the accuracy required for stable signal light detection.
Solution Approach 2:
The patent uses a lightweight image stabilization box identification approach that consumes fewer computational resources compared to traditional feature extraction methods. This simplified approach uses less energy and computational power, making it suitable for continuous processing without requiring expensive or complex computational resources.
3Reliability
If traditional image stabilization method is used, then position offset can be corrected, but the system is sensitive to illumination variations
Solution Approach 1:
The patent introduces image stabilization boxes as an intermediary element that mediates between the image content and the stabilization process. These boxes are designed to be less sensitive to illumination variations compared to traditional feature points, thereby reducing the harmful effect of lighting changes on stabilization reliability while maintaining accurate position offset correction.
Data Source
AI summary
A method and an apparatus for stabilizing an image, a roadside device, and a cloud control platform are provided. The method may include: calculating, for a current image in an image sequence, a first offset between each image stabilization box of the current image and a corresponding image stabilization box of a template image; adding a first offset meeting a first preset condition to an observation list of a corresponding image stabilization box of the template image until there is an observation list meeting a second preset condition; and determining a position of the signal light of the current image, based on a first offset in an observation list of at least one image stabilization box and a position of the signal light of the template image.