Seamless Bird-View Image Generation Using Camera Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image-based vehicle maneuvering assistant systems face challenges in combining surrounding images from different image sensing devices into a seamless bird-view image, often resulting in visual burdens and increased manufacturing costs due to the need for ultrasonic radar detectors to address overlapping regions.
Innovation Solution
A method and system that utilize a calibration process and bird-view conversion algorithm, assisted by known ground features, to identify and combine surrounding images from various cameras mounted on a carrier, creating a seamless bird-view image with distance estimation aids, thereby eliminating the need for ultrasonic radar detectors and enhancing driving safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional image-based vehicle maneuvering assistant systems display multiple surrounding images from different image sensing devices, then driving assistance is provided, but visual burden increases and manufacturing cost increases due to need for ultrasonic radar detectors
Solution Approach 1:
The patent combines multiple surrounding images from different image sensing devices into a single synthesized bird's-eye view image. This merging process integrates the visual information from multiple cameras into one unified display, eliminating the need to show multiple separate images simultaneously and thereby reducing visual burden while maintaining comprehensive driving assistance
2Ease of manufacture
If mask lines are used to divide bird-view image into plural regions, then image processing is simplified, but seams appear in the bird-view image and ultrasonic radar detectors are required
Solution Approach 1:
The patent applies local quality adjustment by using weighting values to selectively emphasize or de-emphasize specific regions in the overlapped areas of adjacent bird's-eye view images. This allows different parts of the image to have different processing characteristics, enabling seamless blending in critical areas while maintaining processing efficiency
Solution Approach 2:
The patent changes the weighting parameter of image features in overlapped regions to achieve seamless combination. By adjusting the weighting values assigned to corresponding features from adjacent images, the system can smoothly transition between regions and eliminate visible seams without requiring mask lines or additional sensors
3Manufacturing precision
If weighting values are assigned to features in overlapped regions, then seamless combination is achieved, but processing time increases
Solution Approach 1:
The patent applies partial action by selectively assigning weighting values only to specific features in overlapped regions rather than processing the entire image uniformly. This targeted approach focuses computational resources on the critical seam areas, achieving seamless combination while minimizing overall processing time
Data Source
AI summary
The present invention provides a method and system for generating a surrounding seamless bird-view image with distance assistance interface, which process a plurality of surrounding images associated with a carrier captured by different image sensing devices. After that, an assistant marking is determined according to the relationship between acquired image space and the real space and thereby is formed on the surrounding seamless bird-view image so as to assist the driver in the carrier for avoiding collision or events affecting the safety being occurred.