Omnidirectional Image Rectification for Vehicle Cross Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Omnidirectional camera images used in vehicle imaging systems are distorted, making it difficult for drivers to accurately determine the distance and speed of surrounding vehicles, limiting safety benefits.
Innovation Solution
A computer-implemented method and system that transforms omnidirectional images into rectilinear images using a panel model, reducing distortion and allowing for more accurate distance and speed assessments by mapping omnidirectional image pixels to rectilinear image pixels and world coordinates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of moving object
If an omnidirectional camera is used to capture ultra-wide angle views, then the field of view is improved and cross-traffic detection is enhanced, but image distortion increases making distance and speed determination difficult
Solution Approach 1:
The omnidirectional image is divided into multiple regions of interest (ROIs), each corresponding to a specific viewing direction. Each ROI is transformed independently into a separate rectilinear image, allowing precise measurement in each segment while maintaining the overall wide field of view coverage.
Solution Approach 2:
A coordinate transformation system acts as an intermediary between the omnidirectional camera output and the display system. The system uses spherical coordinate system conversions and panel models to mathematically transform distorted omnidirectional images into undistorted rectilinear images, preserving measurement accuracy.
2Device complexity
If the omnidirectional image is displayed directly to the driver, then the system complexity is reduced, but the driver's ability to assess distance and speed accurately deteriorates
Solution Approach 1:
The system performs coordinate transformations and image rectification in advance, before displaying the images to the driver. By pre-processing the omnidirectional images into rectilinear format with accurate spatial relationships, the system eliminates the need for complex real-time calculations during driver interaction while ensuring measurement precision.
3Measurement precision
If multiple rectilinear images are generated from different regions of interest, then the measurement accuracy is improved, but the processing time and computational load increase
Solution Approach 1:
The omnidirectional image is segmented into multiple regions of interest, each processed independently into a rectilinear image. This segmentation allows parallel processing of different ROIs, reducing overall processing time while maintaining high measurement accuracy in each region through dedicated coordinate transformations.
Solution Approach 2:
The system processes only the necessary regions of interest at any given time rather than the entire omnidirectional image. By focusing computational resources on relevant ROIs based on vehicle context and driver needs, the system achieves sufficient measurement accuracy with reduced processing time and computational load.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
System and method of performing various transformations between an omnidirectional image model and a rectilinear image model for use in vehicle imaging systems. In particular, a panel transform system and method for transforming an omnidirectional image from an omnidirectional camera positioned on a vehicle to a rectilinear image. The rectilinear image is then displayed to the driver for use while performing vehicle maneuvers. The panel transform system and method also provide a rectilinear image model based on the omnidirectional camera for use with existing rectilinear imaging processing systems. The rectilinear image model is determined based on a variable set of input parameters that are defined by both automatic and manual system inputs such as a steering wheel angle sensor and a user interface.