Fisheye Image Rectification Using Dual Lookup Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fisheye lenses used in cameras produce images with significant spatial distortions, which conventional rectification techniques often address at the cost of losing data in central or peripheral regions, and require computationally intensive three-dimensional rendering.
Innovation Solution
A computer-implemented technique that applies first and second lookup functions to fisheye images to generate rectilinear images from specific portions, mitigating distortions without data loss and reducing computational resources, suitable for use in autonomous vehicles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional computer-implemented rectification techniques are applied to fisheye images, then spatial distortions are removed, but data is lost in central or peripheral regions
Solution Approach 1:
The fisheye image is divided into multiple regions (central region and peripheral regions), and different rectification functions are applied to each region. This segmentation allows the patent to preserve data in all regions while achieving accurate rectification, as each region can be processed with an appropriate rectification function that maintains its unique characteristics.
2Manufacturing precision
If three-dimensional rendering techniques are used to rectify fisheye images, then spatial distortions are removed, but computational resources are significantly consumed
Solution Approach 1:
The patent uses computationally efficient lookup tables and simple rectification functions instead of complex three-dimensional rendering techniques. These lighter computational methods achieve sufficient rectification precision while consuming significantly less energy, making them suitable for real-time processing in autonomous vehicles.
3Device complexity
If a single rectification function is applied to the entire fisheye image, then processing is simplified, but data loss occurs in certain regions
Solution Approach 1:
The patent applies different rectification functions to different regions of the fisheye image based on their specific characteristics. The central region uses one rectification function while peripheral regions use different functions, allowing each region to be processed with the most appropriate method that preserves its data without excessive complexity.
Data Source
AI summary
Various technologies described herein pertain to rectification of a fisheye image. A computing system receives the fisheye image. Responsive to receiving the fisheye image, the computing system applies a first lookup function to a first portion of the fisheye image to mitigate spatial distortions of the fisheye image. The computing system also applies a second lookup function to a second portion of the fisheye image to mitigate the spatial distortions. The first lookup function maps first pixels in the first portion to a first rectilinear image corresponding to the first portion when viewed from a first perspective of a first virtual camera. The second lookup function maps second pixels in the second portion to a second rectilinear image corresponding to the second portion when viewed from a second perspective of a second virtual camera. The computing system then outputs the first rectilinear image and the second rectilinear image.