Vertical Fisheye Stereo Camera Using Equirectangular Depth Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing RGB-D cameras face challenges in achieving wide viewing angles, particularly in horizontal and vertical directions, which limits their effectiveness in applications like human-computer interfaces, autonomous driving, and robotics, due to the use of rectilinear or fisheye lenses that hinder efficient stereo matching and depth information acquisition.
Innovation Solution
A stereo camera apparatus with two fisheye lenses arranged vertically to improve both horizontal and vertical viewing angles, utilizing an equirectangular projection scheme for image conversion, enabling efficient stereo matching and depth extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a rectilinear projection lens is used, then the image appears similar to human perception with straight lines remaining straight, but the viewing angle is limited to about 120 degrees horizontally
Solution Approach 1:
The patent employs a fisheye lens with spherical projection characteristics to capture a viewing angle exceeding 180 degrees. The lens projects the real world into an image with barrel distortion, allowing the camera to achieve a horizontal viewing angle greater than 180 degrees while maintaining the ability to perform stereo matching through coordinate transformation
2Area of stationary object
If a fisheye lens is used to achieve a viewing angle greater than 180 degrees, then the viewing angle is improved, but efficient stereo matching through 1D search becomes difficult due to barrel distortion
Solution Approach 1:
The patent transforms the coordinate system from spherical coordinates to Cartesian coordinates through a mapping function. This parameter transformation converts the fisheye image with barrel distortion into a format where epipolar lines become horizontal, enabling efficient 1D search for stereo matching while preserving the wide viewing angle capability
Solution Approach 2:
The patent introduces an intermediate coordinate transformation step that acts as a mediator between the fisheye projection and stereo matching. By converting spherical coordinates to Cartesian coordinates through a mapping function, the system creates an intermediate representation that enables both wide viewing angle and efficient stereo matching
3Area of stationary object
If two cameras are arranged horizontally to increase horizontal viewing angle, then the horizontal viewing angle is improved, but the baseline distance increases which reduces depth measurement accuracy
Solution Approach 1:
The patent changes the arrangement dimension from horizontal to vertical, placing two fisheye cameras one above the other. This vertical arrangement allows the baseline to be perpendicular to the horizontal viewing direction, enabling the system to achieve wide horizontal viewing angle while maintaining short baseline distance for accurate depth measurement in the horizontal plane
4Measurement precision
If patterned light or structured light is used to increase stereo matching speed and accuracy, then the stereo matching process is improved, but the device complexity and cost increase
Solution Approach 1:
The patent replaces the mechanical/optical system of structured light projection with a computational approach. By transforming the coordinate system and utilizing the natural stereo parallax in fisheye images, the system achieves accurate depth measurement without requiring additional light projection hardware, thereby reducing device complexity
Data Source
AI summary
Disclosed are a wide viewing angle stereo camera apparatus and a depth image processing method using the same. A stereo camera apparatus includes a receiver configured to receive a first image and a second image of a subject captured through a first lens and a second lens that are provided in a vertical direction; a converter configured to convert the received first image and second image using a map projection scheme; and a processing configured to extract a depth of the subject by performing stereo matching on the first image and the second image converted using the map projection scheme, in a height direction.