Single Camera Visual Navigation Using Micro-Lens Array Disparity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional visual navigation methods for robots, such as those using binocular stereo cameras, face challenges with calibration difficulties, high data processing requirements, and reduced real-time processing capabilities, leading to lower accuracy and increased costs.
Innovation Solution
A visual navigation method utilizing a single camera with a micro-lens array to obtain ambient sub-images, performing region-matching processes to determine disparity, and pre-processing images to eliminate noise, enabling efficient real-time processing and improved accuracy through stereo matching and three-dimensional reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a binocular stereo camera is used for visual navigation, then positioning accuracy is improved, but device complexity and data processing requirements increase
Solution Approach 1:
The patent divides the image capture function into multiple sub-images captured by different regions of a single camera sensor, rather than using multiple complete cameras. This segmentation approach achieves stereo vision functionality while reducing device complexity and calibration requirements.
Solution Approach 2:
The patent combines multiple image capture functions into a single camera device, integrating the capabilities of what would traditionally require separate cameras into one unified system. This merging reduces overall device complexity while maintaining positioning accuracy.
2Measurement precision
If a binocular stereo camera is used for visual navigation, then positioning accuracy is improved, but processing speed decreases
Solution Approach 1:
By segmenting the image into multiple sub-images within a single camera frame, the system reduces the total data volume compared to multiple complete cameras, enabling faster processing while maintaining stereo vision accuracy for real-time navigation.
3Measurement precision
If GPS and INS are combined for robot navigation, then positioning accuracy is improved, but reliability decreases in obstructed environments
Solution Approach 1:
The patent replaces GPS-based mechanical/electronic positioning with optical-based visual navigation using image processing and stereo matching. This substitution enables navigation in environments where GPS signals are blocked, improving reliability while maintaining positioning accuracy through visual feature matching.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances navigation accuracy and reduces processing time and costs by leveraging a single camera system, allowing for precise depth information determination and effective navigation in complex environments.
Implementation Method 1
obtaining a plurality of ambient sub-images for a collected ambient image based on the collected ambient image, wherein the ambient image is collected by a single camera and the plurality of ambient sub-images is obtained by a micro-lens array included in a primary lens of the camera
Data Source
AI summary
The present disclosure provides a visual navigation method and a visual navigation device. The method may include steps of: obtaining a plurality of ambient sub-images for a collected ambient image based on the collected ambient image; performing a region-matching process on each of the ambient sub-images to determine a disparity between at least two of the ambient sub-images; and determining coordinates of an object in the ambient image based on the determined disparity, and performing a navigation based on the determined coordinates. In the present disclosure, the ambient image is collected by a single camera, the plurality of ambient sub-images is obtained by the micro-lens array based on the collected ambient image, and then stereo matching and three-dimensional (3D) reconstruction are implemented based on the ambient sub-images to perform the visual navigation. In addition, since the ambient image is collected by the single camera, the amount of the data being gathered is smaller, the processing speed is easy to be controlled, and thus it is good in real-time processing. Furthermore, depth information of an object may be determined in a more accurate manner based on information on a vertical disparity and information on a horizontal disparity of the object that may be obtained simultaneously, so that an accuracy of the navigation is effectively improved.