Dual-Vision Sensing Layout for Wider UAV Depth Perception
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for enhancing the sensing range and accuracy of movable platforms, such as unmanned aerial vehicles, either limit the depth information calculation to overlapping visual ranges or require complex, costly, and weight-intensive sensor deployments.
Innovation Solution
Deploying two vision sensors with partially overlapping visual ranges, allowing for the calculation of positional information of objects in both overlapping and non-overlapping areas, using a combination of binocular and monocular algorithms, and semantic prediction to improve sensing accuracy and range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple vision sensors are deployed to expand sensing range, then sensing coverage is improved, but device complexity and cost increase
Solution Approach 1:
The patent combines binocular vision (overlapping visual ranges) and monocular vision (non-overlapping visual ranges) into a unified sensing system. Two vision sensors are deployed such that their visual ranges partially overlap, allowing the system to process images from both sensors together to obtain positional information across the entire environmental area, including both overlapping and non-overlapping regions.
Solution Approach 2:
The vision sensors are designed to serve multiple functions: they can operate independently in their non-overlapping visual ranges using monocular algorithms, and they can work together in their overlapping visual ranges using binocular algorithms. This multi-functionality allows a single sensor configuration to achieve comprehensive environmental coverage without requiring separate sensor systems for different sensing modes.
2Area of stationary object
If multiple vision sensors are deployed to expand sensing range, then sensing coverage is improved, but weight and cost increase
Solution Approach 1:
The patent merges binocular and monocular vision capabilities into a single integrated system using only two vision sensors. By strategically positioning the sensors to create partial overlap in their visual ranges, the system achieves comprehensive environmental coverage without requiring additional sensors, thereby minimizing weight while maximizing sensing capability.
Solution Approach 2:
The patent transitions from a single-sensor monocular system to a dual-sensor system with spatial overlap, adding a dimensional aspect to the sensing architecture. This spatial arrangement allows the system to leverage both overlapping and non-overlapping visual ranges, achieving three-dimensional environmental understanding with minimal sensor count.
3Measurement precision
If binocular algorithm is used for depth calculation, then depth information accuracy is improved, but sensing range is limited to overlapping visual ranges
Solution Approach 1:
The patent segments the environmental area into overlapping visual range regions and non-overlapping visual range regions. For overlapping regions, binocular algorithms provide accurate depth information. For non-overlapping regions, monocular algorithms with semantic prediction extend depth estimation capability. This segmentation allows the system to apply the most appropriate algorithm to each region, maximizing both accuracy and coverage.
Solution Approach 2:
The patent introduces semantic prediction as an intermediary mechanism that bridges binocular and monocular vision systems. By using semantic information from overlapping regions as a reference, the system can predict and extend depth information to non-overlapping regions, effectively mediating between the precise but limited binocular system and the broader but less accurate monocular system.
Data Source
AI summary
An image processing method may be applied to a movable platform, and the movable platform may comprise a first vision sensor and a second vision sensor. The method may include obtaining a first localized image of the first vision sensor within an overlapping visual range, obtaining a second localized image of the second vision sensor within the overlapping visual range; acquiring an image captured by the first vision sensor at a first moment and an image captured at a second moment, the first vision sensor being positioned in space at the first moment differently than at the second moment; and determining a relative positional relationship between an object in the space where the movable platform is located and the movable platform based on the first localized image, the second localized image, the image captured at the first moment and the image captured at the second moment.