Pose Estimation via Image Surface Intersection Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current pose estimation methods for air and underwater vehicles face challenges such as increased complexity, accuracy issues due to noise, and the difficulty in distinguishing true surfaces from false surfaces in noisy conditions, particularly when geographical features are unclear or symmetrical surfaces are present.
Innovation Solution
A data processing device and method that estimates pose by dividing images into candidate and background regions, extracting pose parameters, computing surface parameter formulas, and identifying significant candidate surfaces through intersection detection and weighting computation to accurately determine the reference surface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple sensors (pose measurement sensors, image sensors, radar, sonar) are mounted on air or underwater vehicles to improve navigation and inspection capabilities, then the functionality and measurement capability are enhanced, but the system complexity, size, weight, and power consumption increase significantly
Solution Approach 1:
The patent makes the image sensor perform dual functions: capturing images for navigation/inspection and simultaneously measuring pose parameters (pitch, roll, yaw, altitude, depth). By processing image data to detect reference surfaces and calculate pose, the system eliminates dedicated pose measurement sensors, reducing overall system complexity while maintaining both imaging and navigation capabilities
Solution Approach 2:
The patent combines the pose measurement function with the image sensor by integrating image processing algorithms that extract pose information from captured images. This merging of functions allows a single sensor system to provide both visual data and spatial orientation data, reducing the number of separate components needed
2Measurement precision
If traditional pose measurement sensors using inertia are used to measure pitch, roll, and yaw, then pose parameters can be obtained, but accuracy severely deteriorates in the presence of vibration
Solution Approach 1:
The patent replaces mechanical inertia-based pose sensors with an optical/image processing-based system. By using image data from the camera and processing it through reference surface detection algorithms, the system obtains pose information without being susceptible to vibration interference that plagues mechanical inertial sensors
Solution Approach 2:
The patent introduces image data as an intermediary medium to measure pose. Instead of directly measuring physical motion with inertial sensors, the system captures images of the environment, processes these images to detect reference surfaces, and derives pose parameters from the geometric relationships in the images, thereby avoiding direct mechanical measurement in vibrating conditions
3Weight of stationary object
If image sensors are used to substitute for pose measurement sensors to reduce size and weight, then the system becomes more compact and energy-efficient, but the ability to accurately measure pose in noisy conditions and distinguish true surfaces from false surfaces deteriorates
Solution Approach 1:
The patent segments the image processing into distinct stages: candidate region extraction, reference surface detection, and pose parameter calculation. By dividing the processing task and applying specific algorithms at each stage (including weighting computations and intersection detections), the system maintains high accuracy in pose estimation while using lightweight image sensors instead of heavy mechanical sensors
Solution Approach 2:
The patent employs a composite approach combining multiple processing techniques: image capture, candidate region extraction, reference surface detection, weighting computations, and intersection detections. This multi-layered processing methodology enables accurate pose measurement from image data, achieving the precision previously requiring heavier dedicated sensors
4Adaptability or versatility
If basic information such as the ground surface or ocean surface is used as a reference for pose estimation in locations lacking geographical features, then navigation becomes possible in featureless environments, but the ability to determine yaw angle deteriorates due to surface symmetry
Solution Approach 1:
The patent applies partial action by determining that yaw measurement is not always necessary for safe navigation. The system prioritizes measuring altitude, pitch, and roll using surface reference, and only attempts yaw determination when additional features or asymmetric surface characteristics are detectable, thus maintaining functionality in featureless environments while acknowledging yaw measurement limitations
Data Source
AI summary
In a pose estimation for estimating the pose of an object of pose estimation with respect to a reference surface that serves as a reference for estimating a pose, a data processing device: extracts pose parameters from a binarized image; identifies a combination of pose parameters for which the number of cross surfaces of parameter surfaces that accord with surface parameter formulas, which are numerical formulas for expressing a reference surface, is a maximum; finds a slope weighting for each of cross pixels, which are pixels on each candidate surface and which are pixels within a prescribed range, that is identified based on the angles of the tangent plane at the cross pixel and based on planes formed by each of the axes of parameter space; and identifies the significant candidate surface for which a number, which is the sum of slope weightings, is a maximum, as the actual surface that is the reference surface that actually exists in the image.