3D Object Positioning Using Feature Matching and Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine vision systems face challenges in image processing speed and reliability, particularly with electro-optic image sensors, which are time-consuming and unreliable for robotics and autonomous applications due to the need for matching different orientations and sizes of templates.
Innovation Solution
Integration of electro-optic image sensors with a global positioning system/inertial measurement unit and a laser ranger to derive attitude and orientation data, enabling real-time object positioning by using fiducial feature and corner detection methods, and combining these with line and circle detection for complex object identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional image processing methods (correlation, gray level matching, graph cut) are used to match all image pixels, then complete image matching is achieved, but processing time increases significantly and reliability decreases
Solution Approach 1:
The patent extracts only the essential feature points (corners and fiducial features) from the complete image for matching, rather than processing all pixels. This extraction approach maintains positioning reliability by focusing on distinctive features while dramatically reducing processing time by eliminating redundant pixel comparisons.
Solution Approach 2:
The patent segments the image processing task by dividing it into distinct stages: feature detection (corner detection, fiducial feature detection), feature matching, and 3D positioning calculation. This segmentation allows each stage to be optimized independently, improving overall efficiency and reliability.
2Measurement precision
If template matching with different orientations and sizes is performed, then object detection accuracy is improved, but processing complexity and time increase
Solution Approach 1:
The patent changes the approach from varying template parameters (orientation, size) to varying the sensor's pose parameters (attitude and orientation from GPS/IMU, range from laser ranger). This allows accurate object detection while simplifying the matching process by using fixed templates and adjusting sensor parameters instead.
Solution Approach 2:
The patent performs preliminary actions by pre-obtaining sensor attitude, orientation, and range data from GPS/IMU and laser ranger before image matching. This preliminary information is used to pre-rotate and pre-scale templates, eliminating the need for iterative orientation and size adjustments during matching.
3Productivity
If feature matching is performed without integrated navigation data, then system simplicity is maintained, but positioning speed and reliability decrease
Solution Approach 1:
The patent merges multiple subsystems (electro-optic image sensors, GPS/IMU integrated navigation system, and laser ranger) into a unified positioning system. The navigation system provides attitude and orientation data while the laser ranger provides range data, both of which are integrated with image matching to achieve fast and reliable 3D positioning.
Solution Approach 2:
The patent introduces an intermediary data processing layer that receives raw data from GPS/IMU and laser ranger, processes this information into useful parameters (attitude, orientation, range), and provides these parameters to the image matching algorithm. This intermediary layer enables efficient positioning while managing system complexity through modular design.
Data Source
AI summary
An object positioning solves said problems encountered in machine vision, which employs electro-optic (EO) image sensors enhanced with integrated laser ranger, global positioning system/inertial measurement unit, and integrates these data to get reliable and real time object position. An object positioning and data integrating system comprises EO sensors, a MEMS IMU, a GPS receiver, a laser ranger, a preprocessing module, a segmentation module, a detection module, a recognition module, a 3D positioning module, and a tracking module, in which autonomous, reliable and real time object positioning and tracking can be achieved.