Sequential Surface Image Registration for GPS-Free Motion Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Standard localization methods, such as GPS, are ineffective in environments like underground parking lots or moving vehicles, as they require prior knowledge of the environment and unique visual features, and struggle with real-time motion and non-smooth surfaces with similar visual features.
Innovation Solution
A system using an imaging sensor to capture and register sequential images of a surface, allowing for real-time computation of an object's location relative to a reference point, without the need for prior mapping or unique features, by analyzing the overlap between current and previous images, and feeding this location into a controller for movement control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If standard localization methods (GPS, landmark recognition) are used, then localization can be achieved in open environments, but they fail in enclosed environments like underground parking lots or moving vehicles
Solution Approach 1:
The system uses the object's own motion through the environment to create localization signals. By capturing sequential images of the passing environment and registering them, the system generates localization information from the object's self-motion, eliminating dependency on external positioning infrastructure like GPS or pre-mapped landmarks.
Solution Approach 2:
The system transitions from static localization methods to dynamic localization by continuously capturing sequential images during motion and registering them in real-time. The localization is updated dynamically as the object moves, using the changing visual environment to compute new position estimates without requiring prior environmental knowledge.
2Measurement precision
If image registration is performed on non-smooth surfaces with similar visual features, then localization can be achieved without unique features, but the visual features lack distinctiveness for reliable matching
Solution Approach 1:
The system segments the image registration problem into identifying and tracking multiple discrete feature points across sequential images. By detecting individual feature points and their spatial relationships, the system can perform registration even when overall surface textures are similar, as long as the relative configurations of feature points are consistent.
Solution Approach 2:
The system moves from relying on single-feature distinctiveness to utilizing multi-dimensional spatial relationships among multiple features. By considering the geometric configuration and relative positions of multiple feature points rather than relying on unique visual characteristics of individual features, the system achieves robust registration on visually similar surfaces.
3Measurement precision
If sequential images are captured and registered in real-time during motion, then high accuracy localization is achieved, but the computational processing time increases
Solution Approach 1:
The system performs partial image registration by focusing computational efforts on identifying and matching key feature points rather than processing entire images. This selective approach captures sufficient localization information while significantly reducing computational time and resources required for real-time operation.
Solution Approach 2:
The system pre-identifies and tracks feature points across sequential images, maintaining a running list of detected features and their positions. This preliminary processing of feature detection and tracking enables faster registration computations, as the computationally intensive feature identification is performed incrementally as images are captured rather than retrospectively on full image sets.
Data Source
AI summary
There is provided a computer implemented method of controlling movement of an object, comprising: accessing a current image of a surface relative to an object at a current location, wherein an imaging sensor is set to capture the current image depicting an overlap with a previously captured image of the surface when the object was at a previous location, registering the current image to the overlap of the previously captured image, computing the current location of the object relative to a reference location according to an analysis of the registration, and feeding the current location into a controller for controlling movement of the object.