Visual Localization Mapping Using Twilight Features in Low Light
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous mobile robots face challenges in creating efficient maps for navigation during night-time due to the limitations of conventional cameras in low light conditions, which affects their ability to reliably extract visual features for localization and mapping.
Innovation Solution
A method is developed to generate a universal map by capturing images during twilight, which allows for the determination of both day-time and night-time features, using twilight straight lines and urban lights that are common to both, enabling accurate positioning and alignment of features for efficient navigation regardless of light conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional cameras are used for visual feature extraction, then the mapping process is simple and straightforward, but the reliability of visual features deteriorates in low light conditions during night-time
Solution Approach 1:
The patent performs mapping during twilight conditions before night-time fully sets in, when both day-time and night-time features are still visible. This preliminary action allows the robot to capture and register features in optimal lighting conditions, ensuring reliable visual features are extracted before the deterioration of light conditions occurs.
Solution Approach 2:
The patent changes the temporal parameter of mapping operations by performing them during twilight rather than during full day or full night. This parameter change optimizes the lighting conditions for feature extraction, allowing conventional cameras to capture both day-time and night-time features with high reliability.
2Reliability
If separate day-time and night-time maps are created, then the accuracy of localization for specific conditions is improved, but the device complexity and mapping effort increase
Solution Approach 1:
The patent merges day-time features and night-time features into a single unified map by performing mapping during twilight when both types of features are visible. This merging eliminates the need to create and maintain separate maps, reducing device complexity while maintaining localization accuracy through the inclusion of both feature types.
Solution Approach 2:
The patent creates a universal map that serves multiple functions for both day-time and night-time localization. By capturing features during twilight, the map becomes multi-functional, supporting navigation and localization across different lighting conditions without requiring separate specialized maps.
3Productivity
If mapping is performed during twilight, then a universal map for both day and night is generated with reduced effort, but the time window for mapping is limited
Solution Approach 1:
The patent performs mapping during twilight as a preliminary action before night-time fully sets in. This timing allows the robot to complete the mapping task during the limited window when both day-time and night-time features are visible, maximizing mapping efficiency while accepting the constrained time availability.
Data Source
AI summary
A method comprises generating a map comprising day-time features and night-time features, wherein the position of night-time features relative to the day-time features is determined by at least one image captured during twilight. The invention also relates to a corresponding processing unit configured to execute such a method.