Visual Odometry Using Fixed Light Sources in Low Illumination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional visual odometry systems fail in low illumination conditions due to insufficient discernible features and texture, leading to incorrect estimation of object movement.
Innovation Solution
The method involves capturing images with a camera mounted on a mode of transportation, identifying fixed light sources and ground planes, and using inverse perspective maps to determine movement parameters such as translation and rotation by transforming image data to map intersections, enabling real-time or near real-time movement tracking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional visual odometry is used in low illumination conditions, then the system can operate without additional lighting, but the feature extraction and matching become insufficient due to lack of discernible features and texture
Solution Approach 1:
The patent introduces fixed light sources (street lights, lamps) as intermediary objects to provide discernible features in low illumination conditions. These light sources serve as mediators between the camera and the environment, creating visible reference points that enable feature extraction and matching when ambient lighting is insufficient. The light sources act as stable, identifiable features that bridge the gap between the imaging system and the dark environment.
Solution Approach 2:
The patent utilizes the brightness and intensity characteristics of fixed light sources to create discernible features. By detecting and tracking the position and intensity of these light sources across multiple images, the system can extract meaningful features even in dark conditions. The light sources effectively change the illumination characteristics of the scene, creating visible contrast and texture that were previously absent.
2Device complexity
If feature extraction is performed at corner points as in typical visual odometry, then the method is simple and computationally efficient, but the features become unstable across changes in scale and orientation
Solution Approach 1:
The patent changes the fundamental parameter being tracked from geometric corner points to luminous intensity characteristics of fixed light sources. Instead of relying on spatial coordinates and corner detection algorithms, the system detects light sources based on their brightness profiles and intensity distributions. This parameter change makes the features inherently more stable because light sources maintain consistent luminous characteristics across scale and orientation changes, unlike corner points which are highly sensitive to such transformations.
3Adaptability or versatility
If standard monocular visual odometry is used, then the system requires highly textured regions for matching, but low illumination conditions provide insufficient texture and discernable features
Solution Approach 1:
The patent converts the harmful effect of low illumination (which removes texture and features) into a beneficial situation by focusing specifically on fixed light sources as the primary features. Rather than trying to recover lost texture information, the system embraces the dark conditions and uses the light sources themselves as the sole basis for matching. This approach transforms the information loss into a simplified feature set that is actually more reliable in low light, as light sources remain consistently visible when other features disappear.
Data Source
AI summary
First and second image data is captured comprising a first and second image, respectively. A fixed light source is identified in each of the first and second images. A first ground plane is determined in the first image data. A first (second) intersection is determined, wherein the first (second) intersection is a point in the first image where a virtual lamp post corresponding to the fixed light source in the first (second) image intersects with the first (second) ground plane. The first image data and the second image data are transformed to provide a first and second inverse perspective map (IPM) comprising a first transformed intersection and a second transformed intersection, respectively. Movement parameters are determined based on the location of the first transformed intersection in the first IPM and the location of the second transformed intersection in the second IPM.