Dual-Camera VSLAM Scale Correction for Mobile-Body Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing VSLAM techniques using monocular images struggle with unknown distance units in generated environment maps, necessitating auxiliary sensors for scale adjustment, which complicates self-localization and map generation.
Innovation Solution
Employing two spherical cameras mounted on a mobile body to determine their positions on an environment map and calculate a correction coefficient based on their known reference distance, allowing for scale correction of the map to align with real-space distances.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If monocular images are used for VSLAM, then the system complexity is reduced, but the distance unit of the generated environment map becomes unknown requiring auxiliary sensors
Solution Approach 1:
The patent introduces a second imaging device as an intermediary element to mediate between the monocular imaging system and the environment map generation. This second imaging device provides additional depth information that acts as a bridge, enabling scale determination without requiring auxiliary sensors like wheel odometry, thus resolving the contradiction between system simplicity and measurement accuracy
Solution Approach 2:
The patent transitions from purely two-dimensional monocular image analysis to incorporating three-dimensional depth information by introducing a second imaging device. This dimensional enhancement allows the system to determine absolute distances and scale in the environment map, solving the unknown distance unit problem while maintaining relative system simplicity
2Measurement precision
If auxiliary sensors are added for scale adjustment, then the distance measurement accuracy is improved, but the device complexity increases
Solution Approach 1:
The patent makes the imaging devices serve multiple functions: they not only capture images for environment map generation and self-localization but also provide depth information for scale determination. This multi-functionality eliminates the need for dedicated auxiliary sensors, achieving accurate scale measurement without increasing device complexity
Solution Approach 2:
The imaging devices on the mobile body serve themselves by providing both localization data and scale information. The system uses its own imaging resources to determine the environment map scale through depth information, rather than relying on external auxiliary sensors, thus achieving self-sufficiency and avoiding additional hardware
3Ease of operation
If the position is determined based on monocular images only, then the ease of operation is maintained, but the localization accuracy deteriorates due to unknown scale
Solution Approach 1:
The second imaging device acts as an intermediary that provides depth information to bridge the gap between simple monocular operation and accurate localization. This intermediary enables the system to maintain ease of operation with image-based processing while achieving accurate localization through the additional depth data it provides
Data Source
AI summary
An information processing system includes circuitry to read from a memory an environment map depicting a surrounding environment in which a mobile body including first and second imaging devices to capture monocular images moves. The circuitry determines a first position of the first imaging device on the environment map based on a monocular image captured by the first imaging device, determines a second position of the second imaging device on the environment map based on a monocular image captured by the second imaging device, generates correction information for correcting a scale of the environment map based on a distance between the first position and the second position on the environment map and a reference distance between the first imaging device and the second imaging device, and determines a position of the mobile body using the environment map and the correction information.