Bionic Binocular Vision Active Positioning for SLAM Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot vision systems struggle with active information acquisition and precise gaze positioning, leading to tracking failures in SLAM tasks due to sparse textures or dynamic objects.
Innovation Solution
An active positioning method for bionic binocular vision that utilizes a panoramic camera to capture a 360° image, detect key scene information, and create a panoramic value map. This map guides the active gaze control of binocular bionic eyes, ensuring they focus on high-value information areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If visual sensors are fixed directly to the robot, then the system structure is simple, but the visual perception capability is insufficient and cannot achieve active information acquisition
Solution Approach 1:
The visual system is divided into peripheral vision (panoramic camera) and central vision (bionic eye cameras) components. The panoramic camera captures the entire scene and identifies high-value information areas, while the bionic eye cameras focus on these specific areas for detailed observation. This segmentation enables active information acquisition by directing central vision resources to the most important regions identified by peripheral vision.
Solution Approach 2:
The panoramic camera performs preliminary scanning of the entire scene to identify high-value information areas before the bionic eye cameras begin detailed observation. The system pre-processes the visual information to determine where attention should be focused, allowing the bionic eyes to proactively position themselves at optimal locations rather than passively recording whatever is in their fixed field of view.
2Reliability
If the robot moves in an unknown environment, then it can explore the environment, but tracking failures occur when images have sparse textures or numerous dynamic objects
Solution Approach 1:
The panoramic camera performs preliminary scanning of the entire scene to identify high-value information areas with sufficient texture and feature points before the bionic eye cameras begin detailed observation. This preliminary assessment allows the system to proactively position the bionic eyes at locations guaranteed to provide adequate visual features for SLAM tracking, preventing tracking failures in advance.
Solution Approach 2:
The system uses the panoramic camera to continuously monitor the entire scene and provide feedback about the quality of visual features in different areas. Based on this feedback, the bionic eye cameras dynamically adjust their positions to maintain optimal viewing conditions for SLAM tracking, ensuring reliable operation even in challenging environments with sparse textures or dynamic objects.
3Adaptability or versatility
If the bionic eye cameras passively receive image data, then the system is simple to operate, but the system cannot actively extract key information or control motion to shift gaze
Solution Approach 1:
The system performs self-service by automatically identifying high-value information areas through panoramic scanning and autonomously controlling the bionic eye cameras to position themselves at optimal locations. The active positioning algorithm calculates the necessary motion commands to shift gaze toward areas with key information, enabling the system to actively extract meaningful visual data without requiring manual intervention or complex external control.
4Loss of information
If the system focuses on central vision only, then the processing is simple, but key information in the entire scene may be missed
Solution Approach 1:
The visual system is divided into peripheral vision (panoramic camera) and central vision (bionic eye cameras) components. The panoramic camera captures the entire scene to ensure comprehensive information coverage, while the bionic eye cameras focus on specific high-value areas for detailed observation. This segmentation allows the system to maintain complete scene awareness while concentrating processing resources on the most important regions.
Solution Approach 2:
The system merges the wide-field coverage of the panoramic camera with the high-resolution detailed observation capability of the bionic eye cameras. By combining these two visual components, the system achieves both comprehensive scene monitoring and focused detailed analysis, ensuring that no key information is missed while maintaining efficient processing through coordinated operation of multiple sensors.
Data Source
AI summary
Disclosed is an active positioning method for bionic binocular vision during execution of a simultaneous localization and mapping (SLAM) task. The method includes: capturing a panoramic image using a panoramic camera and detecting key scene information; assigning values to pixels in the panoramic image to obtain a panoramic value map; projecting field-of-view areas of left and right bionic eye cameras separately onto the panoramic value map, to obtain a current binocular field-of-view projection area; calculating a mean value of the current binocular field-of-view projection area in the panoramic value map; comparing the mean value of the current binocular field-of-view projection area with a value threshold, and obtaining a binocular field-of-view projection area with a mean value higher than the value threshold by moving the binocular bionic cameras; and finally using a high-value image captured by the left and right bionic eye cameras as an input of a SLAM system.