AR Glasses Camera Angle Adjustment for Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
AR glasses currently have low object recognition accuracy, particularly in the edge regions of the shooting range, as they focus on central objects and ignore those in peripheral areas.
Innovation Solution
The implementation of an image display method in AR glasses, featuring a first camera for detecting eye rotation and a second camera for adjusting its shooting angle to capture target images, allowing for improved object recognition and dynamic adjustment of the observation region based on the user's gaze, along with features like region mapping, safety monitoring, and high-altitude object detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AR glasses focus on displaying preset virtual scenes and recognize objects only in the central region, then the display function is simple and device complexity is low, but object recognition accuracy is poor especially in edge regions
Solution Approach 1:
The patent divides the shooting range into multiple regions (central region and edge regions) and assigns different recognition priorities. The system segments the field of view to identify and process objects in edge regions separately from central objects, thereby improving overall object recognition accuracy without requiring complete system redesign
Solution Approach 2:
The patent implements dynamic adjustment of the shooting angle based on eye rotation information. The second camera's shooting angle is dynamically adjusted to follow the user's gaze direction, enabling the system to adaptively focus on different regions including edge areas where objects might be missed in static configurations
2Measurement precision
If the second camera uses a fixed shooting angle, then device complexity is low and ease of operation is high, but the camera cannot track user gaze and object recognition accuracy deteriorates
Solution Approach 1:
The patent employs feedback mechanisms where eye rotation information detected by the first camera is used to adjust the second camera's shooting angle. This closed-loop feedback system automatically tracks user gaze and adjusts the viewing direction, improving object recognition in the direction of user attention while maintaining automatic operation
Solution Approach 2:
The system performs self-adjustment by automatically modifying the second camera's shooting angle based on detected eye rotation, without requiring manual intervention. The device serves itself by autonomously tracking gaze and adjusting its observation focus, thereby maintaining ease of operation while improving recognition accuracy
Data Source
AI summary
An image display method applied to AR glasses is disclosed. An inner frame of the AR glasses is provided with a first camera, and an outer frame of the AR glasses is provided with a second camera. The image display method comprises: detecting eye rotation information of a glasses wearer by using the first camera (S101); adjusting a shooting angle of the second camera according to the eye rotation information, and acquiring a target image collected by the second camera after adjusting the shooting angle (S102); judging whether there is a target object in the target image (S103); and if yes, displaying an image including position information of the target object via a lens of the AR glasses (S104). The object recognition accuracy of the AR glasses is improved.