Keystone Correction Using Marker Feedback for Viewing Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image projection systems struggle to accurately correct image distortion caused by differences in the angle between the image projection device and the screen, as well as the user's viewing direction.
Innovation Solution
An image projection device and method that utilize a processor to project a first corrected image, receive a captured image from a user device, generate second keystone correction information based on marker positions in both images, and project a second corrected image to align with the user's viewing direction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If keystone correction is performed based on the angle between the image projection device and the screen, then the image distortion is corrected, but the correction does not align with the user's viewing direction
Solution Approach 1:
The system captures an image of the projected reference image using a camera, detects marker positions in both the reference image and captured image, and uses this feedback to calculate a second homography matrix that aligns the correction with the user's actual viewing direction. This feedback loop enables continuous refinement of the correction accuracy.
Solution Approach 2:
The system transitions from using a single homography matrix based on device-to-screen angle to using a second homography matrix that incorporates user device orientation parameters. This parameter change allows the correction to adapt to different user viewing angles and positions, achieving precise alignment with the user's perspective.
2Ease of manufacture
If a single homography matrix is used for keystone correction, then the correction process is simple, but the recognition rate of projected patterns is insufficient
Solution Approach 1:
The correction process is segmented into two stages: first calculating a homography matrix based on the angle between the image projection device and the screen, then calculating a second homography matrix based on the angle between the user device and the screen. This segmentation allows each stage to focus on specific correction aspects, improving overall recognition accuracy while maintaining process manageability.
Solution Approach 2:
The system first projects a reference image with markers and captures it using a camera before performing the final correction. This preliminary action provides the necessary data to calculate the second homography matrix, ensuring that the correction is based on actual captured pattern positions rather than assumptions, thereby improving recognition rate.
3Productivity
If keystone correction is performed without considering user device orientation, then the correction is fast, but the image does not align with the user's viewing direction
Solution Approach 1:
The system performs preliminary correction using the first homography matrix based on the angle between the image projection device and the screen before incorporating user device orientation. This preliminary action provides a baseline correction that can be quickly applied, while the second homography matrix refines the alignment to match the user's viewing direction, achieving both speed and precision.
Data Source
AI summary
An image projection device is disclosed. An image projection device according to an embodiment of the disclosure may comprise: memory storing information about a reference image, an image projection unit comprising a projector configured to project an image or a video, a communication unit comprising communication circuitry, at least one processor, comprising processing circuitry, connected to the memory, the image projection unit, and the communication unit. In an embodiment, at least one processor, individually and/or collectively, may be configured to control the image projection device to: project a first corrected image generated based on first keystone correction information and the reference image, onto a screen, receive, from a user device, a captured image obtained by capturing the first corrected image projected onto the screen, generate second keystone correction information based on information about positions of at least four first markers included in the reference image, information about positions of at least four second markers included in the captured image, and the first keystone correction information, and project a second corrected image generated based on the second keystone correction information and the reference image. The at least four second marker may be disposed at positions corresponding to the positions of the at least four first markers.