Video Camera Digital Zoom for Stable AR Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern videogame devices face challenges in maintaining video frame rates and processing resources due to the varying size of users or fiduciary markers in video images, which complicates detection and tracking, especially when they move outside the camera's field of view.
Innovation Solution
A video camera system with a wide field of view lens and a high-resolution imaging sensor, coupled with a graphics processor, generates secondary images of specific regions or resolutions requested by a host device, allowing for digital panning, tilting, and zooming without mechanical components, and performs 3D to 2D mapping to correct distortions and enhance feature detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the video camera uses a fixed field of view to capture images, then the processing complexity varies when users or markers move within the scene, but if the field of view is made variable through digital zooming, then the processing resources can be stabilized while maintaining consistent feature detection quality
Solution Approach 1:
The system dynamically adjusts the field of view by generating secondary images with digital zooming based on the detected position and size of users or markers in the primary image. This allows the processing complexity to adapt to the actual scene requirements, stabilizing frame rates while maintaining detection quality.
Solution Approach 2:
The system changes the resolution and region of interest parameters by generating secondary images at different zoom levels. When users or markers are small in the primary image, the system generates high-resolution secondary images of specific regions to maintain feature detection quality, thereby stabilizing processing resources.
2Measurement precision
If the video camera captures the entire scene at high resolution, then feature detection accuracy is maintained, but the processing resources and computational load increase significantly
Solution Approach 1:
The system extracts only the regions containing users or markers from the full primary image to generate secondary images. This extraction approach maintains feature detection accuracy for the regions of interest while significantly reducing the computational resources required compared to processing the entire high-resolution scene.
Solution Approach 2:
The system applies different processing quality levels to different regions of the image. High-resolution secondary images are generated only for regions containing users or markers, while other regions are processed at lower resolution or not at all, thereby maintaining detection accuracy where needed while reducing overall computational load.
3Productivity
If the video camera uses a narrow field of view to focus on specific areas, then processing resources are reduced, but users or markers moving outside the field of view cannot be tracked
Solution Approach 1:
The system adds a temporal dimension to tracking by generating a sequence of secondary images over time. When users or markers move outside the current field of view, the system can detect their new positions in subsequent primary images and generate new secondary images, thereby maintaining tracking capability while keeping processing efficient at each individual frame.
Data Source
Figure 1~2
Figure 3A
Figure 3B
AI summary
An imaging system comprises a video camera that in turn comprises a wide field of view lens, an imaging sensor operable to form a first image having a first pixel area, and an interface operable to receive commands from a host entertainment device; in which the video camera is operable to receive a command requesting at least one secondary image having a second pixel area for a 3D mapping of a portion of the first image; and the video camera system comprises a graphics processor operable to perform the 3D mapping to generate the or each secondary image from the first image, in which the 3D mapping transforms a section of the first image depicting a surface having a normal in a first direction to form a secondary image in which the normal of that surface faces directly out of the secondary image.