Seamless Mosaic Generation for Distributed Aperture Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current software cannot create seamless and uniform panoramic images in real-time from multiple cameras on high-performance aircraft, such as those with infrared or visible light cameras, which are essential for providing a wide-angle view to pilots in a real-world-registered mosaic format.
Innovation Solution
A method and apparatus that determine a virtual line-of-sight and field-of-view for an output mosaic, obtain input images from multiple cameras, and map contributions from these images to create a seamless mosaic, using transformation matrices and coordinate systems to ensure seamless stitching across camera boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional panoramic stitching software is used to combine multiple camera images, then a wide-field-of-view mosaic can be generated, but the processing cannot be performed in real-time and is limited to post-processing environments
Solution Approach 1:
The system performs preliminary actions by pre-calculating and storing transformation matrices for each camera's field-of-view and line-of-sight relationships. During real-time operation, these pre-computed matrices are directly applied to map input images to the output mosaic without requiring complex iterative processing, enabling real-time stitching performance.
Solution Approach 2:
The patent segments the image processing task by dividing the wide-field-of-view mosaic into discrete regions corresponding to each camera's field-of-view. Each camera's contribution is independently mapped using its own transformation matrix, and the results are combined without requiring global optimization across all images simultaneously, enabling real-time processing.
2Area of stationary object
If multiple cameras are distributed to provide wide-angle coverage, then the field-of-view increases, but achieving seamless and uniform stitching across camera boundaries becomes extremely difficult
Solution Approach 1:
The system applies local quality by using camera-specific transformation matrices that are tailored to each individual camera's optical characteristics, mounting position, and field-of-view geometry. Each camera's images are mapped with locally optimized parameters, ensuring seamless transitions at boundaries while maintaining uniformity within each camera's contribution region.
Solution Approach 2:
The patent employs parameter changes by dynamically adjusting transformation matrix parameters based on the virtual line-of-sight and field-of-view geometry. The system modifies mapping parameters for each camera according to its specific orientation and optical properties, enabling seamless stitching across diverse camera configurations while maintaining overall mosaic uniformity.
3Productivity
If a virtual line-of-sight and field-of-view system is implemented to map multiple input images to an output mosaic, then real-time seamless stitching is achieved, but the device complexity and computational requirements increase
Solution Approach 1:
The system uses copying by creating and storing transformation matrix representations (virtual models) of each camera's geometric and optical properties. These matrix copies enable rapid computation during real-time operation without requiring access to the physical cameras or their complex calibration data, simplifying the processing system while maintaining accuracy.
Data Source
AI summary
Distributed Aperture Systems use multiple staring sensors distributed around a vehicle to provide automatic detection of targets, and to provide an imaging capability at all aspects. The sensor image data is “stitched” to make the camera joints transparent to the operator. For example, images from three different cameras may be combined into a single seamless mosaic. The output mosaic is suitable for rendering on a head-steered helmet mounted display or a multifunction console display.