Image Stitching System Using Custom Code Modules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image stitching technologies struggle to seamlessly combine images from multiple cameras, resulting in disjointed effects due to differing perspectives and lack of computational efficiency, which can lead to image distortion and incomplete situational awareness.
Innovation Solution
A system utilizing custom code modules with FPGAs, DSPs, and edge processors to coordinate imagers, employing bow-tie warping and adaptive matching algorithms to reduce computational load, ensure full 360-degree coverage, and accurately stitch images while identifying hazards and keep-out regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If multiple images from different perspectives are combined, then situational awareness is improved, but image distortion and disjointed effects increase
Solution Approach 1:
The system divides the field of view into multiple overlapping regions captured by separate imagers. Each imager captures a specific segment (e.g., 120-degree FOV) and the system processes these segmented images independently through custom code modules before combining them, allowing precise control over each segment while achieving comprehensive coverage.
Solution Approach 2:
Custom code modules written in HDL act as intermediaries between the imagers and the final stitched output. These modules perform coordinate transformations, image processing, and stitching operations, serving as a mediator that reconciles the different perspectives while minimizing distortion through algorithmic correction.
2Loss of information
If conventional image stitching methods are used, then image combination is achieved, but computational load increases
Solution Approach 1:
The patent merges the image capture and processing functions into integrated custom code modules that combine multiple imager outputs simultaneously. By merging the processing of overlapping regions and performing coordinated transformations in unified HDL code, the system reduces redundant computations and achieves efficient stitching with lower computational load.
Solution Approach 2:
The system replaces conventional software-based image stitching with hardware-implemented custom code modules in HDL. This substitution of mechanical/computational approach with dedicated hardware logic significantly reduces computational load and processing time while maintaining comprehensive image coverage.
3Manufacturing precision
If overlapping fields of view are used, then seamless stitching is improved, but device complexity increases
Solution Approach 1:
The system applies different processing qualities and transformations to different regions of the stitched image. Overlapping regions receive specialized processing through custom code modules that handle coordinate transformations and blending locally, while non-overlapping regions maintain their original quality. This local quality approach ensures seamless stitching without uniformly increasing complexity across the entire system.
Data Source
AI summary
A system for stitching images together is disclosed. The images are sometimes referred to as frames, such as frames in a video sequence. The system comprises one or more imagers (e.g. cameras) that work in coordination with a matching amount of custom code modules. The system achieves image stitching using approximately one third the Field of View (FOV) of each imager (camera) and also by increasing the number of imagers to be above a predetermined threshold. The system displays these stitched images or frames on a computer monitor, either in a still-image context but also in a video-context. Normally these tasks would involve a great detail of computation, but the system achieves these effects while managing the computational load. In stitching the images together, it is sometimes necessary to introduce some image distortion (faceting) in the combined image. The system ensures no gaps in any captured view, and assists in achieving full situational awareness for a viewer.