Pattern Stitching Algorithm for Machine Readable Symbol Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Imaging systems face challenges in decoding machine-readable symbols when the images are corrupted or partially unreadable due to factors like movement, illumination, or surface features, as existing methods struggle to align and combine multiple images effectively.
Innovation Solution
The system employs a data stitching algorithm to convert symbol data regions from multiple images into binary matrices, accumulating them to generate a decodable representation, and uses correlation techniques to adjust positions and scales, enabling the stitching of images even when the symbol has moved, thus overcoming the limitations of prior art.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple images are captured to improve decoding reliability, then the reliability of symbol decoding is improved, but the complexity of image processing and alignment increases
Solution Approach 1:
The patent segments the symbol data region from multiple images and processes each segment individually. Binary matrices are generated for each image's symbol data region, then these segmented binary matrices are accumulated through a stitching algorithm to form a complete decodable representation, reducing the overall processing complexity while maintaining reliability
Solution Approach 2:
The patent introduces binary matrices as an intermediary representation between the captured images and the final decoded symbol. By converting image data to binary matrices and accumulating these matrices, the system creates a simplified intermediate form that facilitates reliable decoding while managing processing complexity
2Loss of information
If multiple images are captured to compensate for movement and corruption, then the readability of the symbol is improved, but the difficulty of aligning and stitching images increases
Solution Approach 1:
The patent replaces complex mechanical alignment systems with a data-level stitching approach. Instead of physically aligning images through complex mechanical means, the system uses a data stitching algorithm that operates on binary matrices, automatically establishing correspondence between images captured at different times and positions
Solution Approach 2:
The patent changes the parameter space by working with binary matrices rather than raw image data. This transformation simplifies the alignment and stitching process by converting continuous image parameters into discrete binary representations that are easier to match and accumulate across multiple images
3Productivity
If individual images are used for detection, then the processing speed is maintained, but the decoding accuracy decreases when images are corrupted
Solution Approach 1:
The patent merges binary matrices from multiple individual images into a single accumulative binary matrix. This combining process preserves the processing speed advantage of individual image detection while improving decoding accuracy by accumulating data across multiple images, allowing corrupted regions in one image to be compensated by corresponding regions in other images
Data Source
Figure 1
Figure 2~8
Figure 9~10
AI summary
A system for establishing a correspondence between a first image and a subsequent image of a machine readable symbol, the system comprising: an imaging device, the imaging device adapted to acquire the first image and the subsequent image, the first image including a first candidate region for the machine readable symbol, and the subsequent image including symbol data; and a processor that is operatively coupled to the imaging device and is adapted to execute a data stitching algorithm that: generates a synthetic model of the machine readable symbol, the synthetic model being a model of a plurality of known features of a particular symbol; correlates the synthetic model and the symbol data to find a position association between the first candidate region and the symbol data; and based on the position association, refines the subsequent image to produce refined subsequent image, such that the first image and the subsequent image can be stitched together for decoding.