Barcode Recognition Using Pre-calculated Blur Reference Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for optically detecting linear barcode patterns using portable devices, such as mobile phones with cameras, fail with defocused images, as they require sharp images and are computationally intensive, making them unreliable and time-consuming.
Innovation Solution
A method that captures a raw image of a barcode, preprocesses it to generate output patterns, compares the initial pattern with pre-calculated reference patterns simulating different focus conditions, and determines the symbol sequence by finding the greatest similarity, allowing for reliable detection even with blurry images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If image processing methods are used to recognize barcodes in blurred images, then barcode detection becomes possible with fixed focus cameras, but the computational complexity increases significantly requiring minutes to analyze an image
Solution Approach 1:
The patent pre-calculates and stores reference patterns for various blur levels and barcode types before actual barcode recognition occurs. During operation, the system only needs to compare the captured blurred image against these pre-computed references, transforming a complex deblurring problem into a simple pattern matching task that executes in seconds rather than minutes
Solution Approach 2:
The system anticipates various blur conditions by pre-generating reference patterns for different degrees of defocus and barcode configurations. This beforehand preparation creates a library of expected patterns that cushions against the computational burden during actual recognition, allowing rapid matching even when the captured image is blurred
2Measurement precision
If complex deblurring algorithms are applied to blurred barcode images, then image quality improves, but processing time increases to minutes per image
Solution Approach 1:
Instead of attempting to restore the original sharp image through computationally intensive deblurring algorithms, the patent creates simplified copy representations called reference patterns that capture the essential characteristics of blurred barcodes. These reference patterns are pre-computed copies that enable rapid comparison without requiring actual image restoration
Solution Approach 2:
The system changes the approach from attempting to reverse the blur physical parameter to working directly with the blurred state by creating reference patterns that incorporate blur characteristics. This parameter change transforms the problem from one requiring complex inversion mathematics to simple pattern similarity comparison
3Reliability
If reference patterns for multiple distances are pre-calculated and stored, then detection accuracy for defocused images improves, but memory requirements and initial processing increase
Solution Approach 1:
The patent segments the reference pattern library by organizing patterns according to blur level categories and barcode types. This segmentation allows the system to store reference patterns efficiently in structured groups and to quickly select only the relevant subset for comparison, reducing both storage overhead and access time while maintaining comprehensive coverage of possible blur conditions
Data Source
Figure 1
Figure 2~6
Figure 4
AI summary
The invention relates to a method for optical recording of a linear barcode pattern and determination of a symbol sequence as defined by the barcode pattern from a blurred raw image and comprises the following steps: taking a raw image (6) of the barcode pattern, preprocessing (7) the raw image (6) to generate a series of output patterns (9), comparing(10) the output pattern (9) with stored reference patterns (5), wherein the reference patterns (5) correspond to blurred images of sharp nominal patterns (2) and each reference pattern is provided with at least one symbol, determining for each of the output patterns (9) the reference pattern (5) which most closely resembles the output pattern (9) outputting, transmitting or storing the symbols corresponding to the determined reference pattern (5) and following the sequence of output patterns (9), as a symbol sequence.