Value Bill Identification Using Stable Color Sub-segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional value bill identification methods based on color data face challenges due to instability in color features, leading to low accuracy, especially when the surface is non-solid-colored or has varying textures, causing incorrect identification of easily confused objects like 'O' and 'D' characters.
Innovation Solution
A value bill identifying method that uses a stable sub-segment mean set of color data to address color shift issues, involving preprocessing, feature extraction as a one-dimensional vector with smaller hue variation, and template matching to accurately identify value bills by simulating the operation mode of a color sensor and converting data into HSL color space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If color data is collected from non-solid-colored surfaces, then the identification method can handle diverse bill surfaces, but the color data becomes unstable and deviates from true color
Solution Approach 1:
The patent divides the color data into multiple sub-segments along the scanning direction and selects stable sub-segments with smaller hue variations. This segmentation approach allows the system to handle non-solid-colored surfaces by focusing on the most stable color portions, thereby maintaining reliability while preserving adaptability to diverse bill surfaces.
Solution Approach 2:
The patent applies different processing strategies to different regions of the color data. By identifying and selecting sub-segments with smaller hue variations (stable sub-segments) versus those with larger variations (unstable sub-segments), the system applies local quality filtering to maintain color data stability in regions where it is most reliable, while still capturing color information from diverse surfaces.
2Adaptability or versatility
If color features are extracted from regions with different textures, then the method can process varied bill surfaces, but the reflected intensity varies causing color data instability
Solution Approach 1:
The patent segments the color data into multiple sub-segments and evaluates the hue variation within each segment. By selecting sub-segments with smaller hue variations, the system effectively filters out regions with high texture variation that cause intensity fluctuations, thereby improving color data accuracy while maintaining the ability to process varied surfaces.
Solution Approach 2:
The patent extracts only the stable sub-segments with smaller hue variations from the overall color data. This extraction approach removes the problematic portions (regions with large hue variations caused by texture differences) while retaining the useful color information from stable regions, thus improving measurement precision without sacrificing processing capability.
3Measurement precision
If high-resolution image collection is used, then identification accuracy improves, but hardware cost and complexity increase
Solution Approach 1:
The patent extracts stable color features from the collected image data by selecting sub-segments with smaller hue variations. This feature extraction approach allows the system to achieve high identification accuracy using standard hardware by transforming the color data into a more stable and discriminative representation, thereby improving measurement precision without requiring high-end hardware.
Solution Approach 2:
The patent changes the parameter representation from raw color values to stable sub-segment mean values with reduced hue variation. This parameter transformation enhances the discriminative capability of the features, allowing standard hardware to achieve accuracy comparable to high-resolution systems by optimizing how color information is represented and processed.
Data Source
AI summary
Provided is a value bill identifying method, which includes: step 1, collecting, by a color collection device including multiple color sensors, color data of a to-be-detected value bill and preprocessing the collected color data; step 2, extracting a feature from the preprocessed color data; step 3, matching the extracted feature with feature template sets corresponding to each type of value bills, to obtain matching scores, and regarding a feature template with the highest score as a matched template of the color data; and step 4, determining a type of the value bill based on a matching result.