Banknote Management System Using Neural Network Classification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing banknote management systems face challenges in accurately and efficiently collecting and identifying banknote information, particularly with issues related to orientation, character positioning, data volume, and robustness under conditions of damage or dirt, leading to low identification efficiency and robustness.

Innovation Solution

A banknote management method that collects features using image, infrared, fluorescence, and magnetism, and employs edge detection, image rotation, adaptive binarization, moving window registration, and normalization to improve the accuracy and efficiency of prefix number identification, using a convolutional neural network for classification.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If simple binarization and size processing are used for character segmentation, then the processing is simple and fast, but the character range is large leading to wrong segmentation and heavy data volume that reduces identification efficiency

Engineering Contradiction:
Improveidentification efficiencyVSAvoiddata volume
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The patent applies segmentation by dividing the banknote image into multiple scanning lines and processing each line independently. The character segmentation is achieved through line-by-line scanning with neural networks, breaking down the large character range into smaller manageable units, which reduces wrong segmentation and decreases the effective data volume for processing.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimension of processing by using line-by-line scanning instead of processing the entire image at once. This dimensional approach transforms the 2D image processing into a sequence of 1D line processing tasks, reducing the computational burden and data volume while maintaining identification accuracy.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If DSP identification method with template matching and neural networks is used, then identification can be achieved, but the network transmission efficiency is limited and identification speed is slow

Engineering Contradiction:
Improveidentification accuracyVSAvoididentification speed
Core Design Contradiction:
Measurement precisionVSSpeed

Solution Approach 1:

The patent applies preliminary action by performing edge detection and orientation correction before the main identification process. The banknote orientation is determined and corrected in advance, and the image is preprocessed with edge detection to highlight character boundaries. This preliminary processing reduces the complexity of subsequent neural network identification and improves overall identification speed.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional mechanical DSP processing with a more efficient neural network-based approach. Instead of using conventional template matching and mechanical image processing methods, the system employs neural networks for orientation determination and character recognition, which significantly improves identification speed while maintaining or enhancing accuracy.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If edge detection and orientation correction are performed, then the banknote orientation can be identified, but the rapid slope change caused by banknote delivery cannot be well adapted and slope correction is delayed

Engineering Contradiction:
Improveorientation identification accuracyVSAvoidslope correction time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies continuity of useful action by performing orientation determination and slope correction in a continuous, integrated manner throughout the image processing pipeline. Rather than treating these as separate discrete steps, the system continuously adjusts for slope changes during the scanning and identification process, ensuring that orientation correction keeps pace with rapid slope changes caused by banknote delivery variations.

Inventive Principle:
Principle #20Continuity of useful action

4Adaptability or versatility

If traditional identification algorithms are used, then they can handle complete banknotes, but the identification robustness of damaged banknotes is low and no specific processing methods are provided

Engineering Contradiction:
Improvehandling complete banknotesVSAvoididentification robustness of damaged banknotes
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent applies parameter changes by adjusting the neural network parameters and processing thresholds based on the condition of the banknote. For damaged banknotes, the system modifies identification parameters and uses adaptive thresholding to accommodate missing or degraded features. This allows the system to maintain high identification robustness across both complete and damaged banknotes by dynamically adjusting processing parameters.

Inventive Principle:
Principle #35Parameter changes

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

The method achieves high-efficiency and accurate banknote information collection and identification, enhancing robustness and reducing data processing volume, while ensuring fast operation and compatibility with existing equipment like ATMs.

Implementation Method 1

collects, identifies and processes banknote features of banknotes in corresponding services by ways of image, infrared, fluorescence, magnetism and thickness

Methodology Applied
Scientific EffectImage: Photography

Implementation Method 2

collects, identifies and processes banknote features of banknotes in corresponding services by ways of image, infrared, fluorescence, magnetism and thickness

Methodology Applied
Scientific EffectInfrared: Infrared Radiation

Implementation Method 3

collects, identifies and processes banknote features of banknotes in corresponding services by ways of image, infrared, fluorescence, magnetism and thickness

Methodology Applied
Scientific EffectFluorescence: Fluorescence

Implementation Method 4

collects, identifies and processes banknote features of banknotes in corresponding services by ways of image, infrared, fluorescence, magnetism and thickness

Methodology Applied
Scientific EffectMagnetism: Magnetism

Data Source

PatentEP3460765B1Banknote management method and system
Publication Date: 2023.02.01 JULONG CO LTD
  • EP3460765B1 patent drawingFigure 1~2
  • EP3460765B1 patent drawingFigure 3~5
  • EP3460765B1 patent drawingFigure 6

AI summary

Provided in the present invention is a banknote management method. The method comprises: acquiring, identifying, and processing banknote features by a banknote information processing apparatus, so as to obtain banknote feature information; transmitting the banknote feature information, service information, and information about the banknote information processing apparatus together to a main control server; and the main control server processing the received information and classifying banknotes. Also provided is a banknote management system for the banknote management method. The method of the present invention can enhance robustness of identification while maintaining an operation speed, thus ensuring accuracy and practicability in actual applications.