Metal Detection Using Multi-Frequency Machine Learning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional metal detectors face challenges in achieving high discrimination accuracy while minimizing false alarms, particularly when detecting small metal contaminants amidst product effects.

Innovation Solution

A computer-implemented method and system that utilizes a metal detector unit with a source and receiver unit for electromagnetic fields, coupled with a machine learning module trained to process detection data and identify the presence of metal with high accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the sensitivity of the detector unit is increased to detect smaller metal contaminants, then the discrimination accuracy improves, but false alarms due to product effect increase

Engineering Contradiction:
Improvediscrimination accuracyVSAvoidfalse alarm
Core Design Contradiction:
Measurement precisionVSObject-generated harmful factors

Solution Approach 1:

The detection process is segmented into multiple independent frequency channels. The detector unit operates at multiple frequencies simultaneously, allowing the system to analyze different signal components separately. This segmentation enables the machine learning module to distinguish between product effects (which may dominate at certain frequencies) and actual metal contaminants (which manifest differently across frequencies), thereby reducing false alarms while maintaining high discrimination accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system changes the operating parameters by utilizing multiple frequencies instead of a single frequency. The machine learning module is trained to recognize patterns across these different frequency parameters. By analyzing the spectral characteristics and phase relationships at multiple frequencies, the system can differentiate between product effects and metal contaminants more effectively, resolving the contradiction between sensitivity and false alarm rate.

Inventive Principle:
Principle #35Parameter changes

2Ease of operation

If conventional discrimination algorithms are used, then the system operation is simple, but the ability to distinguish small metal contaminants from product effects is insufficient

Engineering Contradiction:
Improvesystem operation simplicityVSAvoidcontaminant detection accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent replaces conventional mechanical discrimination algorithms with a machine learning-based computational system. The machine learning module automatically learns complex patterns and relationships in the multi-frequency detection data, substituting simple rule-based algorithms with an intelligent system that can handle the complexity of distinguishing small metal contaminants from product effects while maintaining ease of operation through automated processing.

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

Solution Approach 2:

The discrimination approach combines multiple detection frequencies and signal parameters into a composite analysis framework. The machine learning module processes a composite of electrical signal components, phase information, and spectral characteristics together, rather than analyzing single parameters in isolation. This composite analysis significantly improves detection accuracy while the automated nature of the machine learning process maintains operational simplicity.

Inventive Principle:
Principle #40Composite materials

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 solution enables the detection of small metal contaminants with improved sensitivity and reduced false alarms, effectively addressing the limitations of conventional metal detection systems.

Implementation Method 1

a source unit for an electromagnetic field and a receiver unit electromagnetically coupled thereto, wherein said source unit, said receiver unit and said transport path are arranged relative to each other so as to output an electrical signal dependent on a metal content within said object

Methodology Applied
Scientific EffectElectromagnetic induction: Electromagnetic Induction

Data Source

PatentEP4560362A1Computer-implemented method, system and computer program for detecting the presence of metal in an object
Publication Date: 2025.05.28 METTLER TOLEDO SAFELINE LTD
  • EP4560362A1 patent drawingFigure 1~2
  • EP4560362A1 patent drawingFigure 3~4
  • EP4560362A1 patent drawingFigure 5~6

AI summary

The present invention relates to a computer-implemented method for detecting the presence of metal in an object (100) transported, preferably by a belt conveyor (101), along a transport path through a metal detector unit (10), said method comprising: providing detection data to electronic data processing means (20), said detection data being created by the metal detector unit (10), the metal detector unit (10) comprising a source unit (11) for an electromagnetic field and a receiver unit (12) electromagnetically coupled thereto, wherein said source unit (11), said receiver unit (12) and said transport path are arranged relative to each other so as to output an electrical signal dependent on a metal content within said object (100), said metal detector unit (10) being further operative to derive said detection data from said electrical signal; processing the provided detection data by a metal detection software module executed by the electronic data processing means (20, 331, 421), said metal detection software module comprising a first machine learning module (22) receiving said detection data as an input and having been trained to automatically identify the presence of metal in the object (10) on the basis of said detection data, said first machine learning module (21) being further configured to output identification data indicative of the presence of metal identified by said first machine learning module (22), and to a system and to a computer program for detecting the presence of metal in an object (100).