Visual and Acoustic Event Detection for Industrial Process Monitoring

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional analytics for industrial processes are limited by reliance on expert-defined patterns, lack of real-time analysis, and high costs associated with intrusive sensor installations, leading to unnoticed abnormalities and delayed remediation.

Innovation Solution

Implementing visual and acoustic analytics using commodity cameras and microphones to acquire data, which is analyzed in real-time by a machine learning engine that adapts to detect events and generate control signals, allowing for automatic identification and control of industrial processes without manual intervention.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional instrument-based data acquisition is used, then measurement reliability is improved, but device complexity and installation cost increase due to intrusive sensor installations and wiring

Engineering Contradiction:
Improvemeasurement reliabilityVSAvoiddevice complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent replaces traditional mechanical/electrical sensors with cameras and microphones that capture visual and acoustic data. This substitution eliminates the need for intrusive physical sensors and wiring while maintaining monitoring capabilities through non-contact data acquisition methods.

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

Solution Approach 2:

The patent uses cameras to create visual copies (images and videos) of the industrial process environment, and microphones to capture acoustic copies of operational sounds. These copies serve as surrogate data sources that replace direct physical measurement by traditional sensors, enabling monitoring without physical intrusion.

Inventive Principle:
Principle #26Copying

2Measurement precision

If expert-defined predetermined patterns are used for event detection, then detection accuracy for known events is improved, but adaptability to unknown abnormalities deteriorates

Engineering Contradiction:
Improveevent detection accuracyVSAvoidadaptability to unknown abnormalities
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent implements a machine learning engine that automatically learns and adapts to the specific industrial environment without requiring manual programming of detection patterns by experts. The system serves itself by continuously training on environment-specific data, enabling automatic adaptation to both known and unknown abnormalities.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent transforms the static, fixed pattern-matching approach into a dynamic system where the machine learning model continuously adapts and evolves based on new data. The detection patterns are not predetermined but dynamically generated and updated through ongoing machine learning training processes.

Inventive Principle:
Principle #15Dynamics

3Measurement precision

If manual analysis of recorded data is performed, then detection thoroughness is improved, but productivity and real-time response deteriorate

Engineering Contradiction:
Improvedetection thoroughnessVSAvoidanalysis speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces manual human analysis with an automated machine learning-based analysis system. This substitution enables continuous real-time processing of visual and acoustic data streams, eliminating the bottleneck of manual review while maintaining or improving detection accuracy through algorithmic pattern recognition.

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

Solution Approach 2:

The patent implements continuous real-time analysis of incoming data streams from cameras and microphones. The machine learning engine processes data continuously without interruption, enabling immediate detection and response to events as they occur, rather than batch processing or periodic manual review.

Inventive Principle:
Principle #20Continuity of useful action

4Device complexity

If data is not analyzed in real-time, then system complexity is reduced, but loss of time for remediation increases

Engineering Contradiction:
Improvesystem complexityVSAvoidtime to detect and remediate abnormalities
Core Design Contradiction:
Device complexityVSLoss of time

Solution Approach 1:

The patent performs preliminary actions by continuously pre-processing and analyzing data in real-time before critical events occur. The machine learning engine is constantly trained and updated with incoming data, so when abnormalities occur, the system is already prepared to detect and respond immediately, reducing remediation time.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3469434B1Automatic visual and acoustic analytics for event detection
Publication Date: 2023.07.19 EXXONMOBIL TECHNOLOGY & ENGINEERING CO
  • EP3469434B1 patent drawingFigure 1
  • EP3469434B1 patent drawingFigure 2
  • EP3469434B1 patent drawingFigure 3

AI summary

Systems and methods are provided for detecting events in industrial processes. An acquisition system may include one of a camera and an audio recorder to acquire monitoring data in the form of one of imaging data and acoustic data, respectively. A computer system, may include a machine learning engine and may be programmed to classify the monitoring data under a classifier, quantify, based on the classifier, the monitoring data with at least one quantifier, and detect an event when the at least one quantifier satisfies a predetermined rule corresponding to the at least one quantifier.