Case Study

Replacing the Human Eye - Automated Detection of Phosphine Gas Release Through Liquid Color Change

Chemical Manufacturing | India

Gas Safety Continuous Monitoring Color Change Detection Operator Alert

Monitoring Method (Before)

Manual

Person watching continuously

Detection Reliability (Before)

Human dependent

Fatigue risk

Monitoring Method (After)

Automated

AI watches all beakers

Detection Speed (After)

Seconds

Alert on change confirmed

THE CHALLENGE

Phosphine gas release is indicated by a gradual color change in indicator beakers - from transparent to blue. The change happens slowly over hours and unpredictably across multiple beakers at different facility points. Someone had to watch these beakers continuously to catch the moment the color shifted.

That someone was a person staring at a camera feed for hours across shifts. The moment they looked away or fatigued - which is inevitable for any passive monitoring task - there was a window where a color change could go unnoticed. For a Phosphine gas release indicator, that window carries serious safety consequences.
  • Unpredictable timing made continuous monitoring mandatory

    Color change could occur at any point across a multi-hour window with no pattern to anticipate

  • Human attention not suited for passive continuous monitoring

    Fatigue and shift handovers created inevitable gaps in a task that required zero gaps.

  • Multiple beakers required simultaneous watch

    A single person could not reliably monitor all indicator points across the facility at the same time.

  • Gradual change easy to miss

    Slow transparent to blue transition made it difficult to pinpoint the exact moment without constant focus.

THE SOLUTION

We deployed a computer vision system that continuously monitors all indicator beakers across the facility for the transparent to blue color change signalling Phosphine gas release. The system tracks all monitoring points simultaneously and alerts the operator the moment a confirmed change is detected.
Flow: Live beaker camera feed -> Baseline transparent color established -> Continuous color state monitored -> Blue change confirmed -> Operator alerted -> Logged with image and timestamp

WHAT CHANGED AFTER

Continuous beaker monitoring is no longer a human task - AI watches all points simultaneously without fatigue.

Color change detected the moment it is confirmed - no gap between change and operator awareness.

All monitoring points covered simultaneously - no beaker unobserved across any shift.

Complete detection log for every event - image and timestamp for safety audits.

Explore more case studies

View All Case Studies