Taiwan environmental authority fines factory using AI smoke detection
Taiwan's Pingtung County environmental authority deployed an AI smoke detection system using image analysis and opacity measurement to identify factory emissions. The system detected illegal black smoke from a rubber factory that had evaded traditional inspections, leading to the country's first AI-based pollution fine of 210,000 NT dollars in October 2025.
What happened
Pingtung County's Environmental Protection Bureau deployed an AI smoke detection system combining surveillance imagery, cloud computing and opacity analysis to monitor factory emissions. The system automatically identifies smoke plume characteristics and uses opacity measurements directly as enforcement evidence.
The AI system detected illegal black smoke emissions from a rubber factory, which had generated multiple public complaints but evaded detection by traditional inspection methods due to intermittent discharge. The Ministry of Environment verified the data using a new standard method incorporating AI image analysis aligned with ASTM protocols.
Reported impact
The factory was issued Taiwan's first fine based on AI smoke detection technology under the Air Pollution Control Act, fined 210,000 NT dollars, which was paid in October 2025.
Index analysis
Traditional air pollution inspections were limited by the momentary nature of emissions and environmental conditions; inspectors often arrived after pollution had ceased, making violations difficult to confirm. The AI system overcomes these constraints through continuous long-term monitoring with automated detection, eliminating human subjectivity and strengthening enforcement credibility.
The system processes environmental conditions including smoke visibility from surveillance, applies cloud-based opacity calculations, and records results with digital signatures to ensure data quality and security.
Governance controls
Taiwan's Ministry of Environment established a new standard method for measuring particulate matter opacity in emission channels using image analysis, referencing ASTM methods and incorporating digital signature verification to ensure data integrity.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.