What happened
SENSIA, a company in Leganés near Madrid, makes the RedLook System. Its software uses computer vision and machine learning to detect emissions in video from SENSIA's cameras for optical gas imaging. In September 2025, SENSIA asked EPA to approve the system for periodic methane screening at oil and gas facilities. On 12 June 2026, EPA found that the system with the Caroline FY camera meets the 1 kg/hr screening threshold with component-level resolution. A letter of the same day approved it with the cooled Mileva camera on the same terms. EPA requires that major changes to the detection algorithm be reviewed before field use. In its approved method, SENSIA reports blinded tests at Colorado State University's METEC center in June 2024. There, the system detected 171 of 202 controlled releases that were analyzed. It reached a 90% probability of detection at 0.84 kg/hr. SENSIA notes that the tests did not compare the system with a reference method. EPA gives no test data in its letters. The tests used planned releases, so no field leaks are reported.
Sources
-
Regulatory document, by authors with no tie to the system.
Based on a review of the submitted materials, the RedLook System Caroline FY meets the periodic screening requirements for the 1 kilogram per hour (kg/hr) threshold with component-level spatial resolution. The EPA approves the solution for use by owners and operators at affected facilities for the periodic screening requirements described in 40 CFR 60.5398b(b).
-
Regulatory document, by authors with no tie to the system.
Based on a review of the submitted materials, the RedLook System Mileva meets the periodic screening requirements for the 1 kilogram per hour (kg/hr) threshold with component-level spatial resolution.
-
Report, by the makers.
This method achieved a 90% PoD of 0.48 kg/hr/m/s when corrected for wind speed (i.e. the emission rate variable was normalized for wind speed) and a 90% PoD of 0.84 kg/hr when corrections for wind speed were not made (i.e. the emission rate variable was not adjusted for wind speed).
Details
AI system
- Workflow
-
SENSIA's cameras recorded video of the site for optical gas imaging. Algorithms for computer vision and machine learning detected emissions in the video.
Follow-up
- Conditions for replication
-
Owners and operators must keep to the method's limits on wind speed and camera distance. Major changes to the detection algorithm must be reviewed by EPA before field use.
Entry details
- Sector
-
Climate change mitigation / methane emissions monitoring.
- Purpose
-
SENSIA built the RedLook System to detect methane emissions automatically in video from its cameras for optical gas imaging.
Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). RedLook camera software meets EPA's 1 kg per hour methane threshold (AIBD-0827). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0827.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)