Odakyu to brake trains when cameras see someone inside a crossing

AI system

System: AI踏切画像解析システム (AI crossing image-analysis system) using existing crossing cameras linked to signaling equipment
Model: Not assessed
Release: Not assessed
Application developer: Meitetsu EI Engineer, Toyota Systems and Toho Electric Industries
Role: Not assessed
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Workflow: Existing crossing cameras with sufficient coverage of the enclosed area. Integration with signaling equipment able to command train braking. Three years of trials covering night and rain.
Human oversight: Not assessed
Relation to AI: The sources describe the intended detection-to-brake chain and the trial period; no detections, false alarms or accuracy figures are reported.
AI contribution: Not assessed

Outcome identifiers: Not assessed
Benefit: ITmedia (22 Jun 2026): from 24 Jun 2026 AI analyzes existing crossing cameras, tracks pedestrians, wheelchair users, cyclists and motorcyclists and sends a brake signal to approaching trains; devel...
Category: Increased Product Safety
Status: Not assessed
Beneficiary groups: Society. Organizations. People.
Beneficiaries: Not assessed
Reach: Not assessed
Distribution and access: Not assessed
Duration: Not assessed
Expected?: Yes
Costs and harms: Not assessed

Participants: Developer: Meitetsu EI Engineer, Toyota Systems and Toho Electric Industries. Operator: Odakyu Electric Railway.
Jurisdictions: Developer: Japan. Operator: Japan.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Conditions for replication: Detection accuracy holds in poor weather and lighting. False detections do not stop trains so often that the function is disabled. Signaling interlocks remain in place.
Replication attempts: 2 follow-up entries. Benefit recurred: 1. Outcome unresolved: 1.
Later findings: No development after the 24 June 2026 start was found; railf.jp says only that wider operation will be considered.
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Assessed benefits: Society. Organizations. People.
Depth: Not assessed
Reasons: Society: No outcome is documented; the sources describe the start of operation. Organizations: No operational outcome is documented. People: No individual outcome is documented.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: Independently reported

Kind: Not assessed
Related entries: Not assessed
Field: Railway safety. Level-crossing obstacle detection.
Purpose: The system is built to detect people left inside a closed crossing and stop approaching trains automatically.
Dates: Date basis unassessed: 2026 (precision unassessed). Updated: 21 October 2026.
Stage: Not assessed
Version: SHA-256: cb230cae9b81…

News report Independent author

岡田有花 (22 Jun 2026). 踏切に取り残された人をAIで検知→列車を自動停止小田急が実運用. ITmedia NEWS.
News report Independent author

railf.jp (25 Jun 2026). 小田急，AIを活用した踏切画像解析システムを導入.
Registry label Independent author

OECD AI Incidents and Hazards Monitor, incident 2026-06-22-b4f8: AI System Deployed to Prevent Railway Crossing Accidents in Japan.