AI for EOR

เอไอสนับสนุนงานสำรวจวัตถุระเบิด

Explosive Ordnance Reconnaissance decision support: visual identification, classification, hazard assessment and structured incident documentation for trained personnel.

ระบบสนับสนุนการตัดสินใจสำหรับงาน EOR โดยช่วยค้นหา ตรวจสอบ จำแนก ประเมิน และบันทึกข้อมูล

View ResearchDemoArchitecturePaperDatasetCodeAPI

Problem

Reconnaissance teams must classify unfamiliar ordnance quickly from limited visual evidence, often without reference material in the field.

Research Question

Can retrieval-grounded visual classification improve identification confidence for field reconnaissance?

Objectives

Support search, verification, classification, assessment and documentation
Ground every answer in a retrievable reference record
Keep the system strictly non-operational in scope

Methodology

01Reference corpus construction
02Grounded retrieval
03Expert-in-the-loop evaluation

Architecture

Field EvidenceImages, notes

Captured by trained personnel

Visual IdentificationVision model

Candidate matches with confidence

Knowledge RetrievalGrounded RAG

Cited reference records

Hazard AssessmentRule + model

Risk context for the team

DocumentationStructured report

Auditable incident record

Results & Benchmark

Scope: Identification and assessment only

Limitations

· The system does not and will not describe assembly, modification or operational use of explosive ordnance
· Intended solely for trained personnel as a support tool, never as an authority