Even unseen defects, made perfect with AI vision.

prev
Quality Reliability

AI-based visual inspection detects tiny defects, contamination, and deformation Automatically detects defects and strengthens harness quality reliability

Inspection Efficiency

Compared to traditional inspection, dramatically reduces inspection time Automates repeated and labor-intensive inspection work, improving efficiency

Quality Management System

Inspection results are automatically collected and analyzed Allows defect statistics and root-cause analysis

NEXT

Invisible Limits(Visual Inspection)

  • Long inspection time

    Average 30–40 seconds (SEAL, BRKT, CONN, etc.)
    Total work time: 961 hours

  • High manpower dependency

    Average required personnel: 6 people

  • Quality control limitations

    Missed micro-defects, data omission, etc.

Increased inefficiency, decreased reliability

Smart Insight, Perfect Quality(AI Vision Inspection System)

Software

(Intelligent Analysis & Integrated Control)

  • AI analysis algorithm(Includes defect detection and defect type classification)

  • Integrated quality monitoring system(Real-time quality status & data analysis)

Hardware

(High-precision inspection infrastructure)

  • Inspection cameras(Ultra-high-resolution images)

  • Industrial PC(Powerful AI computation & learning engine)

  • Target components(Wiring harnesses, fuse boxes, etc.)

System Configuration

prev
Vision Inspection Equipment
HW & Control SW

We build an optimal hardware environment and provide a control system integrated with AI inspection modules.

AI-Based Vision Inspection SW

Our self-developed AI determines
defect status in real time.

Integrated Quality Monitoring System

Inspection results are visualized in real time and quality data is provided. Through API linkage, data analysis and part management functions are supported.

NEXT
Applicable Parts
PREV
NEXT

ML Ops-based Algorithm Optimization

MLOps-based algorithm optimization for defect detection in incoming inspections of harness components

STEP 01 Object & Defect Location Labeling

  • Normal / Defective Parts

  • Labeled Data

Parts Data Collection

STEP 02 Harness Part Region Extraction

EfficientNet, YOLO model-based
harness component area extraction

STEP 03-01 Context-Aware Detection

Abnormal Region Detection (Applied Heatmap, PatchCore / PaDim)

STEP 03-02 Context-Aware Detection

Object Detection (PaddlePaddle, RT-DETRv2 Object Detection)

STEP 04 Final Judgment & Feedback

AI-based visualization of inspection results
Generate diagnostic reports

Automating the entire lifecycle of data management, model training, deployment, and monitoring.
PREV
NEXT

AI Vision Inspection Monitoring System

  • Builds a monitoring server that stores and retrieves AI vision inspection results for key harness components and provides a variety of functions.
  • Main functions: dashboard, AI vision inspection equipment management, inspection result search and statistics, system status monitoring, etc.
With innovation in wiring harnesses,
we are beginning with an AI vision inspection system
to drive AX manufacturing automation.