REALTHON
AI Industrial Inspection
All solutionsINDUSTRIAL INTELLIGENCE

AI Industrial Inspection

Edge AI visual inspection for production lines built on the D-Robotics chip platform — defect recognition and quality records for parts and materials, lightweight deployment for online inspection across production lines

Contact a solution expert

Solution overview

This solution builds a production-line edge visual-inspection system on the D-Robotics chip platform. Station cameras capture images of materials and parts, and the on-chip AI runs locally to identify quality issues such as scratches, missing material, and dimensional deviation in real time. Inspection results are recorded and uploaded to the quality-management platform for online inspection and data traceability. The lightweight, low-power design supports rapid deployment on single or multiple production-line stations.

  • On-chip AI inference on D-Robotics hardware, with inspection available offline
  • Real-time capture of material appearance, dimensions, and other quality information
  • Low-power design for 24/7 operation on the shop floor
  • Standard interfaces for MES integration and quality-data traceability

Application scenarios

Material appearance inspection

Defect recognition on raw materials, sheets, and coils — detecting scratches, stains, and missing material.

  • High-speed capture on the line
  • Real-time defect recording and alerts
Part quality inspection

Appearance and assembly checks for parts — detecting damage, deformation, and assembly anomalies.

  • Quality-status determination
  • Results bound to material SN records

Deployment architecture

The system adopts a three-layer edge-connectivity-cloud architecture: station units handle image capture and AI inference, gateways aggregate and securely relay data, and the cloud platform provides unified quality monitoring and system integration — simple, lightweight, and extensible.

Edge: station inspection unit
Industrial cameras capture images and the D-Robotics chip runs AI inference to output inspection results.
Connectivity: transport layer
Industrial gateways aggregate station data and upload it over encrypted links.
Cloud: business platform
Quality monitoring, defect-image storage, yield statistics, and MES integration.

Hardware selection

VS capture terminal
Industrial camera + light source for capturing material and part images

Image capture at production-line stations

D-Robotics AI module
On-chip AI inference, low-power defect detection, result output

Online visual inspection on the production line

Industrial gateway
Data aggregation and forwarding, MES integration

Multi-station networked digital projects

Frequently asked questions

How well does the D-Robotics solution perform in inspection?

Low power consumption and high inference speed support millisecond-level online inspection on production lines.

How are inspection results delivered?

They support local logs and network messages, and can be connected to production dashboards and MES systems.

Can algorithms be customized?

Yes — model conversion and algorithm customization on the D-Robotics platform adapt the system to different materials and parts.

Get solution design and engineering support

Share your application, target platform, and project stage so we can define the technical path and delivery scope together.

Contact a solution expert