Case studies

Manufacturing · Vietnam, South Korea, Hong Kong

Solution Portfolio for the Manufacturing Industry

A roundup of NTQ projects for manufacturing: customs & inventory management, a Smart Warehouse with AGVs, AI detection of PCB/relay defects, QR-based production management for a heat treatment plant, a WMS for a dairy plant, worker action recognition and WiFi IoT devices.

C#.NET Core 6.0ASP.NET BoilerplatePostgreSQLJavaJSPPHPMySQLPythonOpenCVTensorFlowTensorRT

Project overview

  • Industry: Manufacturing
  • Country: Vietnam, South Korea, Hong Kong (depending on the project)
  • Project type: Web Application, AI Solution (Computer Vision), IoT/Embedded
  • The portfolio includes 8 representative projects, detailed below.

Representative projects

1. Customs, code conversion and inventory management system (Order Management System)

  • Country: Vietnam · Scale: 10.38 MM · Project type: Web Application
  • Technology: C#, .NET Core 6.0, ASP.NET Boilerplate, PostgreSQL
  • Key features:
    • Material Master: manage/import materials, customise fields/groups, print goods receipt notes.
    • Master BOM: dynamic BOM management, BOM data import, inventory alert settings.
    • Customs Code Management: manage/import customs codes, convert internal codes for declarations.
    • Product Conversion and Export: manage product lists, automatic aggregation, link exports with customs.
    • Inventory and Reporting: manage physical inventory/customs inventory, print documents, import-export reports.

2. Smart Warehouse – intelligent warehouse system

  • Country: South Korea · Scale: 76.5 MM · Project type: Web Application
  • Technology: Java, JSP, HTML5, CSS3, JavaScript, SQL
  • Problem: an integrated logistics control and warehouse management system to maximise space utilisation and storage efficiency and increase inventory turnover through fast warehouse processing. The client wanted to build a core team for pilot projects ranging from simple to complex, and to train the team for long-term development.
  • Key features:
    • Material (WIP) handling for just-in-time supply and transport between process stages.
    • Building an automated warehouse using AGV transport robots.
    • Maximising cargo volume and warehouse turnover with shuttles and stacker cranes.
    • Managing inbound/outbound data and tracking inventory to ensure accuracy and eliminate unnecessary stock.
    • Monitoring intra-logistics for stable processes and efficient warehouse management.
  • Benefits: greater capacity, higher productivity, flexibility & scalability, safety, cost and energy savings.

3. AI PCB defect detection (PCB Error Detection)

  • Scale: 12 MM · Project type: AI Solution
  • Problem: post-production PCB inspection required many workers checking manually with magnifying glasses; there are about 50 common defects and 10 rare ones.
  • Solution: AI object detection to locate defects on the PCB; image processing to classify defects and make pass/fail decisions.
  • Technology: Python 3, OpenCV, TensorFlow, TensorRT; Object detection, image processing.

4. AI relay defect detection (Relay Error Detection)

  • Scale: 10 MM · Project type: AI Solution
  • Problem: the relay production line relied on manual inspection, with low productivity and wasted labour; 5 types of defect needed to be detected: spike, glue on terminal, standoff height, soldering on standoff, lack pre-soldering.
  • Solution:
    • Object segmentation to crop 2 objects: terminal and standoff.
    • Terminal classification model: OK or NG (Spike, Glue, Lack pre-soldering).
    • Standoff classification model: OK or NG (soldering on standoff); calculating the standoff image height to detect "Standoff height" defects.
  • Technology: Python 3, OpenCV, TensorFlow Lite; Deep Learning: Object detection, Object segmentation, Image Classification.

5. Production management system for a heat treatment plant (Manufacture Management System)

  • Problem: control production at a heat treatment plant; generate QR codes for products and orders and drive management processes with QR codes.
  • Key features:
    • Back End: User & User group management, master data, data settings for functional screens.
    • Front End: W10X – Tag & Lot Management; W20X – Delivery Order Management; W30X – Stock Management; W40X – Return Order Management.
  • Scope of work: Coding, Integration, Test, Deployment, Training.

6. Warehouse management system for a dairy plant (Warehouse Management System for Factories)

  • Client: a reputable dairy products supplier in Hong Kong · Project value: 100K – 500K USD
  • Technology: PHP, Java, HTML, MySQL
  • Key features: Dashboard, Purchase Order management, multi-warehouse operations, supplier management, multi-user, Delivery Tracking, customer management, goods receipt tracking, warehouse transfers, inventory returns.

7. AI worker action recognition (Action Recognition)

  • Scale: 2.5 MM · Project type: AI Solution
  • Problem: detect and recognise whether operators' actions are OK or NG when a machine in the plant malfunctions.
  • Solution: Object Detection to crop people's positions from video; an Action Recognition model based on an LSTM architecture to recognise actions.
  • Technology: Python 3, OpenCV, TensorFlow; Deep Learning: Object Detection, Object Tracking; deployed on Edge Devices.

8. WiFi IoT devices

  • Scope of work: Specification, Design, Coding, Production
  • Key features:
    • Base Board: programmable, connects to a WiFi access point and transmits data over WiFi, receives data from the Extension Board and forwards it to the server.
    • Extension Board: reads sensor data and sends it to the Base Board.
    • Programmer Board: bridge between the computer and the Base Board, helping users flash code onto the Base Board.

Have a similar project?

Contact SSD