EMPOWERING AI WITH ADVANTECH’S SOLUTIONS
Artificial intelligence leverages computers and machines to mimic the problem-solving and decision-making capabilities of the human mind.
Although the concept of thinking machines is not new, it is only in recent years, leveraged by major advances in computing and parallel processing, that artificial intelligence has moved beyond science fiction to become a useful and cost effective real-world technology.
Gartner: “Artificial Intelligence (AI) applies advanced analysis and logic-based techniques, including machine learning, to interpret events, support and automate decisions and take action.”
AI requires a foundation of specialized hardware and software for writing and training machine learning algorithms, together with optimized platforms for edge inference to apply the results of the trained models at scale in a cost effective way.
As a leading provider of cutting-edge technology solutions, Advantech has developed a comprehensive range of AI-based AOI solutions to satisfy the needs of diverse industries.
AI Growth Market Trend
20%⇡
2022-2027 CAGR
ARTIFICIAL INTELLIGENCE APPLICATIONS
ROBOTIC

3D and AI robot vision solutions
MANUFACTURING

AI revolutionizes manufacturing
SURVEILLANCE

AI the New Eyes of Urban
Safety & Security
SMART CITY

Building smart cities
with AI video analytics
Requirements to Future-Proof Automated Optical Inspection (AOI)
3D Vision
- XYZ dimension measurements for accurate size, location, and shape identification
- Precise parts handling and pick-and-place capabilities for dynamic tasks
- Diverse object identification
Deep Learning
- AI model retraining for continuous accuracy improvement
- AI distributed architecture supports one-to-many processing to facilitate intelligent scaling according to computing requirements
Online Deployment
- Edge AI model remote configuration, troubleshooting, and monitoring
- Batch updates and deployment of AI models
- AI model event notifications
Key Factors for AOI Deployment
Data Security

- Private or public cloud adoption
- Assessment of software/hardware-based security, such as TPM2.0
- Security key storage
- Hardware Root of Trust
- Secure Boot
- Data encryption
Data Convergence

- OS compatibility
- Hardware and software integration for seamless communication
- ERP and MES system integration
- Codeless integration of OT devices such as PLC, meters, etc.
- Ability di directly integrate sensor data
Edge System Evaluation

- System platform and GPU solution selection
- Computing power evaluation
- GigE, 10 GigE data transmission bandwidth
- Assessment of edge environment durability and conditions, such as setting, installation, dimensions (space/room in the system), and temperature
- Architecture for remote management, such as backend management system
- Architecture for AI model retraining and deployment
AOI Model Building
Four Methods for AOI AI Model Training

Single-Frame
Analysis

Deep Comparison
Analysis

Deep Anomaly
Analysis

Multiple-Frame
Analysis
AI Computer Systems
Recommended AI Training Servers
Integrated Vision Systems