AI and Machine Vision quality control software

About the project

The client specializes in industrial vision-based software and has more than ten years of experience in this field. The solution was created to support automated quality control in manufacturing by combining Machine Vision with Artificial Intelligence, helping companies detect defects faster and maintain consistent production standards.

Challenge

From a business standpoint, the solution had to address several common challenges manufacturers face:

• The need to launch new products within very short timeframes
• Rising customer expectations for flawless product quality
• Strong cost-efficiency to remain competitive in the market

On the technical side, the complexity of the project involved:

• Building a full-scale MVVM application with a modern user interface, multiple layers, user permissions, and configurable settings
• Managing stable communication between the WPF UI and device-side software written in C++

The target devices were image-based barcode readers equipped with high-resolution sensors. These readers are widely used across industries such as automotive, electronics, intralogistics, catering, packaging, document processing, and healthcare.

AI and Machine Vision quality control software

Solution

Every product passes through multiple stages of quality control. When the manufacturing workflow is optimized, it becomes easier to maintain a stable quality level and reduce defects. Machine vision makes it possible to capture real-time data directly from the shop floor and automate inspection without slowing down production. In real-world environments, the system operates reliably with minimal maintenance and almost no downtime.

We developed a desktop User Interface client application that allows engineers to configure reading devices in both Stand Alone and Master/Slave modes. The software runs on Windows-based PCs (typically laptops) and connects to devices through an Ethernet TCP/IP interface. It also supports visual monitoring of images stored in an Image Database, whether the data is saved locally on the device or on a local or remote computer.

Key features:

• PackTrack™ Calibration for omnidirectional reading and tracking stations used in logistics scenarios
• Station-level statistics and diagnostic insights for better monitoring and troubleshooting
• Device communication powered by Apache Thrift, enabling efficient multilingual service interaction through a binary protocol

To make complex workflows easier for operators and engineers, we created a set of advanced custom controls that support tasks such as:

• Clustering multiple devices and managing them through Multiview
• A visual setup that clearly represents complex I/O interactions across many devices and views at once
• Managing the device learning process using the photos captured during operation

The application automatically adapts its interface and functionality based on device capabilities and user permissions. This makes it easy to switch between different views and use cases, including liquid lens calibration, basic monitoring, admin configuration, and PackTrack™ Calibration.

• Automated collection of statistics and decoded data to support smoother and more efficient operations
• Simultaneous device monitoring from multiple remote PCs
• Three separate user access levels for better control and security
• Real-time configuration of user and session language settings
• System-wide configuration options
• Dynamic content with automatic page updates
• Ability to test different devices using built-in simulators

Technologies

TypeScript

Xsd databases

IoC

Unit Tests

csharp

WPF

MVVM

XAML

C++

Business value

This solution helps manufacturers unlock the full potential of Machine Vision and AI to automate quality control and gain valuable production insights through software-driven analytics. By improving automated inspection, companies can manage a growing range of products more effectively while also reducing the time-to-market cycle. As a result, production becomes faster, more consistent, and easier to scale without compromising on quality.