Cognex In-Sight 3800 Vision System: AI-Powered Quality Inspection and Automation
Key Highlights
- The In-Sight 3800 can inspect up to 1,200 parts per minute, leveraging multi-torch illumination for enhanced surface contrast.
- Built on hybrid AI, it merges AI-based edge learning with rule-based algorithms to handle complex and variable inspection tasks efficiently.
- Multi-spectrum imaging allows the system to project different wavelengths onto targets, revealing surface imperfections invisible under standard lighting.
The In-Sight 3800 Vision System was one of several demonstrations at the Cognex booth at Pack Expo 2025 held in Las Vegas (Sept. 29- Oct. 1, 2025).
The vision system can inspect 1,200 parts per minute while leveraging multi-torch illumination. It essentially enhances image contrast across materials and surface finishes, said Alan Lunardhi, technical sales engineer, Cognex.
Hybrid AI Enables Smarter Inspection
In-Sight 3800 is built on a hybrid AI framework and real-time processing capabilities, noted a Cognex press release. This is a welcome feature for engineers advancing inspection, part tracking and robotic guidance in digitally connected manufacturing environments.
At the system’s core is an embedded suite of inspection tools that merges AI-based edge learning with rule-based vision algorithms. Cognex has transitioned from traditional rule-based inspection methods to example-based learning in this system. This is a form of built-in AI that enables the system to handle more complex or variable inspection tasks, Lunardhi explained.
READ MORE: Cognex Displays Vision Technology That Automate Quality Control Tasks in a Cinch
Whereas traditional vision tools tend to rely on mathematical parameters and look for example at contrast thresholds, brightness and edge detection tools to identify defects or verify dimensions, these approaches generally work well only for highly consistent parts with fixes measurement. However, example-based learning can be applied where acceptable parts naturally vary in appearance or perhaps where defects present inconsistently, Lunardhi said.
Example-based learning enables the system to train on a small set of images. It learns to recognize acceptable variations without extensive datasets or complex programming.
“This is a great way of embedding AI into our systems and getting a Delta where we’re facing more nuanced inspections,” said Lunardhi.
Multi-Spectrum Imaging for Nuanced Defect Detection
At the packaging and processing show, the platform was set up to verify blue label alignment and bottle cap presence on medical vials. Lunardhi pointed out how the changing appearance of the caps in a monochrome display illuminated the system’s use of multi-torch illumination. The technique sequentially projects red, green, blue, white and infrared light onto the target. By capturing the images under each wavelength, the system enhances contrast and shows surface characteristics or imperfections that are undetectable or invisible when using standard lighting conditions.
READ MORE: Currency Inspection Demonstration Using High-fidelity 3D Profile Sensor
The multi-spectrum approach is valuable for color inspections across color-sensitive or material-dependent applications (including medical packaging and food and beverage containers) where consistent detection is essential. Having this flexibility in one integrated vision device expands inspection flexibility and performance for diverse use cases, including cap alignment, label verification and surface analysis, Lunardhi said.
The In-Sight 3800 Vision System is intended for various tasks, such as detecting defects, monitoring production lines, guiding assembly robots and tracking, sorting and identifying parts.
Engineered for Success: Alan Lunardhi, Technical Sales Engineer, Cognex
Step inside the real stories of engineers as they share candid snapshots of their career journeys, challenges and triumphs. This dynamic series builds on Machine Design’s commitment to spotlight the issues shaping today’s engineering workplace—covering everything from leadership strategies and management insights to in-depth profiles of innovators across every engineering discipline. We also highlight voices from equity-seeking groups, ensuring a diverse and inspiring perspective on what it means to thrive in engineering today.
Visit Machine Design’s WISE (Workers in Science and Engineering) hub to explore more topics, expert advice and meet inspiring people in this series.
One to Watch: Alan Lunardhi, Technical Sales Engineer, Cognex
Job Title: Technical Sales Engineer
Company: Cognex
In His Own Words:
My name is Alan Lunardhi and I work for Cognex as a Territory Sales Engineer for the Southern California area.
Currently, as a sales engineer, I manage the entire sales process, from front to back end, from cold calls to developing solutions and overseeing implementation.
I studied bioengineering in college. But my interest in business led me to the intersection of engineering and commerce.
Sales engineering fits perfectly at the crossroads, allowing me to work across industries, such as medical device, aerospace, food and beverage and consumer goods. The manufacturing space is really broad and my expertise in bioengineering has been valuable in supporting customers in my territory.
Looking ahead, there are several avenues I can take beyond sales engineering, whether that means building on more advanced vision systems applications or getting into technical marketing. For me, the intersection of engineering and business is where the most future opportunities lie.