The In-Sight SnAPP vision sensor on display at the Cognex booth at ATX West 2024 was designed to streamline manufacturing processes.
The latest addition to Cognex’s vision sensing portfolio, the technology is an integrated smart sensor that streamlines quality control on production lines.
According to Kyle Maledy, applications engineer, Cognex Corporation, the vision sensor can run without external software requirements and provides “quick sanity checks” to ensure products meet quality standards. “This is an integrated smart sensor bridging the gap between a traditional photo eye and a full-blown vision system or vision camera,” Maledy said.
Aside from typical expected benefits aside—the ability to save time and reduce costs—Maledy highlighted that the sensor addresses a range of quality control challenge, including presence/absence inspections, assembly verification and defect detection.
READ MORE: Manage, Monitor and Synchronize the Entire Hardware Build
He pointed out that vision-based detection is an improvement on conventional laser-based sensors because it can locate parts in any position and can detect subtle defects.
Plug & Play Software for Industrial Machine Vision
In-Sight SnAPP is the first Cognex product that does not require software for installation.
Cognex’s press release stated that the system is trained using just a few examples and users will not need any programming or special vision knowledge. A web-based user interface allows customers to run the application from anywhere using a standard web browser.
Cognex reported that this product alone expands the company’s capabilities into the vision sensing space and adds a whopping $1 billion to its served markets.
Vision Applications with AI-based Automation
During the booth demo, Maledy also demonstrated the Insight D900 deep learning-enabled vision system, designed to solve a range of tasks, including defect detection, assembly verification and optical character recognition. It comes with field-changeable lighting, lenses, filters and customizable covers to match application requirements.
The camera, which launched in 2020, was set up to use the communications for Ethernet, and demonstrated how the system could perform multiple inspections at once.
READ MORE: How Voxel Combines Video Feeds and AI to Assess and Mitigate Workplace Safety
“Here we’re looking at PCBs for defects as small as bad solder joints, and as large as the OCR,” said Maledy. “We’re also reading the texts on the cards for serialization and we’re reading the data matrix code to make sure that data matches and we can output all those results as good or bad and associate those with the unique part ID for track and trace functionality downstream.”
Video footage associated with this article was filmed at ATX West 2024 in Anaheim, Calif.