All the robots in the world won’t help your assembly process, if your parts are not in the right place or the right orientation for pick up. Vision systems can help. They can recognize different parts, determine how they are positioned, and tell a robot exactly how and where to pick them up and drop them off. The newest technologies take advantage of artificial intelligence to overcome inconsistent lighting or changes in optical scale. Here’s a look at four of the latest technologies.

3D Vision Camera for Bin Picking

With its compact, industrial-grade design, short working distance and large field of view, the Ensenso B 3D vision camera is particularly suitable for bin picking applications. Thanks to its high depth of field, it can capture objects over a large area and to localize all parts in a container at once. This optimizes the pick-and-place process and facilitates emptying of containers.

One strength of the camera is its ability to precisely locate objects at close range. It provides a wide field of view and a high depth of field. This means that the area in which an object is sharply imaged is unusually large. At a distance of 30 centimeters between the camera and the object, the Z accuracy is approximately 0.1 millimeter. The maximum working distance is 2 meters.

The housing, which accommodates two 5-megapixel color cameras, a pattern projector and all the electronics, is completely enclosed. This simplifies the cabling. The unit complies with protection class IP65/67. Measuring just 120 by 56 by 104 millimeters, the camera can be mounted to a robot arm.

IDS Imaging Development Systems
en.ids-imaging.com

Smart Camera Has AI Tools

The VS series of smart cameras lets engineers adjust focus, field of view, and resolution with a single click. A series of mechanical zoom lenses allows one camera to be used across a range of mounting distances and fields of view without changing lenses. Customized CMOS sensors are available for resolutions of 1.6 to 15 megapixels.

Using optical zoom, the camera can change the field of view without sacrificing resolution. All adjustments can be completed from the software. No mechanical changes are necessary. The camera has a unique auto-focus algorithm to create the best image. This optimizes images for code reading, measurement, and flaw detection applications that require high resolution.

Engineers can apply both AI and rules-based tools on the same image. Edge AI tools can be programmed with almost any number of images. Unlike typical deep learning tools, edge AI tools can be applied quickly and easily.

The camera’s water-resistant housing is rated at IP-67. Power can be supplied to both the camera and light over a single Ethernet cable. A high-intensity smart ring light features snap-on installation and the ability to cut ambient light. The camera comes standard with advanced algorithms that adapt to changes in ambient lighting.

Keyence Corp. of America
www.keyence.com

High-Speed Camera for Airborne Testing Applications

The Phantom Miro C321 Air is a high-speed camera for airborne testing applications. This 2-megapixel camera leverages the rugged and field-proven design of the Phantom Miro C321, which was originally designed for vehicle crash tests. The C321 Air modifies that design for manned and unmanned airborne testing requirements. Specifically, its high frame rates and advanced imaging features expand testing opportunities and meet industry requirements for data security, environmental durability, EMI and avionics power.

The camera achieves 1,480 frames per second at full HD resolution and up to 94,000 fps at reduced resolutions. Because airborne tests often require lower frame rates, the higher rate unlocks additional testing opportunities while providing detailed, higher-quality images. The camera also features a 10-micron pixel size for high sensitivity, low noise and high dynamic range to capture critical test events with minimal motion blur.

The C321 Air satisfies aerospace-specific specifications including one of the most difficult tests: the vibe test. The camera is also temperature tested to -55 C, includes conformal-coated boards to protect against humidity, and offers a removable CFast solid-state drive, secured with a double-action latching door. Users can download shots to the flash in-flight, saving multiple captures from a mission.

Vision Research
www.phantomhighspeed.com

AI Software Improves Automated Visual Inspections

Kitov.ai announces significant improvements to its flagship software platform with the introduction of version 2.7. The latest version brings about enhanced user experience and expanded functionality, offering a more efficient and user-friendly workflow along with increased control options.

The software suite by Kitov.ai features multiple components designed to enable smart inspection processes. Its improved AI Trainer permits engineers without specialized expertise to rapidly gather and label defect data, thus providing a solid foundation for the software’s neural network learning.

Enhancements in version 2.7 are mainly in the Planner Tool, which integrates seamlessly with various robotic systems. It uses AI to create optimized inspection strategies and simplifies the process of transferring these plans to the inspection equipment. Improvements include a more accessible interface, equipped with intuitive icons, tool tips, contextual menus, and versatile features like an undo function for reversing unintended actions.

CAD2SCAN functionality, a vital part of the Smart Planner, benefits from refined workflows and additional improvements. It allows quality inspectors to define test parameters directly on a CAD model of the part being inspected, making it easier and faster to handle components with complex shapes. The tool now supports an expanded range of quality tests that can be applied to a CAD model, enhancing its ability to detect missing components; mark part edges, patterns, and surfaces.

The Workcell Configurator has also been updated to allow for 70 percent quicker configuration and validation of robotic workcell hardware through a straightforward and intuitive user interface.

Kitov.ai
kitov.ai