Breakthrough Technology for Machine Vision Applications

Zebra Aurora™ Deep Learning offers breakthrough technology for machine vision applications. Designed as an add-on product, it further enhances the already impressive outputs of Aurora™ Vision Studio and Aurora™ Vision Library software with a set of industrial-quality deep learning-based tools, including optical character recognition (OCR) and features & anomaly detection.

In Aurora™ Deep Learning, advanced deep learning capabilities recognize the way imaging is used in more unusual vision applications – and evolve accordingly.

Seven ready-to-use tools, which are trained with 20-50 sample images, can quickly be used to detect objects, defects or features automatically. Internally, this software uses large neural networks designed and optimized by our research team for use in industrial vision systems.

With use cases across multiple industry verticals, this software enables solutions for complex machine vision problems that were previously impossible to achieve with traditional algorithms and approaches. Now, for example, machines and manual operators are able to use it to find complex and irregular surface defects or accurately read blurred, poorly lit, or damaged characters printed on glossy surfaces of product packaging.

View the Spec Sheet to explore the full capabilities of Aurora Design Assistant 10.

Find & Capture 3D Box Dimensions 

If your challenge is capturing accurate parcel dimensions as parcels move in and through your shipping workflows, your solution is Zebra Dimensioning™ - Certified Mobile Parcel and Mobile Parcel. Both are industry firsts — dimensioning solutions that are fully integrated into handheld mobile computers, enabling the anywhere and anytime capture of accurate parcel dimensions.

3D Surface Matcher 

Aurora Design Assistant includes the ModelFinder3D step for finding a surface model– including multiple occurrences–at wide-ranging orientations in a point cloud. A surface model is defined from a point cloud obtained from a 3D camera or sensor, or from a CAD (PLY or STL) file. Various controls are provided to influence the search accuracy, robustness, and speed. Search results include the number of occurrences found and for each occurrence, the scores, center coordinates, estimated pose, and number of points. 

3D Box Finder 

Aurora Design Assistant features the Box Finder step for locating one or more box-like objects in a point cloud. The boxes to find are specified graphically or numerically. The step returns the position and orientation information of a whole box and its faces while also providing dimensions. Box Finder handles situations where a restricted number of faces are visible.

The object detection approach—implemented by the CNNObjectDetect step—locates instances of pre-established classes. It is suited for inspection tasks whose goal is to find, size and count objects or features. The result for each located instance is the most likely class, the score and a bounding box including the corner coordinates, center, height and width. These classification steps are particularly well suited for analyzing images of highly textured, naturally varying, and acceptably deformed goods in complex and varying scenes.

1D and 2D Measurement Steps 

Aurora Design Assistant includes three steps for measuring: Measurement, BeadInspection, and Metrology. These tools are predominantly used to assess manufacturing quality. 

The Measurement step uses the projection of image intensity to very quickly locate and measure straight edges, stripes, or circles within a carefully defined rectangular region. The tool can make several 1D measurements on edges, stripes, and circles as well as between edges, stripes, and circles. 

The BeadInspection step is for inspecting material that is applied as a continuous sinuous bead, such as adhesives and sealants, or its retaining channel. The step identifies discrepancies in length, placement, and width, as well as discontinuities. The BeadInspection step works by accepting a user-defined coarse path as a list of points on a reference bead and then automatically and optimally placing search boxes to form a template. The size and spacing of these search boxes can be modified to change the sampling resolution. The allowable bead width, offset, gap, and overall acceptance measure can be adjusted to meet specific inspection criteria. 

The Metrology step is intended for 2D geometric dimensioning and tolerancing applications. The step extracts edges within defined regions to best fit geometric features. It also supports the construction of geometric features derived from measured ones or defined mathematically. Geometric features include arcs, circles, points, and segments. The step validates tolerances based on the dimensions, positions, and shapes of geometric features. The Metrology step can also analyze profiles obtained from 3D profile sensors or extracted from point clouds and depth maps. 

The Metrology step’s effectiveness is maintained when subject to uneven changes in scene illumination, which relaxes lighting requirements. The expected measured and constructed geometric features, along with the tolerances, are kept together in a template which is easily repositioned using the results of other locating steps.

1D and 2D Code Reading and Verification Steps 

The CodeReader step in Aurora Design Assistant locates and reads 1D, 2D, and composite identification marks. The step handles rotated, scaled, and degraded codes in tough lighting conditions. The step can provide the orientation, position, and size of a code. The alternate CodeReaderIK and DataMatroxAVL steps provide the ability to read multiple codes at once among other traits. In addition, the CodeGrade step verifies the quality of a code based on the ANSI/AIM and ISO/IEC grading standards.

Security Features 

Access to specific Operator Views and the Aurora Design Assistant management portal can be made to require user authentication (i.e., username and password) so only authorized personnel can modify key settings in a deployed project. The access control is implemented using the Keycloak identity and access management service with support for integration with the Microsoft Active Directory service. A project can be locked to a specific Zebra Iris GTX smart camera or vision controller when deployed, preventing it from running on an unauthorized platform. A project can also be encrypted during deployment to a platform, ensuring that the project cannot be read or changed by unauthorized users. Projects locked to a platform are automatically encrypted.

Unlock the full potential of your machine vision applications with Aurora Design Assistant 10! This intuitive, versatile, and extendable integrated development environment (IDE) empowers you to build robust vision solutions without writing a single line of code. With deep learning and 3D tools at your fingertips, you can tackle complex visual inspections, classification, and segmentation tasks with ease.

Key Benefits:

• Flowchart-Based Programming: Simplify your development process by constructing flowcharts instead of traditional coding.
• Hardware Independence: Choose the best platform for your project, whether it’s a Zebra Iris GTX smart camera, vision controller, or a third-party PC.
• Advanced 3D Capabilities: Interface with Zebra AltiZ and other third-party 3D sensors for comprehensive depth map and point cloud analysis.
• Deep Learning Integration: Leverage powerful CNN-based tools for accurate image classification, detection, and segmentation.
• Customizable Operator Interface: Design a web-based operator interface to control and monitor your applications from anywhere.
• Real-Time Feedback: Boost productivity with instant feedback on image analysis and processing operations.

Don’t miss out on the opportunity to revolutionize your machine vision applications. Call Automation Distribution at 1-888-600-3080 or contact us here and get Aurora Design Assistant 10 and experience the future of vision technology today!