Комаров Артём о новых алгоритмах и технологии интеллектуального роботизированного планирования (eng)
Комаров Артём о развитии интеллектуального роботизированного планирования (eng)

Комаров Артём о развитии интеллектуального роботизированного планирования (eng)

Artem Komarov noted that designing and manufacturing components for the automotive industry is a complex and time-consuming process. Metal stamped components for automobile bodies, chassis, enclosures, and brackets—simple and complex—are designed using CAD/CAM software.

Комаров Артём Андреевич, инжиниринг

Precision is key to producing quality parts that are safety-critical. With high demand for zero-defect automotive parts, stamping manufacturers are under pressure to produce large quantities of complex parts quickly without compromising on quality.

Parts undergo a series of quality verification steps to ensure defect-free stampings, but often those checks rely on human inspectors. Because poor-quality components come with a high cost—with stamping manufacturers sometimes forced to absorb the cost of entire shipments if a defective part is found—stampers are turning to automated vision systems to eliminate human error.

Smart inspection planning allows metal stamping operations to boost quality without slowing down the process. Using deep learning, 3D imaging, novel algorithms, and intelligent robotic planning technologies for single-part inspection, vision systems can ensure that only defect-free parts make it off the line.

CAD2SCAN software is designed to simplify the setup of robotic vision inspection systems using CAD-based automated inspection planning.  CAD2SCAN automated planning and inspection software platform, stampers can automate inspection processes that are difficult and time-consuming to program and deploy manually.

How It Works
Particularly useful for parts with complex geometries, CAD2SCAN allows quality managers to define their inspection requirements directly on the part CAD model. This can save weeks or months compared to manual programming. Users mark the surfaces to be scanned on the part CAD model and CAD2SCAN automatically extracts the specific geometric and semantic information for each inspection requirement.

The requirements are exported to the planning software, which creates an optimized inspection plan that guarantees full coverage. It reduces robot motion and the number of images needed to cover all areas of interest. This minimizes the total inspection time.

The information is passed on to the relevant semantic detectors performing the inspection tasks. In addition to built-in semantic detectors such as a surface detector, label detector, screw detector, and existence detector, its open software platform allows integration of third-party detectors that benefit from the planning and reporting services provided by the platform.

The robotic inspection system performs the inspection according to the plan, scanning each region the user marked on the CAD model. The scan results are displayed in the Review Station tool. Stampers can view defects such as cracks, wrinkles, and incorrect sheet thickness side by side with a corresponding image of a part without defects.

The system provides a complete digital record of every part inspected for future reference and simplified root-cause analysis.

The technology is implemented as a plug-in to common CAD software systems, such as SolidWorks.

CAD2SCAN technology improves the inspection of single-material parts with complex stamped 3D geometric shapes; CNC parts, which are difficult and time-consuming to fully inspect manually; and custom-made parts, which involve processes that are extremely hard and expensive to automate.

The technology is implemented as a plug-in to common CAD software systems (currently available for SolidWorks and Creo). It also supports the evolving QIF (quality information framework) ISO standard and can parse visual inspection requirements embedded into it.

This approach to automated vision inspection is designed to give metal stamping manufacturers confidence in the quality of their automotive components, with data to back it up, said Artem Komarov.