Subscribe Button 1

Photogrammetry Increasingly Used for Automated Production Measurements

Over the past two decades, the industrial design and manufacturing technology infrastructure has changed significantly. 2D component drawings have been replaced by complex 3D CAD models. Tooling, molds, and production equipment are today designed with the aid of 3D simulations, and produced using associated CAM systems. Design and free-form surfaces are becoming more complex with significantly expanded design capabilities resulting in improved product designs, ergonomics, and space optimization.

As a consequence of these developments production quality checks must be adapted accordingly. Today’s measurements must be three dimensional, fast and increasingly non-contact. The goal being a total component comparison of the actual produced 3D surfaces using nominal CAD surface models as the reference. Increasingly, single point measurements, using traditional Coordinate Measuring Machines (CMM) are insufficient as the trending demand is for components to be completely digitized and not just a few critical dimensions measured. Requirements for highly precise measuring processes remain unchanged.

Measuring systems have to be flexible and in some cases portable to allow scanning of objects in-situ or on-site. The spectrum of parts to be measured extends from small injection-molded parts to the large metal forming tools and complete vehicle sheet-metal bodies..

One measuring concept currently experiencing significant market success is the free-form acquisition of surface geometry based on the photogrammetric principle. The photogrammetry system key is a freely positional, self-contained 3D fringe projection sensor, which identifies its orientation in a higher-level coordinate system exclusively through markers fixed on the component or associated tooling (reference points). The use of markers eliminates the need for highly precise mechanics to orientate and position the sensor around the part under measurement.

Basics of photogrammetry

In photogrammetry, the position of a point in 3D space can be determined by triangulating multiple bundles of observation rays. With the knowledge of the spatial orientation of each bundle in the object coordinate system, the intersection of the rays delivers the desired 3D object coordinate. Bundle adjustment is used to determine the unknown object coordinates and additionally the parameters of the detecting arrangement. For this, multiple observations from different directions are required, with a partially overlapping image area. Furthermore it is necessary that all desired object points exist in more than one observation. A minimum of two observations of one object point is required to create enough equations for the derivation of its position.

Fringe projection procedures are characterized by the projection of line-like light of differing intensities on the object. Observations of the fringe under a known triangulation angle indicate the 3D contour along the fringe. For a comprehensive measurement, either the fringe has to be moved or multiple parallel fringes have to be projected. By using extensive fringe patterns projected in sequence, high-resolution encoding of the object surface can be achieved. In practice, the phase shift procedure is used whereby the fringe patterns are shifted one-quarter of their width resulting in every camera pixel registering a sequence of four gray scale values which can be assigned to the fringe phase shift.

Additional, absolute information is required to achieve unique encoding of the fringe number. The component is illuminated with a sequence of fringe blocks of differing widths and every camera pixel registers a sequence of digital information that stands for a specific fringe number. The combination of procedures results in an absolute encoding of the object surface that can be acquired independently for each camera pixel.

GOM ATOS 3D Digitizers are based on the combination of fringe projection and stereo camera configuration built into a compact and robust sensor head. The sensor projects different fringe patterns onto the part surface which is acquired by two cameras. ATOS uses the phase shift procedure to generate extensive three-dimensional measurement data. Up to 16 million independent measuring points are captured within 1 to 2 seconds with the measuring data characterized by a very high detailed reproduction enabling even very small component features to be measured.

Parts to be measured are often larger than the measurement field of the Digitizer. In addition, a single measurement tends to be insufficient to obtain a complete point cloud for complex components. For these reasons, partial measurements always have to be combined.

To accommodate large parts and need for overlapping measurements GOM utilizes circular markers affixed on or around the component. These small circular markers are uniquely identified by the stereo sensor as real 3D points, and therefore take on the function of fixed features.

With the growing demand for automated metrology, and the comfort zone manufacturing companies have with industrial robots due to their cost and reliability, GOM has adapted it ATOS 3D Digitizers and use of markers to its ScanBox Measuring systems that provide shop-floor production measurement cells.

The reference markers are affixed directly to the part fixture or holding device allowing the robot metrology to be eliminated from the overall system measuring accuracy since the robot derived sensor position is determined at every static sensor pose. The GOM ScanBox Optical CMM is available in different sizes with the largest ScanBox 8 comprising dual robots mounted on linear rails and capable of measuring a complete automotive body-in-white structure. Optical 3D coordinate measuring systems providing complete 3D part deviations against CAD data.

The GOM Virtual Measuring Room (VMR) software supplied with ScanBox offers a comprehensive tool for automated 3D part digitalization. VMR is central control for all elements of the ATOS measuring cell and offers functional representation of the real measurement environment in a virtual simulation. Users can work with the system without need of specific robot programming skills. All robot movements are simulated in VMR and checked for safety before being executed. The VMR software handles all measuring procedures:

The optical production CMM has become a reality and providing front-line production measurements in automotive, aerospace and appliance manufacturing industries.

For more information: