Data fusion of areal measured microtopographies with plane of reference using optical sensors (DFG)
The Projekt „Data fusion of areal measured microtopographies with plane of reference using optical sensors” of Leibniz Universität Hannover (LUH) is funded by the German Research Founda-tion (DFG) and has the aim of impriving methods to increase the field of view of image processing microscopes by stitching. Stitching is a technique that uses the fusion of partially overlapping meas-urements with high resolution to form a high resolution, large field of view measurement. Due to a number of deviations like distortion, missing measurement data, axial positioning errors and rotation between the axes of the CCD-chip and the translational axis the results of image registration (alignment) and fusion are often unsatisfactory. Commercially available systems often lack sufficient correction of these deviations. In the project we researched the various deviations of 3D-microscopes with positioning stages on their influence to the creation of mosaics of measurements. We developed a software module that allows the characterization of radial distortion by measuring chessboard-like surface standards. Us-ing resolution, structure and flatness standards we measured the various rotational and translational positioning errors as well as the rotation of the CCD-chip against the positioning axes. The measured errors were used in a model based simulation of 3D-microscopes to characterize the influence of the positioning errors on the image registration.Measurements by 3D-microscopes, such as white light interferometers or confocal microscopes often exhibit false or missing data due to high surface gradients or bad lighting. These measurement data have to interpolate consistently in different measurements to allow an accurate image registration. We researched the resulting similarity (measured by cross-correlation) of surfaces with varying degrees of statistical and deterministic bad measurement values and outliers after interpolation with a number of pixel and triangulation based interpolation methods. To allow for a robust method of image registration of overlapping measurements with unknown overlap we developed a method of image registration based on the calculation of one dimensional feature functions. These functions are calculated along the direction perpendicular to the direction of movement. By estimating the overlap using the simultaneous correlation of the feature functions and subsequently registering the measurements by correlation by Fourier transform we were able to decrease the necessary overlap from 35 to 20 % down to 10 % of the measurement area. This was tested using a model based simulation of a white light interferometer and simulated polished surfaces. A similar test was conducted for the characterization of registration by phase cor-relation which was also verified by a test conducted using a high precision piezo positioning stage.
Contact person: Dr.-Ing. Dipl.-Phys. Markus Kästner
Multiscale Geometry Measurement (SFB871-A2)
The assessment off worn parts the key to successful regeneration. Defects occur due to different kinds of thermal, chemical and mechanical stress. Tactile measurement with single point-wise or line-wise reading are not sufficient to detect defects than can occur randomly distributed over the whole surface. Optical sensors with an areal measurement can digitalize the surface with a good ratio of duration to resolution.
Geometric defects vary in size and can simply be a global deformation of the whole work piece or a defect in micro structured surface. It is not possible to digitalize a work piece in a reasonable time with a single micro scale sensor. The approach to this problem is to use different sensors with overlapping scales, to minimize the measurement duration.
The core of the system is the multi sensor head with a camera for photogrammetry and a structured light sensor. For positioning the sensor head is mounted to an industrial six axis robot arm. For an exact reference the position and orientation of the sensor head is measured with a Lasertracker. Algorithms for the detection of defects in image and ranged data are developed. For the assessment of the shape the difference of the measured data to CAD data is calculated. Because of the huge amount of measured data methods for storing, processing and visualization of this data are in the focus of research. In cooperation with the other projects of the SFB871 parameters for product regeneration are developed.
Contact person: M.Sc. Yinan Li
Fiber-endoscopic 3-D sensors for industrial inspection (SFB/TR 73, TP B6)
Fringe projection offers a great variety of applications fields in geometry of free form elements. From large measuring areas that can be found in automotive car body parts down to geometry elements with sizes in the millimeter range such as gearings, it can be used for areal, holistic measurements. With advanced deviation analyses methods errors in fabrication lines can be found immediately after their appearance and rejections can be minimized. That makes it to an excellent tool the quality control in the production metrology.
However, there are still fields that cannot be covered by fringe projection. One of these fields is filigree form elements on narrow or internal carrier geometries, which cannot be captured by current optical measuring devices due to shadowing.
To overcome this drawback, a fiber-endoscopic micro fringe projector was developed. The new device is capable of resolutions of less than 20 µm with uncertainties of about 5 µm in a workspace of 7x7x5 mm³ and is used to measure form elements of Sheet-Bulk Metal Forming (SBMF) tools. 3-D measurements of typical geometries of SBMF-tools can be captured in less than 1 s.
Contact person: Dipl.-Ing. Steffen Matthias
An analysis of the potential and qualification of the process control system of a local distributed force-sensing technology (AiF)
The research project “An analysis of the potential and qualification of the process control system of a local distributed force-sensing technology", funded by the German Federation of Industrial Research Associations (AiF) is jointly developed by Institute of Measurement and Automatic Control (IMR) and Institute of Metal Forming and Metal Forming Machines (IFUM).
The purpose of this project is to enforce the overall monitoring and detection of the productive process with a newly developed local distributed force transducer and multichannel measurement System. With the measured data being analyzed with machine learning algorithms, the mould can be fine-tuned so that the optimized clearance can be achieved between punch and die. Thereby, a high reject rate can be avoided.
For the mass production of components of limited size with high forming degrees, multistage press systems are frequently used. The allowed manufacturing tolerances for every press stage are very low because the workpiece is thin. The measurement of the process force in the individual process stages of a punching process is so indispensable (for example, for the purpose of process monitoring), that a system for the production of force transducers with absolute calibration capability need to be developed. The basis principal is segmentation of the integrated spring body of the force transducer, which makes it possible to locally resolve force measurements and the transmitted form punch can be measured simultaneously.
By using such sensors for process monitoring, consequently, the information contained in the distribution of the force would not be lost. Conversely, the information can be used for an improved control of the process by the detection of process disturbances, which could offer a better protection for the machine with faster shutdown at faults and help to find the faulty set up of a mould by measuring an unexpected distribution of the force. By using the information provided by the locally resolved force measurement, process monitoring, which is being devolped in this for project.
Contact person: M.Sc. Yinan Li
Multiscale Methods for the Dimensional Characterisation of Surfaces and Interfaces (MARIO)
In automotive applications new approaches concerning innovative plasma spray surface coatings for liquid-lubricated high performance contact pairs, like cylinder liners in combustion engines, deal with porous material layers to optimise the tribological behaviour and lubrication in view of fuel efficiency.
The aim of this project is the research and development of methods for the dimensional characterisation of these porous surface coatings using multiscale methods for data acquisition and analysis. For a reliable measurement of the form, surface topography and volume information of the porous material layers, it is necessary to develop appropriate calibration and measurement strategies based on a combination of surface metrology and xray micro-tomography. Based on the generation of a volumetric data model, methods for a quantitative analysis of the dimensional parameters of the geometric features on or within the porous surface coatings should be developed. The correlation of these parameters with the tribological behaviour of the surface gives the key information for an optimisation of the plasma spraying process.
Contact person: M.Sc. Yibo Zou
Functionally integrated porous implant coatings (Ulderup Foundation)
The implantation of artificial joints, endoprostheses, is one of the most frequently performed operations in Germany and is by now regarded as a routine procedure. A critical factor of endoprostheses is the quality of the sliding surface. As wear particles can cause foreign body reactions, which in turn can lead to a loosening and thus the failure of the endoprosthesis, the endurance of the implants depends largely on the wear resistance and lubrication of the sliding surface. With currently used material combinations a life span of about 15 years can be achieved. However, this means that especially young patients are often affected by costly and risky revision operations.
The aim of this research project is the development of functionally integrated, porous implant coatings, which possess a high wear resistance due to the coating material itself and further ensure an active friction reduction due to a targeted and controlled release of a biocompatible lubrication fluid. In addition, a "local-drug-eluting" system to release anti-inflammatory drugs is supposed to be integrated.
At the IMR both the manufacturing concomitant and the final characterization of the implant surfaces is done by the use of imaging techniques.
Contact person: M.Sc. Nina Loftfield