Oberflächencharakterisierung auf Basis optische Messtechnik
Final Report Abstract
6.6 Conclusion Optical measurement techniques allow for a fast and nondestructive evaluation of roughness. To adapt optical sensors to the challenge of surface characterization of micro structured cylinder liners new measurement and data processing methods were developed and implemented. The complex geometry of the microstructures to be characterized necessitates new approaches for slruclure identificafion and form removal. The optical sensors used for measuremeni of the micro structured surfaces were a confocal chromatic sensor, a point sensor using chromatic aberration for distance measurement and a while light interferometer, which uses incoherent light and image slacks for areal height measurements. Optical measurements inside of Intact cylinder liners are complicated due to the small size of the cylinder opening compared lo the size of the sensor. To overcome this problem a sensor setup was constructed using a coordinate measuring machine (CMM) and a specially tailored sensor head using a chromatic point sensor and an optical fiber. The CMM Is used to posifion the sensor head inside the cylinder liner at Ideal measuring distance and orientafion. The point sensor measures the sensor to object distance while the CMM rotates the cylinder liner. The synchronous positioning information by the CMM and the chromatic sensor are fused to give 3D coordinates of the cylinder liner surface. The main merit of this technique is that the cylinder form is removed by the measurement setup. To analyze the form of micro dimples high resolution measurements using a white lighl interferometer were made, which, due to a small field of view with respect to the dimples, needed to be aligned among each other to form a larger measurement data set. The proper alignmenl was achieved by identifying the relative shift between two images by evaluafing the area cross-section with respect lo the lateral position. Using Surface characterization based on optical metrology 121 a least square fit the area cross-sections of each single measurements data sets where aligned. Based on this, the single measurement data sets were fused to form a large scale, high resolution Image of the micro dimples. These measuremenis allow for an analysis of smaller features of the dimples then the measurements using the chromatic sensor. The measurements using a white light interferometer were carried out on samples with underlying cylinder form, which had to be removed for proper evaluation of the dimples. Because simple form removal by polynomial fit to the surface gives flawed results due to the influences of the micro structures a new method had to be designed which removes the form regardless of the structured surface parts. In a first step the form was removed by a polynomial fit after which the structures where segmented and replaced with other height values which better fit the plane of reference. Now a second polynomial fit is used on the remaining data. Both surface fits are then removed from the original measurement and the relevant surface micro structure becomes accessible for evaluation. Investigation of a large number of structured surface samples made automated segmentation algorithms necessary. Microstructures in the measurement data were identified by segmentation using watershed and threshold methods. The bounding box surrounding the structures was extended and used to search for ridges at the edges of the structures by threshold methods. To find correlafions between the form ofthe micro structures on the surface samples and their respective tribological behavior the identified structures were characterized by parameters describing their individual forni as well as more statistical parameters like areal and volumetric parameters. Micro dimples were also compared to the blade used In the cutting process by fitting an idealized blade to the structure or to actual measurement data ofthe used blades. To analyse microstructures for possible undercuts, which would increase the value of a number of areal and volumetric parameters used for surface evaiualion, methods were developed which allow for undercut detection. The sensor head of a chromatic sensor was modified to allow for an adaptation of the surface to sensor angle, which in theory allows for the measurement of undercuts. The undercuts can then be identified by evaluafion of large measurement data sets obtained by 3D data fusion of several smaller data sets. Another method was developed for the recognition of undercuts using a white light interferometer, by automatically measuring overlapping high resolution fields of view and merging them to form a larger field of view. This is repeated for different surface to sensor angles, the resulting measurements are then aligned and transformed into 3D binary data and merged. With the aim of identifying undercuts automatically a new method was developed to skeletonize 3D binary data. This method Is based on the thinning of 2D binary data which removes morphologically irrelevant pixels. This was generalized to a 3D binary voxel space. This algorithm has great potential for the analysis of 3D surface data for 122 Surface characterization based on optical metrology hills and valleys as well as undercuts, which can be identified by the endpoints of the 3D surface skeleton.
