Patch and curvature specific estimation of efficient sampling scheme for complex surface inspection

Osama Abdulhameed,Syed Hammad Mian,Abdulrahman Al-Ahmari,Hisham Alkhalefah

DOI: https://doi.org/10.1007/s00170-020-06063-6

IF: 3.563

2020-09-22

The International Journal of Advanced Manufacturing Technology

Abstract:The parts with freeform surfaces are extensively employed in the production industries. Consequently, the quality inspection of sculptured surfaces becomes increasingly important. Typically, the Coordinate Measuring Machine (CMM) is utilized for inspection of these irregular geometries because of its precision. Certainly, the complex and non-rotational geometry makes it difficult to obtain the freeform surfaces. The meticulous analysis will indeed require the explicit reconstruction of the measuring surface. In view of the fact that only a well-chosen sample of inspection points and their locations can accurately define the measurand, the development of an appropriate sampling strategy becomes crucial. This work proposes a novel adaptive sampling plan for the precise extraction of surface form. It recursively and adaptively computes the relevant sampling scheme to achieve an effective inspection using the CMM. The input to the algorithm is a random sample size depending on the inspection point spacing and measuring surface dimensions. This algorithm employs a knot vector–based segmentation, followed by curvature-based grading of the segmented patches. A newly established formulation is utilized to assign a specific sample size for each segment. The points within each patch are allocated by exploiting distinct point distribution algorithms. This algorithm intends to abbreviate the variance between the reference and generated surface obtained through the specific sampling plan. Indeed, the befitting sampling policy provides the least error between the two surfaces. The outcome in the present study demonstrates that the proposed algorithm can significantly reduce the inspection points as well as maintain the precision of surface modeling.