Quality evaluation of honing surface groove features based on improved Level-Set analyses

Honing groove features of a cylinder liner surface contain helpful information on surface performance. Traditional honing surface evaluation methods do not focus on groove features, which is significant to surface functions. Hence, a topographical feature segmentation algorithm and corresponding evaluation parameters are needed. This paper proposes an improved level-set-based segmentation algorithm for groove feature extraction. The proposed algorithm was tested on a real honing surface dataset compared to some state-of-art algorithms, including the classical height thresholding and watershed segmentation. The results show that the proposed method is robust against initial conditions and improves both segmentation accuracy and computational speed. Specifically, a 30% accuracy improvement over the best-tested level-set algorithm has been observed. Another case study on fold-mental characterization was conducted, which also validated the accuracy and speed performance of the proposed algorithm.