Abstract

In this paper, an improved adaptive watershed segmentation method is proposed based on the characteristics of the optical fiber faceplate Firstly, median filtering and morphological contrast enhancement are performed on the defect images, and then the gradient of the image is obtained by multi-scale morphology.In the improved watershed algorithm, the local minimum is first removed which the depth is lower than H.Then, the local minimum of the depth larger than H as the seed point are extended.Finally,the gradient image is modified by the forced minimum method .The modified gradient image is used to make the watershed Segmentation to get the final segmentation result .The experimental results show that the method can effectively suppress the over-segmentation, and the defects can be extracted well.

© 2017 OSA, SPIE, ICO, IEEE