To address missing features and reflections in the extraction process of ancient ceramic patterns, a pattern extraction method combining sharpening-smoothing and the whale-type k-means algorithm is proposed. By analyzing the reflection phenomenon of ancient ceramic images, a sharpening-smoothing image enhancement method is designed. It effectively improves image detail and texture expression. In addition, by analyzing the characteristics of missing graphic features in ancient ceramics, a whale-type k-means algorithm is constructed to achieve accurate extraction. The experimental results show that the accuracy of this method in ancient ceramic pattern extraction reaches 99.319%. F1 Score, MIoU, and Recall are 93.13%, 93.84%, and 87.15%, respectively. This method demonstrates superior performance in the accuracy and robustness of ancient ceramic pattern extraction. Meanwhile, it provides reliable technical support for the digital protection of cultural heritage and academic research.