<p>Archaeological line drawings are essential and widely used in archaeological research to depict artefact morphology and key features. Traditional manual drafting is time-consuming and labour-intensive, while generating line drawings from images using automated tools remains challenging due to intricate ornamentation, material wear, and varying illumination. To solve these problems, we propose a generative framework that transforms artefact images into accurate line drawings using limited reference images, reducing data requirements while enabling personalized styles that reflect individual expert preferences. Based on a DiT architecture and fine-tuned with LoRA, our method enables joint visual-textual conditional generation, preserving structural fidelity while meeting professional drawing standards. Comprehensive experiments demonstrate that our method achieves high fidelity and clarity in generating line drawings of bronze, stone, and bone artefacts, closely matching expert hand drawings and outperforming existing sketch-extraction and image-to-sketch techniques. The method offers an efficient and scalable solution for cultural heritage conservation and research workflows.</p>