Extracting surfaces from 3D point clouds is significant in reconstructing and transforming these discrete points into their corresponding models. Scanned point clouds are often accompanied by noise, and the existing methods mainly rely on local feature similarities for surface extractions. Errors in estimating the feature information may lead to incorrect surface detection. In this paper we propose a surface extraction and boundary detection method based on clustering technique. The method can be described in three steps: In the first step, a normal correction is carried out using the information from the neighborhood of points with sharp features. The second step is to cluster the points that meet the coplanar condition of the local quadric surface (LQS). In the third step, surface merger is performed by merging the local surfaces satisfying the merging conditions. Experimental validation is carried out to determine the effectiveness of the proposed method. The experimental results show improved surface extraction accuracy of the proposed method in comparison to RANSAC, RG, LCCP, C2NO and HT methods. •A normal estimation method based on neighborhood information reconstruction is proposed to provide reliable feature information for subsequent segmentation.•A subdivision strategy is designed to extend the DBSCAN clustering algorithm by incorporating a density-reachable condition. This condition is based on the proximity of points to the quadric fitting surface, addressing the issue of over-segmentation.•Based on the characteristic information of normal vector and normal distance (ND), the merging criterion of adjacent quadratic patches is established.