Abstract
Accurate and effective classification of lidar point clouds with discriminative features expression is a challenging task for scene understanding. In order to improve the accuracy and the robustness of point cloud classification based on single point features, we propose a novel point set multi-level aggregation features extraction and fusion method based on multi-scale max pooling and latent Dirichlet allocation (LDA). To this end, in the hierarchical point set feature extraction, point sets of different levels and sizes are first adaptively generated through multi-level clustering. Then, more effective sparse representation is implemented by locality-constrained linear coding (LLC) based on single point features, which contributes to the extraction of discriminative individual point set features. Next, the local point set features are extracted by combining the max pooling method and the multi-scale pyramid structure constructed by the point’s coordinates within each point set. The global and the local features of the point sets are effectively expressed by the fusion of multi-scale max pooling features and global features constructed by the point set LLC-LDA model. The point clouds are classified by using the point set multi-level aggregation features. Our experiments on two scenes of airborne laser scanning (ALS) point clouds—a mobile laser scanning (MLS) scene point cloud and a terrestrial laser scanning (TLS) scene point cloud—demonstrate the effectiveness of the proposed point set multi-level aggregation features for point cloud classification, and the proposed method outperforms other related and compared algorithms.
Highlights
Lidar sensors have been widely used in many fields
locality-constrained linear coding (LLC), Point set features fusion of LLC-latent Dirichlet allocation (LDA) and LLC-MP LLC, Point set features of LLC-LDA LLC, Point set features of LLC-MP DKSVD, Dictionary-based sparse representation LCKSVD1, Sparse representation based on saliency dictionary LCKSVD2, Sparse representation based on saliency dictionary
This paper presents a novel point set features extraction method via multi-level global and local features aggregation for point cloud classification
Summary
Lidar sensors have been widely used in many fields. Classification of laser scanning point clouds is an important technology in the applications of automatic driving, intelligent city, mapping, and remote sensing [1,2,3,4]. The single point-based methods mainly consist of neighborhood selection, feature extraction, and classifier for each single point classification [5,6,7,8,9]. As the growth criterion construction and the low-level features selection have huge impact on the point clouds segmentation, the region growing-based algorithms usually are less robust. It is difficult to segment the point cloud objects of different scales based on a clustering algorithm. The proposed method first uses the DBSCAN (density-based spatial clustering of applications with noise) [27] algorithm to coarsely segment the point cloud. (3) A multi-level LLC-LDA and LLC-MP aggregation feature extraction and fusion method of the point set is proposed. Once the local LLC-MP and the global LLC-LDA aggregation features are generated, we fused them together to obtain the final discriminative point set features
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have