A Saliency-Based Sparse Representation Method for Point Cloud Simplification

Abstract

High-resolution 3D scanning devices produce high-density point clouds, which require a large capacity of storage and time-consuming processing algorithms. In order to reduce both needs, it is common to apply surface simplification algorithms as a preprocessing stage. The goal of point cloud simplification algorithms is to reduce the volume of data while preserving the most relevant features of the original point cloud. In this paper, we present a new point cloud feature-preserving simplification algorithm. We use a global approach to detect saliencies on a given point cloud. Our method estimates a feature vector for each point in the cloud. The components of the feature vector are the normal vector coordinates, the point coordinates, and the surface curvature at each point. Feature vectors are used as basis signals to carry out a dictionary learning process, producing a trained dictionary. We perform the corresponding sparse coding process to produce a sparse matrix. To detect the saliencies, the proposed method uses two measures, the first of which takes into account the quantity of nonzero elements in each column vector of the sparse matrix and the second the reconstruction error of each signal. These measures are then combined to produce the final saliency value for each point in the cloud. Next, we proceed with the simplification of the point cloud, guided by the detected saliency and using the saliency values of each point as a dynamic clusterization radius. We validate the proposed method by comparing it with a set of state-of-the-art methods, demonstrating the effectiveness of the simplification method.

Fast subsampling strategy for point cloud based on novel octree coding

“…For structured point clouds, Markovic et al [36] introduced a feature-sensitive subsampling technique based on insensitive support vector regression, which can efficiently locate and save points in high-curvature areas while minimizing points in flat areas. Leal et al [37] constructed eigenvectors by calculating the normal vector and curvature of each point to estimate the saliency of the points and used the saliency values as a guide for point cloud simplification. Along with standard features such as normal and curvature, it also contains point cloud subsampling that uses characteristics such as edge form and density.…”

Point Density Variations in Airborne Lidar Point Clouds

“…Even though high-precision DTMs generally do not need the full density of the current point cloud [ 24 , 40 , 76 , 77 , 78 , 79 ], redundant points do not negatively influence the resulting DTM. However, redundant points can significantly increase the processing time or storage requirements [ 80 , 81 , 82 ], causing practical issues for many applications.…”

“…Decimation techniques range from simple ones, such as random sampling, to complex decimations based on shape of the objects described by the point cloud [ 82 , 89 , 90 , 91 , 92 , 93 , 94 , 95 , 96 ]. Random sampling can be based on the ordinal number of a point within the point cloud (count-based decimation).…”

A Scraper Conveyor Coal Flow Monitoring Method Based on Speckle Structured Light Data

“…This kind of method can directly simplify the point cloud data as a whole, and the efficiency is relatively high. The second category is to divide the original point cloud data into different grids [26,27]. By simplifying the point cloud data in each grid, such as the uniform grid method, this method requires grid construction before data simplification.…”