基于小波技术的散乱点云自适应压缩算法

Adaptive Reduction Algorithm of Scattered Point Clouds Based on Wavelet Technology

摘要:

提出基于小波技术的散乱点云自适应压缩算法.利用快速成型理论中的切片技术，将三维空间点云数据降为二维平面点云数据，并对排序后的点云数据进行小波变换，利用小波系数峰值，自适应地保留能够反映目标特征和细节信息的点，实现散乱点云的快速压缩.借助于实验，验证切片的分割厚度选为采样间隔的2~3倍时，可以实现快速高质量的散乱点云压缩.结果表明：算法在特征保留上具有明显的优势，能够最大限度地保留特征信息，压缩效果更为理想，且无需设置阈值，同时还具有自适应的特点，有助于实现压缩的自动化.

Abstract:

An adaptive reduction algorithm of scattered point clouds based on wavelet is proposed, in which the 3D point clouds are converted into point sets on the 2D plane firstly by using the slicing technology in rapid prototyping theory, and then the wavelet coefficients of sorted point clouds data after the wavelet transform can be obtained whose peaks represent the points to be reserved. According to the experiments, the rapid and high quality reduction of scattered point can be performed while the slice thickness is chosen as 2 or 3 times of the sampling interval. The result indicates that this algorithm has obvious advantages in terms of the feature preserving. It can preserve the feature information ultimately, thus the reduction results are more ideal. Due to peaks of wavelet coefficient can adaptively identify the objects’ details and features, this algorithm needlessly set a threshold, which explains the adaptability of the algorithm and also contributes to realizing the automatic reduction.