Rate Control Optimization for Joint Geometry and Attribute Coding of LiDAR Point Clouds
Yang Wang, Wei Gao, Xingming Mu, Hang Yuan
Abstract
With the increasing utilization of dynamically acquired Light Detection and Ranging (LiDAR) point cloud data, the significance of LiDAR point cloud compression has grown substantially. Bit rate control plays a pivotal role in LiDAR point cloud compression. This study introduces a novel bit rate control approach for LiDAR point cloud compression. By constructing rate and distortion models based on the geometry and the attribute of the LiDAR point cloud, as well as developing a comprehensive distortion model encompassing both geometry and attribute, this approach transforms the bit allocation problem for geometry and attribute into a constrained optimization problem. Consequently, it determines the optimal bit allocation for geometry and attribute. Furthermore, a rate control algorithm, incorporating parameter updates, is proposed for precise bit control. Experimental results demonstrate that our method achieves a remarkable overall BD-Rate performance, with an average bit error rate of merely 0.93%.