Abstract:Presently, the topics of sensor network information sharing and multi-robot mapping in robotics have gained considerable research interest. This paper introduces a real-time map compression methodology developed to address the time-intensive problem associated with large-scale map cross-system transmission. This method employs neighborhood filtering, a modified Moore-Neighborhood tracking algorithm, and polygonal Boolean operation technology to extract the inner and outer boundaries of a 2D raster map, thereby enabling map reconstruction. The proposed methodology specifically focuses on the removal of burrs from the outer boundary and reoptimization of the inner boundary search process. This is done to enhance the efficacy of the map reconstruction. Additionally, the extracted inner and outer boundaries are utilized for the reconstruction of the map. The performance of the proposed algorithm was evaluated using several types of maps. The results indicate a relatively low execution time of 300 ms for a map comprising 2 million pixels, which essentially fulfills the real-time requirements for large-scale map transmission and reconstruction. Furthermore, the quality of the map reconstruction was assessed using the Universal Image Quality Index (UIQI), Multi-Scale Structural Similarity (MS-SSIM), and Peak signal-to-noise ratio (PSNR) image evaluation metrics. The reconstruction results reveal that the map effectively satisfies the requirements with a UIQI exceeding 0.9984, an MS-SSIM surpassing 0.9282, and a PSNR above 27.7089. Simultaneously, the map’s compression rate exceeded 99.0%, thereby demonstrating that the algorithm can accomplish a significantly high compression rate. This evidence suggests the algorithm’s potential for effective deployment in real-world applications requiring large-scale map transmission and reconstruction.

Real-time acquisition of compact volumetric maps

Efficient screen space approach for high complex dynamic scenes in real-time rendering

VDBblox: Accurate and Efficient Distance Fields for Path Planning and Mesh Reconstruction

3D Map Building for Mobile Robot under Unknown Outdoor Environments

Partition Based Real-Time Rendering Method for Large-Area Terrain Data

Mesh-LOAM: Real-time Mesh-Based LiDAR Odometry and Mapping

Adaptive Decomposition and Shared Weight Volumetric Transformer Blocks for Efficient Patch-Free 3D Medical Image Segmentation.

3D LiDAR Map Compression for Efficient Localization on Resource Constrained Vehicles

An Efficient Volumetric Mesh Representation for Real-time Scene Reconstruction using Spatial Hashing

Coxgraph: Multi-Robot Collaborative, Globally Consistent, Online Dense Reconstruction System

Multi-View Depth Map Sampling for 3D Reconstruction of Natural Scene

Efficient volumetric mapping of multi-scale environments using wavelet-based compression

A Smart Voxelization Algorithm Zhao Dong Wei Chen Hujun Bao Hongxin Zhang Qunsheng Peng State Key

Huge texture mapping for real-time visualization of large-scale terrain

Real-Time Globally Consistent Dense 3D Reconstruction with Online Texturing.

Adaptive Block Elevation Mapping for Large-scale Scene.

DP-MVS: Detail Preserving Multi-View Surface Reconstruction of Large-Scale Scenes

Real-Time Map Compression Method Based on Boolean Operation and Moore-Neighborhood Search.

Real-time 3D reconstruction at scale using voxel hashing

Real-time Underwater 3D Reconstruction Based on Monocular Image

3DFusion, A real-time 3D object reconstruction pipeline based on streamed instance segmented data