Intelligent terrain generation considering global information and terrain patterns

Simulated terrains can provide rich information for landform and terrain research, disaster prediction, rescue and disaster relief, and national security. Quickly generating an accurate simulated terrain for target areas is of great importance. However, for existing data-driven terrain generation methods, it is difficult to balance modeling accuracy and the amount of data required. To overcome this gap, this study proposes a deep learning method that integrates global information and pattern features of the local terrain (IGPN) to realize terrain generation. In our proposed IGPN method, we first apply both terrain patterns (ridge line and drainage line) and elevation points as input data; thus, the terrain features (i.e., channel and peak) and elevation of the target terrain can be provided. The local information extraction module (LIEM) and global information extraction module (GIEM) are then applied to generate local and global terrain features, respectively. Thus, global and local terrain features can be supplemented, and a more accurate terrain can be generated. Experiments show that the IGPN method performs state-of-the-art terrain-generation tasks. Specifically, compared with existing terrain generation methods (IETA, SRResNet, Bicubic, Kriging, FEN, and ERFFN), the terrain generated by IGPN is closer to the real terrain and can retain more local terrain features.

Automated summary

Simulated terrains are important for landform and terrain research, disaster prediction, rescue and disaster relief, and national security. This study proposes a deep learning method, IGPN, that integrates global information and pattern features of the local terrain to generate accurate simulated terrains quickly.