SynPhoRest - A Procedural Generation Tool of Synthetic Photorealistic Forest Datasets

The development cadence of Deep Learning algorithms is extremely high, with the development of autonomous navigation systems and especially autonomous vehicles in the forefront. Growing environmental awareness has directed efforts toward the development of autonomous systems for the maintenance and preservation of forested areas. Unlike urban areas, available datasets on these environments are scarce and incomplete. In addition, the complex and unstructured nature of forested areas and the tedious labeling process lead to a high rate of mislabeling. Given the success that the use of synthetic data has had in model training, this work proposes an approach that helps overcome these limitations. The SynPhoRest simulator can generate photorealistic synthetic data in the form of RGB images, semantic segmentation maps, and depth maps from procedurally generated virtual forest environments. The system supports both manual and automatic extraction and can generate 100 sets of data in about an hour.