SatelliteSurfaceReconstructionEvaluation

Evaluation scripts of 3D Surface Reconstruction From Multi-Date Satellite Images

Installation

• Requirements
  • Install and configure SatelliteSurfaceReconstruction
    • Including the python libraries defined in SatelliteSurfaceReconstruction/requirements.txt
    • Including Meshlab
  • Install Pyntcloud by running pip install -r requirements.txt
• git clone https://github.com/SBCV/SatelliteSurfaceReconstructionEvaluation.git

Configure the SatelliteSurfaceReconstructionEvaluation Project

• Add SatelliteSurfaceReconstruction to the SatelliteSurfaceReconstructionEvaluation project
  • For example in Pycharm
    • Open SatelliteSurfaceReconstructionEvaluation
    • Open SatelliteSurfaceReconstruction and attach it to SatelliteSurfaceReconstructionEvaluation
    • Go to File / Settings / Project: SatelliteSurfaceReconstructionEvaluation / Project Dependencies
      • Select SatelliteSurfaceReconstructionEvaluation
      • Make sure that SatelliteSurfaceReconstructionEvaluation is selected

Prepare the Ground Truth Data

Obtain the data from here.

Run the Evaluation Pipeline

• Run ssr_eval/run_ssr_evaluation.py
• The first time ssr_eval/run_ssr_evaluation.py is executed, it will create a configuration at ssr_eval/configs/eval.cfg using the template located at ssr_eval/configs/eval_template.cfg
• Adjust the paths in the config file ssr_eval/configs/eval.cfg such as
  • meshlab_server_fp
  • meshlab_temp_dp
• Select the required parameters such as
  • sampling_method
  • algo_target_dn_list
  • vissat_dataset_site_idp
  • ...
• Run ssr_eval/run_ssr_evaluation.py