Evaluating NeRFs for 3D Plant Geometry Reconstruction in Field Conditions
Authors:
Muhammad Arbab Arshad,
Talukder Jubery,
James Afful,
Anushrut Jignasu,
Aditya Balu,
Baskar Ganapathysubramanian,
Soumik Sarkar,
Adarsh Krishnamurthy
Abstract:
We evaluate different Neural Radiance Fields (NeRFs) techniques for reconstructing (3D) plants in varied environments, from indoor settings to outdoor fields. Traditional techniques often struggle to capture the complex details of plants, which is crucial for botanical and agricultural understanding. We evaluate three scenarios with increasing complexity and compare the results with the point clou…
▽ More
We evaluate different Neural Radiance Fields (NeRFs) techniques for reconstructing (3D) plants in varied environments, from indoor settings to outdoor fields. Traditional techniques often struggle to capture the complex details of plants, which is crucial for botanical and agricultural understanding. We evaluate three scenarios with increasing complexity and compare the results with the point cloud obtained using LiDAR as ground truth data. In the most realistic field scenario, the NeRF models achieve a 74.65% F1 score with 30 minutes of training on the GPU, highlighting the efficiency and accuracy of NeRFs in challenging environments. These findings not only demonstrate the potential of NeRF in detailed and realistic 3D plant modeling but also suggest practical approaches for enhancing the speed and efficiency of the 3D reconstruction process.
△ Less
Submitted 15 February, 2024;
originally announced February 2024.
