Agisoft Metashape Photoscan - Orthophot from Mesh or Dem

 Usually we recommend to generate DEM from the dense cloud and use this surface as input for the orthomosaic generation, thus the meshing step can be skipped.

For very large areas most likely it wouldn't be possible to generate a single-block mesh, therefore the DEM is recommended for large areas covered.

Using Mesh surface may be reasonable for custom plane projections (like building facades, for example) or for small areas, when mesh surface produce more accurate results, for example, by the edges of the roofs.

Agisoft Metashape Photoscan - DEM from Mesh or Dense Cloud

 When you choose to build a DEM based on the Dense Cloud, Photoscan will generate a DEM on the specified resolution, build from the the dense points directly with an "Inverse Distance Weighting" (IDW) interpolation.

The DEM resolution that is specified by Photoscan is directly related to the Quality mode that was chosen when building the Dense Cloud. Where a higher quality level means a higher density 'Dense Point Cloud' and thus a higher resolution that's available when building a DEM (without upsampling the data).

When you choose Mesh, the points get interpolated as a "Triangulated Irregular Network" (TIN) where it linearly interpolates the Dense Points as small triangles. The higher the specified resolution is, the smaller the triangles are.

 

 

odm docker

 docker run -it -v /home/thiago/datasets/project/code/images:/code/images -v /home/thiago/datasets/project/code/gcp_list.txt:/code/gcp_list.txt -v /home/thiago/datasets/project/code/orthophoto/:/code/orthophoto --rm  opendronemap/odm  --dsm --orthophoto-resolution 2 --gcp gcp_list.txt  --crop 0 --fast-orthophoto --resize-to 1024 --project-path /

docker run -it -v /home/thiago/datasets/project/code/images/:/code/images -v /home/thiago/datasets/project/code/gcp_list.txt:/code/gcp_list.txt -v /home/thiago/datasets/project/code/odm_orthophoto/:/code/odm_orthophoto -v /home/thiago/datasets/project/code/images_resize/:/code/images_resizes -v /home/thiago/datasets/project/code/odm_georeferencing/:/code/odm_georeferencing -v /home/thiago/datasets/project/code/odm_meshing/:/code/odm_meshing -v /home/thiago/datasets/project/code/opensfm/:/code/opensfm --rm  opendronemap/odm  --dsm --orthophoto-resolution 2 --gcp gcp_list.txt  --resize-to 2048 --project-path / --time --mesh-size 550000 --mesh-octree-depth 10 --texturing-nadir-weight 30

tirar o fast orthophoto melhor a qualidade da imagem, deixa menos blur

sem o resize ele trava, falta memoria ram para 850 fotos