Agisoft Metashape Professional 2 'link' -

The quality of your final piece depends entirely on your input.

Since you are using the Professional version, you have access to advanced tools that distinguish it from the Standard edition: agisoft metashape professional 2

Automatic (AI) Background Masking: Modern machine learning tools now assist in automatically removing backgrounds, a huge time-saver for studio-based photogrammetry or small-object scanning. The quality of your final piece depends entirely

While both editions share core photogrammetry functions like mesh generation, the Professional edition is required for advanced outputs and industrial applications. Professional Georeferencing (GCPs, RTK/PPK) DEM & Orthomosaic Outputs Multispectral & Thermal Processing Python Scripting API Network & Cloud Processing LiDAR Data Support Comparison and Value for Existing Users Agisoft Metashape Change Log LiDAR & Laser Scan Integration: The most significant

Conclusion

• Surveying and mapping professionals
• Architecture, engineering, and construction (AEC) industry professionals
• Researchers and scientists working with 3D data
• Photogrammetry enthusiasts and hobbyists

LiDAR & Laser Scan Integration: The most significant addition is native support for terrestrial and aerial LiDAR. Users can now co-process images with laser scans to improve the accuracy of 3D models and digital elevation models (DEMs).

Quick tips

• Use consistent camera settings and avoid motion blur.
• Include scale bars or GCPs for survey-grade accuracy.
• Process subsets when testing parameters to save time.
• Keep backups of projects and use chunking for very large datasets.

Recommended settings (general)

• Image alignment: High accuracy for camera alignment when possible.
• Dense cloud: Medium to High quality; use Aggressive filtering only for noisy data.
• Mesh: Use Poisson for watertight models; Surface-from-depth for thin structures.
• Texture: Use mosaic blending for orthorectified textures; keep texture size to balance detail and file size.
• GPU: Enable CUDA/OpenCL and use tiled processing for very large datasets.