[Notebooks PR] Update reconstructing_3D_buildings_from_aerial_lidar_with_deep_learning.ipynb

[neha-sharma-geoai]

✅ Enhancements Made to the Notebook & Workflow

The notebook and supporting workflow have been upgraded for clarity, reproducibility, and ease of use.
Below is a detailed breakdown of all improvements, organized by category.

• Notebook Language and Documentation Improvements
  • Revised the language throughout the notebook for consistency, readability, and conceptual understanding.
  • Enhanced introduction to the notebook.
  • Expanded explanations across the workflow.
  • Added/edited markdown cells to separate conceptual sections, output analysis, and procedural steps.
  • Reorganized/added a few cells for better flow.
• Inserted comments within code blocks to explain:
  • Purpose of each cell
  • Syntaxes
  • Parameter roles
  • Dependencies and assumptions
  • Expected outputs
• All 8 input and intermediary datasets used in the notebook were identified, collected, and published to the api_data_owner ArcGIS Online account. Item IDs in the notebook were updated to reference the newly published items. Added cells for fetching the items in the notebook.
• Added thumbnails and descriptions for each item on the ArcGIS Online portal.
• Modified script to automatically replace hardcoded file paths containing employee username with a generic placeholder <USER_HOME>. Ensures the notebook can be safely shared publicly or across teams without exposing personal directory structures.
• The trained model from extended epochs was missing. Re-trained the Mask R-CNN instance segmentation model at high epochs. Exported and published this trained model as a new item on ArcGIS Online. Added a cell to load the model directly from ArcGIS Online
• Explained the custom toolbox methodology and added steps to add and run a custom toolbox with an embedded Python Script in ArcGIS Pro
• Ensured that it is a run-all notebook.

Datasets -

• Guneet's project folder - \ndhfs2\Swap\Guneet Mutreja\3D_building_reconstruction
• All the required datasets for this notebook are maintained here - \ndhfs2\Data\Jupyter_Sample_Notebooks_Data\Reconstructing_3D_buildings_from_Aerial_LiDAR_with_Deep_Learning
• On the api_data_owner portal, all assets are published and placed here - https://geosaurus.maps.arcgis.com/home/content.html?sortField=modified&sortOrder=desc&view=table&folder=cb6315c4282246e59ab5895ac327c0f8#my

Please let me know if any other change is required.
Renaming of the notebook is required, which will be done in a separate PR.
Currently, the custom toolbox is not generating the expected 3D results; an issue has been created for it.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB