There is a simple idea that is currently lacking across most LAS and point cloud editing software, in both GIS and 3D specialization platforms (even with converting LAS to another format and back again), i.e., the ability to paint (colorize) custom permanent RGB values into LAS point clouds without the need for aerial or oblique imagery alone. The "RGB LAS Paint Brush" 🖌 would be implemented by having a 3D spherical paint brush of any specified diameter that could perform direct color replacements to a point cloud, when viewed in a Scene.
The tool could also modify, rather than replace, RGB values with shading, darkening, contrasting, tinting, etc. There would be much value behind the RGB LAS Paint Brush in that it could be used to correct errors caused by aerial imagery (such as shadows or lidar and imagery not being of the same date and being mismatched for some areas). It would allow users to edit RGB points instead of imagery first. It could help with creating more realistic base coloration of pavilion buildings or removing shadows from the top of structures where they do not make sense.
The RGB LAS Paint Brush could be used for even terrestrial lidar data colorized by ground photos. There is also potential for the tool to create more abstract and visually appealing RGB point clouds for informational, educational, or aesthetic reasons, but mainly the idea is for color correction or replacement of point color values. The tool could paint visible-only points (surface), internal points (spherical reach), selected-only points (isolate points for color edits), or adjusting the color characteristics of the entire LAS file (e.g., brightness/contrast, hue, etc.). The RGB LAS Paint Brush could repurpose point clouds used for analysis and create better and more realistic visualizations without needing to go into advanced 3D specializations outside the scope of most GIScientists. Also, this tool would have nothing to do with classification, and would not require using classes to custom colorize a cloud using symbology. Instead, RGB values would remain in their own symbolization category. It would work in real-time with a Save Edits function, and not be a process to run like Colorize LAS.
Attached are examples of how orthorectified aerial imagery can falsely colorize the top of rooves in an aerial lidar dataset, given the rooves overlap with a shaded or ground region. The process can also double-stamp building imagery colors above and below their rooftops if ground data exists, which the ground could be shaded out manually with such a tool. Being able to color correct the building top would be very useful and not require a more painstaking approach of editing aerial imagery first to see how it turns out later. It would give users more leeway in painting RGB values from a ground perspective (without ground photos) and not just the top-down colorization that is currently available.
Hope others have imagined something similar, would find the tool useful, and thanks for reading!
