Non-Optimal Illumination Conditions
In circumstances where the lighting conditions are less than ideal, or where the settings of the camera capture were not optimal for the level of light in the data, we suggest implementing pre-processing on the photos prior to import into PhotoMesh for final processing. Pre-processing is especially vital when the raw data is in full bit format (12-bit or 16-bit) and the end-result photos are in 8-bit format, (TIFF or minimally compressed JPG).
Unprocessed photos typically exhibit a more contrasted histogram, which means areas with large brightness variations are emphasized. This can be observed in instances such as bright regions adjacent to shadows, or dark roads situated alongside light sidewalks.
While the photo may appear aesthetically pleasing at first glance, it may not be optimal.
The deficiencies in the photos become more pronounced when inspecting the photo for pixel-based correlation. This process necessitates maximum detail at full native resolution, a task which is particularly, challenging in areas of limited or excessive illumination.
Even areas with large contrast may have shadows that are excessively dark, or illuminated surfaces that are too bright. The ability to correlate 3D features at the pixel level in these areas will be greatly limited.
Applying corrections on the histogram edges, specifically through shadow and highlight adjustments, followed by a minor adjustment in saturation to accentuate non-gray features, can help achieve a more uniform distribution of actual pixel values of captured objects. This results in a "flatter" photo in terms of histogram, reducing the impact of shadows or overexposure.
Below are some recommended settings for photo correction, which include a white balance adjustment to achieve a more neutral gray color scheme:
These adjustments improve the 3D correlation in the treated areas and facilitate better texture blending in scenes composed of both shaded and illuminated photos of the same object.
These areas now contain more detail for 3D correlation. Even though one was previously in dark shadow and the other in bright illumination, the scenes are now much closer in histogram level. This leads to improved geometry and a more neutral texture that can be adjusted with environmental settings to recreate various shading and lighting conditions.
Additional tweaks applied to this photo include dehaze and minor color balance adjustments in the shadows and highlights to account for Rayleigh Scattering.
The degree of dehaze correction hinges on the visibility conditions during data collection. The color correction in the shadows and highlights depends on the light intensity and contrast between illuminated and shaded surfaces.