My request is for an application that would create 3d models either directly on a phone or help gather sufficient data to generate the model on a PC. An example of the end functionality that I want can be seen at: insight3d(dot)sourceforge(dot)net
Using several images, a point cloud is generated that can be turned into a polygonal 3d model. I am not concerned about textures. The end output file should be .stl (Standard Tessellation Language). I understand the processing requirements for such an application would be very high, however, I believe the latest generation of mid to high-end phones are nearing the power requirements. If not, thats where the helper app would come in.
Desired features:
-Capture photos of macro (building size) or micro (soda can and smaller) objects
-Scale reference in view finder
-Object orientation reference (see Camera3D in the Android market)
-Previous image overlay in view finder (Camera3D)
-Auto create a folder that contains the photos generated for one object. The goal for this is to simplify transfer to a PC
-Render models, that canbe saved in an stl format, from photos (see insigh3d on sourceforge and netfabb for PC and Android)
-Warning to indicate when there is insufficient data or blindspots for the model
-Object detection (I imagine this might simplify the generation of the model, or provide better accuracy. In essence, the app user would draw a box around the object they wish to model in each photo they take.)
-Ability to scale models post render
-It may be of benefit to gather accelerometer and gyroscope data (where available) to give relative location data for each photo. If there is enough resolution from GPS (specific location) and/or the above sensors (relative location) then a significant amount of math for determining camera perspective could be dropped.
As the 3D printer community continues to grow, an app (or apps) with a feature set similar to the above could stand to be extremely profitable.
EDIT:
insight3d is open source.
Trimensional 3D scanner for the iphone has essentially the correct functionality
EDIT 2:
Also see Objectify in the Android Marker
Using several images, a point cloud is generated that can be turned into a polygonal 3d model. I am not concerned about textures. The end output file should be .stl (Standard Tessellation Language). I understand the processing requirements for such an application would be very high, however, I believe the latest generation of mid to high-end phones are nearing the power requirements. If not, thats where the helper app would come in.
Desired features:
-Capture photos of macro (building size) or micro (soda can and smaller) objects
-Scale reference in view finder
-Object orientation reference (see Camera3D in the Android market)
-Previous image overlay in view finder (Camera3D)
-Auto create a folder that contains the photos generated for one object. The goal for this is to simplify transfer to a PC
-Render models, that canbe saved in an stl format, from photos (see insigh3d on sourceforge and netfabb for PC and Android)
-Warning to indicate when there is insufficient data or blindspots for the model
-Object detection (I imagine this might simplify the generation of the model, or provide better accuracy. In essence, the app user would draw a box around the object they wish to model in each photo they take.)
-Ability to scale models post render
-It may be of benefit to gather accelerometer and gyroscope data (where available) to give relative location data for each photo. If there is enough resolution from GPS (specific location) and/or the above sensors (relative location) then a significant amount of math for determining camera perspective could be dropped.
As the 3D printer community continues to grow, an app (or apps) with a feature set similar to the above could stand to be extremely profitable.
EDIT:
insight3d is open source.
Trimensional 3D scanner for the iphone has essentially the correct functionality
EDIT 2:
Also see Objectify in the Android Marker