Brainstorm Technology LLC
3D Photography Software Solutions
Tutorial
PhotoSketch is a photo-centric urban 3D modeling plugin for SketchUp. Three tutorial videos are presented below to demonstrate the versatilty of PhotoSketch. A full description of this tutorial is available in the user manual.

NEW PROJECT / CAMERA POSE RECOVERY / FLOOR ALIGNMENT
In the first video shown below, a tutorial is presented describing how to set up a new project in PhotoSketch. We show how PhotoSketch proceeds to extract and match geometric features derived from a collection of overlapping photos of an urban scene. This is used to perform camera pose recovery in which the positions and orientations of the cameras that acquired the photos is computed. The purpose of this stage is to infer where the photographer was standing in the scene so that the resulting image can be properly projected onto the created 3D model. Modeling is enhanced by using these projected images as tracing paper upon which the user can sketch the footprint of the model. In this manner, we use these images as visual cues for standard SketchUp extrusion and taper operations.





2D FOOTPRINT SKETCHING / EXTRUSION

In the next video, a tutorial is presented describing how to use the SketchUp interface along with an
extended set of PhotoSketch plugin tools to build 3D models. The PhotoSketch workflow requires the user to draw 2D rectangular footprints directly on the calibrated photo. These footprints are then extruded to embed 3D volumes directly into the 2D scene photos. The photos thereby serve as visual cues as well as textures that are projected onto the extruded geometry to produce photorealistic 3D urban models.




TAPER / REFINEMENT / ADVANCED TEXTURE MAPPING
The next video describes how to use the SketchUp interface along with an
extended set of PhotoSketch plugin tools to build 3D models. This video segment focuses on the use of extrusion, taper, and advanced texture mapping tools.




MANUAL FEATURE TRACKING / CAMERA POSE RECOVERY
In the event that automatic feature tracking does not recover the camera poses, the user has the option to perform manual feature tracking. The next video describes this process.



FLOOR ALIGNMENT USING VERTICAL-LINE CORRESPONDENCE
This tutorial shows how to perform floor alignment by drawing line segments on the vertical edges of façades or windows among two photos in the path. In many scene photos, it is often easier to find vertical line segments than finding corresponding points on a horizontal plane. The goal of this tool is to draw vertical line segments on a photo and match these segments across photos in the sequence. The user is free to start from any photo in the sequence. Technically, the user only needs to draw one set of line correspondences to align the floor. However, drawing more than one pair of line correspondences is recommended to distribute the floor alignment error among all photos in the scene.