This project demonstrates a method for visualizing SfM derived pointcloud and image data in a web browser.
Small details, which were not resolved by the SfM pointcloud, can still be visualized using the images.
These data were collected as part of a 2017 Hurricane Irma and Maria Disaster Reconnaissance Survey in the US Virgin Islands by Oregon State University. Funding for data acquisition was provided by the National Science Foundation through awards CMMI-1761461, CMMI-1661315. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation
The main sections of the project are the interactive map in the upper left, the toolbar across the top, and the main 3D scene. The main scene has both the pointcloud and image pyramids.
"Image Pyramids" are placed in the scene to represent where an image was taken from. The image plane is placed at a position and orientation computed through SfM processing, such that it aligns with the pointcloud when viewed from the right position.
The view can be simplified to turn off the map and images using the toggle buttons on the toolbar.
Use the left mouse button to orbit around a point. Use the scroll wheel to zoom. Use the right mouse button to translate the view. Double click a point to fly to it.
"Look Through" mode is a view where the viewpoint is placed at the exact position an image was taken from. The image plane is placed at a position and orientation such that it aligns with the pointcloud. This can be seen as holding a picture frame in front of a person. The two ways to enter this mode are: 1) Click a desired image pyramid. 2) Click the Forward / Backward arrow on the toolbar. This will move the current view to align with either the selected image plane, or the previous/next image.
Use the left mouse button to pan/tilt the camera. Use the scroll wheel to zoom in and out (this is actually changing the field of view). Use the left mouse button to translate the camera position, which will exit the "look-through" view.
The interactive map box (which can be resized by dragging the lower right corner), contains a high resolution orthophoto of the scene overlaid on top of a simple basemap. The orange polygon represents the approximate footprint of the current view, while the cyan polygon represents the footprint of the current "look-through" image. There is a small, draggable dot which reprents the horizontal position of the current view.
The measurements toolbar enables 4 basic measurements on the pointcloud (Point, Distance, Height, and Angle.) These measurements are displayed on the screen until the "Remove All" button is clicked.
The "Look at" button can be used to filter the image planes by images that can see that point. Using the filter button enables and disables filtering by images. Using the "Visible Point" button turns the magenta dot off, but maintains the functionality. As you step through different "look through" views using the arrow keys, the camera will stay locked onto this point, and skip images which don't see the point.
There are various appearance sliders and buttons which you can use to change the quality and color of the pointcloud.
When in "look through" mode, the opacity of the image can be changed with the appearance slider.
The downloads dropdown can be used to download the pointcloud, orthophoto, and metadata. *Note: Rather that taxing the OSU research server with Gb of file transfers, small text files are used to demonstrate the functionality.