Home
Who We Are
Project Overview
Building Instructions
Hardware Testing & Results
The Flight
Results: Panoramas
Other Results
Download Data
Acknowledgments and Links
In the News
Image Recovery
Image recovery occurred after HASP landed and consisted of the removal of the pictures from the camera, digitally stitching of the pictures, and comparing satellite images with those obtained from GEOCAM.
Comparing Satellite Imagery and GEOCAM Data
When images obtained from Google Maps are compared to those obtained from GEOCAM, the level of detail afforded by GEOCAM is comparable to the next best resolution provided by Google Maps. Image 1 is a picture of Moriarty, NM taken by GEOCAM, while Image 2 is a picture provided from Google Maps of the same area.
 |
 |
Image 1: GEOCAM
Picture |
Image 2: Google
Maps Picture |
One pixel on GEOCAM was estimated to cover approximately 3.3 feet. Potentially, the amount of land GEOCAM could provide coverage over could increase as in this flight the camera was tilted at a 20 degrees angle from nadir.
PhotoStitching of Images
GEOCAM data was used to see which stitching algorithms performed well with data obtained from flight. An algorithm was chosen based on how accurately the data could be stitched together based on the GPS data obtained from the flight log. The photos also needed to be placed together so that visual obstructions, such as clouds, would minimally affect the outcome of the new panorama. A RANSAC algorithm was determined to be a good fit for GEOCAM, and was provided by the AutoStitch website.
First, five images were stitched together to obtain a reduced file of 200KB. Shown below in Image 3 is the outcome panorama of five pictures in a general area.
Image 3: Panorama of 5 pictures
We then went on using about 100-200 frames per panoramas and obtained the following view:
Image 4: Large panorama of 100-200 pictures
One of the main issues when stitching a larger amount of pictures together seemed to be that the overlap comparison between frames was based on cues from clouds. When the clouds changed locations, some of the panoramas became stretched accordingly. This is physically evident in the bottom-right side of the panorama shown in Image 4.
Eventually, the images obtained through panoramas will be
placed on Google Maps with a circle placed on the area of the map to show
the general area that the pictures were taken. Future integration with
GoogleEarth is planned for the future.