Balloon Mapping 1

Introduction: 

 Earlier this year we did some preparation work in order to launch a weather balloon to obtain aerial photography of UW-Eau Claire campus. Thanks to this work done a head of time we were prepared to launch our rig attached to the balloon with no problem. It was decided not to use the soda bottle rigs instead we would use the high altitude balloon launch (HABL) rig as a dry run for the real HABL launch later this semester.

Methods:

In order to have a successful launch the class split up into several different groups to complete these tasks. In order to get the rig into the air we had to get the camera rig ready, get the helium tank down to the shed, get the balloon filled, and measure out 400 ft of rope so we would know how far we know the balloon was up in the air. We also had a person whose sole job was to take pictures and anther to take video.

Below Figure 1 shows the helium the tank down to shed so we could fill the balloon up. Down at the shed the used some plastic tubing to fill the balloon up with helium. They used zip ties to make sure seal the balloon close. Figure 2 shows the filling of the balloon. Figure 3 shows that the balloon is almost filled. Someone had to hold on to the balloon while it was being filled. Figure 4 shows the zip ties being used to seal the balloon.

Figure1: The transportation of the helium tank

Figure 2: Filling up the balloon

Figure 3: Balloon almost filled

Figure 4: Zip ties used to close of the balloon

With the balloon being filled, we needed to get the string measured out so we would know how high the balloon was in the air. To do this we unwound a real of string. At every 50ft interval we marked it with a permanent marker.

Figure 5: We used Red to mark 50ft and Black to mark 100ft

Figure 6: We used the tiles on the floor, which were one foot long, as a measuring unit to measure out the string

 Earlier this semester we had built two different rigs to hold our camera. However it was decided to use the HABL rig for this launch. By doing this Joe and the rest of the balloon rig engineers would get an idea on how it would react in the real world. Figure 7 shows the two rigs we built earlier this year.  Figure 8 shows the HABL rig that we used for this launch.

Figure 7

Figure 7

Figure 8: the HABL rig. We actually had two rigs built just in case

With all the jobs done we were ready to launch the balloon. We choose to launch the balloon in the center of the green area. UWEC had undergone a large face lift this past year and half. We lost our old Davis center and gained a new one. We are also in the process of building a new education building. With the imagery we would collect with the launch we could get a new aerial view of our new campus. Figure 9 shows us launching the balloon. Figure 10 shows us watching for the 400ft mark that we had mark on the string. 

Figure 9: The launch of the balloon rig

Figure 10: Counting out the feet

After we got the balloon up in the air we walked around the green for awhile. We then took the balloon down so we could switch out the cameras. The first one was a regular camera set on continuous mode. This time we would be using a flipcam that would take video instead of pictures. Figure 11 shows our the flipcam that we used.

Figure 10: The flipcam we used in the second rig

 This time we decided to take the balloon across the foot bridge over the river. We manged to get across the bridge with no problem. The real issue came when we began to reel in the line. The string snapped and the rig fell into the river. The video below shows the string snapping.

Luckily Joe was able to retrieve the rig from the river. The camera was fine and so was the tracking beacon we placed inside. Figure 11 shows Joe climbing back up to the sidewalk after getting the rig out of the water.

Figure 11: Joe climbed down to the waters edge and used a branch to snag the rig out of the water.

Georeferencing and Mosaic:

Now that we had our images we could georeference them. To georeference an imagine means we take a regular picture and establishes it's location on the Earth. I used ArcMap to georeference the images. I used a reference picture to help me georeference the images. This reference image is already georeferenced and is close to the same resolution of our images, making it ideal for us. When georeferencing you want to get your REMs error as low as possible. Anything around 2 or 1.5 is very good. I tried to get each of my image's REMS error to be around 2 as possible. After the images were georeferenced it was time to mosaic them together. To mosaic an image means to take several smaller images and stitch them together into a larger image. I could have used ArcMap to mosaic the images but instead I used Erdas IMAGINE. It took me about 2-3 hours to get this final image. Figure 12 is my final mosaic over laying the reference image.

Figure 12: Final image

Discussion:

We ran into several problems during this launch and during the post processing. The first problem that we ran into was that it had been very windy that day. Because of the wind our rig was thrown all around and did not take many perpendicular pictures. We need perpendicular images to have a successful mosaic. Figure 13 is an example of an image that is not perpendicular to the ground. While these are interesting to look at they are not useful when mosaicing.

Figure 11: A picture of campus and the River

Another problem that we ran into was the lack of ground points or ground references. Because of all the recent construction we did not have an updated image of what the campus green looks like right now. It is difficult to georeference an image without a good ground references points. It is possible by georeferencing one image very well. Using this image you can then reference the rest of your images by making sure they over lap by at least 60%.

Results:

 In conclusion there are many steps to in making a successful launch for taking aerial photos. It does help to have a lot of people so that these tasks can be split up into more manageable tasks. It is also important to do as much of the prep work before the actual launch date to save yourself headaches and problems. For the actual launch, you want to do it on a day that is clear and has little to no wind. It can take a long time to georeference your images into a good final image. You need a lot of patience and willingness to sometimes through out images to get a perfect final mosaic.