Introduction
Search and rescues are not a new concept. When a person is reported missing, search teams analyze the area and conduct the search. There are many ways to going about a search; to name the methods: containment, hasty, grid, choke point, and track trap. In all of the methods, with the exception of hasty, they are all time consuming. The time-consuming methods all take great deals of planning to coordinate the teams and to also organize volunteers into the teams and as support staff. By the time search teams are deployed, the missing person can be up to miles away from the area that was planned to be searched.
In Lab 12, planning software for search and rescues was used. Loc8 is the name of the software. It is currently in a beta stage, but the main purpose is to use a predetermined RGB value and to find that color within a mass amount of pictures. For example, if a victim that was reported missing was last seen wearing blue jeans, Loc8 can use an RGB value of the pants and search multiple photos that match the same value.
Methods
For the lab, the main purpose was to become familiar with the software Loc8. The students were instructed to find a pair of jeans and a black shirt during the lab. To locate these two clothing items, they were instructed to use images of the two to create an RGB range to find the two items. Once the range was created, the software had to be prompted to search through over a thousand images to find matches of the RGB value.
Results
During the lab, for the pants, a more lighter spectrum was used. Since there were numerous shadows present in all the images, it made the search a lot easier. Figure 1 shows the location of the pants. |
| Figure 1: Within the red circle, the pants were located. |
For the software, it was quite easier for the pants to be found because they are a lighter blue. For the black shirt, it was a little more challenging because it was a darker color, and all of the shadows had a good amount of the same RGB value within them. Figure 2 will show the location of the black shirt.
 |
| Figure 2: This the location of the black shirt: the side of the road. |
Discussion
To find these clothing items, the RGB range will be shown. Figure 3 shows the values of the scale used.
 |
| Figure 3: Above is pictured the spectrum that was used to find the pants. |
To better show the information used for the spectrum, an excerpt from a text file of the settings will be shown below. The spectrum value is to show that the spectrum can go as much as ten above or below the RGB value.
Start Color:
Red: 30
Green: 101
Blue: 133
End Color:
Red: 50
Green: 121
Blue: 153
Spectrum Value: 10
For the shirt, Figure 4 will show the spectrum used.
 |
| Figure 4: This is the spectrum used to find the shirt. |
As before, an excerpt from the text file for the shirt will be shown below as well.
Start Color:
Red: 49
Green: 76
Blue: 103
End Color:
Red: 69
Green: 96
Blue: 123
Spectrum Value: 10
The same spectrum value was used because, although many shadows matched the value, it gave a good range if the shirt would have been lighter or dark when "out in the wild."
Conclusion
During the lab, there were moments when it was challenging. When a darker start RGB value was chosen, almost every image was marked as a potential target because of the shadows. When trying to rerun both of the trials, a lighter RGB value was chosen so that not every image is chosen as a target. Although Loc8 is in its beta stage, the software was very useful for search and rescue. It has great potential to be implemented into the search and rescue industry to reduce plan time before actual search teams are deployed. UAS has a future in search and rescues because they can be used to acquire images and potentially be used for continuous surveillance to find the victim.
Comments
Post a Comment