Week 3 Update and Research


Field Work Friday September 6th, 2019

This was our first day of field work and collecting data to use in Loc8, 3 flights were flown over the Purdue Wildlife Area. The flights were designed to portray a search and rescue mission. Prior to the flight we have taken “last seen” pictures of a team member to act as a missing person. The clothes that were worn in those pictures were then placed randomly in the field. Each flight was flown with similar parameters besides overlap. We are now testing to see how overlap will affect our data and if Loc8 will have any problems identifying the objects. 
Figure 1: Experiment Details

This experiment used a Mavic 2 Pro and the clothes set out in the field were all of different colors to test how darker or brighter colors will be identified. Total flight times were recorded to test which method is the fastest and compare a normal search and rescue mission with a search and rescue mission using UAV’s. Times were also noted for when the operator was able to visually spot each target during the flight.  

Figure 2: Mission 1 Field Notes

Figure 3: Mission 2 Field Notes

Figure 4: Mission 3 Field Notes

Figure 5: Mission 1 Flight Plan

Figure 6: Mission 2 Flight Plan

Figure 7: Mission 3 Flight Plan

We used DJI Ground Station Pro for mission planning, Ground Station Pro allowed us to go down to only 10% overlap in the flight, with the other flights being at 70% and 30% overlap. Due to the length of Mission 3, the flight was stopped with two lines remaining in order to safely land the UAV.

Research

In addition to our first set of flights, a great deal of research was performed in order to get a better understanding of the search and rescue field we are entering. We learned a lot about various methods of SAR missions, as well as many of the limitations various operations have encountered. We came up with a number of ideas that relate to each of the articles we read about how to implement potentially better solutions to the problems that the SAR community is facing. An annotated bibliography about our research has been created, totaling our current 20 sources. That annotated bibliography can be found here.

Comments

Popular posts from this blog

GIS Day Poster

Processing Data in Pix4D, No GCP's