Project Update 10/1/19

-
The group has made steady progress since the last update. Training with the m600 has concluded with preflight checks in place and operational procedures established. The data collection procedure is still undergoing improvements before we are confident that the data that we collect will be of high quality.

On 9/30, Tim Pruss and Luke Mlakar went to Purdue Wildlife Area again in order to test different data collection methods with the m600. The flights took place from 5:45 to 6:30. The same XT2 Thermal Sensor was used but the speed of the UAV while collecting data was decreased to and, in later flights, the angle of the sensor was changed from Nadir to an oblique 60 degrees. The second and third flights, performed with a 60 degree oblique angle were unusable.

Problems to work past

Several issues arose from the first data collection outing on September 7th.

A possible solution to this problem is to capture images at an oblique angle that allows a larger area to be captured in a single image. The problem with this solution is that the flight location would have to be shifted over to cover the same area and would have to take pictures from two different angles to avoid unseen areas. Unfortunately, the use of oblique angles requires a different processing input from a standard nadir collection method.

 Below are the thermal orthomosaics of September 7th and September 30th respectively.
Purdue Wildlife Area 9/7/19
Purdue Wildlife Area 9/30/19
The more recent one incorporates our improvements in the data collection process with a higher collection altitude and a slower movement speed. The trade-off of flying at a higher altitude is that there is a worse pixel per square inch ratio in a camera that is already lacking in resolution. However, by flying higher the UAV was able to collect images covering a larger area. We were also able to avoid a large amount of the overexposure that plagued our first data set and made the RGB orthomosaic so patchy.

 These two orthomosaics also display the high range of temperatures that a single area can reach even on two similar days at relatively the same time.

Below you can see the differences in height.
200ft
400ft 
And below here are the generated orthomosaics.
PWA Ortho 9/7
PWA Ortho 9/30


 With our methods slightly more refined from our first flight we are going to attempt to collect a full diurnal analysis data set.

 We plan on having flights at both sunrise and sunset as well as two more in intervals throughout the day. Having same-day collection will reduce the vast amount of variables that may have caused the differences in the two data collection days that we have had so far.

 As of writing this post we plan to do our first diurnal pattern analysis on Friday, October 4th. We may have to reschedule depending on the weather, however.
Along with progressing in our data collection methodology, we have begun designing the layout for our poster for GIS day.


Metadata for the 9/30/19 flight

 General
------------------
Location: PWA
Date: 9/30/19
Vehicle: M600 Pro
Sensor: FLIR XT2
Battery: blue, magenta
Approval # (LAANC/COA/Waiver)

 Flight Information
-------------------
Flight Number: 1
Takeoff Time: 5:48
Landing Time: 5:57
Altitude (m): 121
Sensor Angle: nadir
Overlap: 90
Sidelap: 90

 Geolocating 
------------------
System used (GCP type/PPK)
Coordinate System: WGS 84 UTM Zone 16 North

 Weather
------------------
Metar Used: KLAF 302054Z 20005G15KT 10SM FEW060 31/19 A3000 RMK AO2 SLP155 T03110189 56020

 Crew
------------------
PIC: Luke Mlakar
VO: Tim Pruss

Metadata: Luke Mlakar

Popular posts from this blog

GIS Day Final Poster

-
Image
After presenting the rough draft of our poster, it was critiqued by our classmates. We wrote down each critique that was given to us and deliberated over the changes that we would make to our poster. Below is a list of critiques that we were given as well as our response to those critiques. Our rough draft of the poster looked like this: Figure 1. Poster Rough Draft "Emphasize the data “results”, should be more like 60%-70% of the poster" We agree with this point and will work to emphasize the results section more by decreasing the size of the other sections and including insets from the main orthomosaics. We incorporated three different subjects at each location and grouped them together to quickly see differences in the thermal imagery at different points in the day. "Combine abstract and introduction into one section" We agreed that the introduction was redundant to the abstract, and was replaced. We also removed the satellite images of our
Read more

Abstract and GIS Day

-
On top of the progress we have made in data collection and processing, we have also written the abstract for our poster and applied to participate in Purdue GIS day on November 5th. The abstract was to be 250 words or less and encompass what our project is about. Abstract: The high-resolution data captured from a UAS, such as thermal imagery, can be utilized in a wide variety of applications. The flexibility of using UAS to capture thermal data allows for multiple sets of data to be captured in a relatively short period of time at an almost on-demand basis. To best display the abilities of thermal imagery, the temperature changes of an area throughout the day were captured with the thermal sensor. These temperature changes are called diurnal heat patterns. Current data collection methods for studying diurnal heat patterns include satellite and aerial photography. However, problems with these methods include cost, speed, timing, and resolution. To address these challenges, a DJI M60
Read more

Whitepaper Conclusion

-
Due to the US covid outbreak in late March, in person classes were cancelled for the rest of the semester and an emphasis was placed on the completion of the whitepaper. Here is the link to the completed whitepaper uploaded onto google drives: UAS Thermal Whitepaper
Read more