Field Notes 2: Sensor Comparison

Intro:

Despite what a lot of people think not all UAS sensors are created equal. There are a large variety of sensors that you can get for just one single UAV. In this post I will discuss the differences in the four sensors we used on three different UAV platforms flying relatively the same mission.

Study Area:

We were at Purdue Wildlife Area,8000 IN-26, West Lafayette, IN 47906. Weather was a perfect 70 degrees Fahrenheit with winds out of the SSE at 3 MPH with no gusts with clear skies at 400 feet 
AGL. We flew the Mavic 2 pro at an altitude of 400 feet AGL and the camera set at 90 degrees down.

Methods:
My crew and I went to fly a smaller section of the same field we flew in "Field Notes 1". We set out five GCP's. One in each corner and one in the middle, starting in the north east corner working in a clockwise direction ending with the middle GCP. We picked the GCP's in the opposite order to make sure that when we got back to the lab the data would be easily collected and the GCP's would be labeled correctly.

We flew a mission with three different UAV platforms with a total of four different sensors, which are labeled in the "Discussion" section. The order in missions went H520, M600, M600 Olympic lens, then Mavic 2 pro. My crew flew the Mavic pro 2, we made sure to get plenty of overlap and to fly past the boundaries of the area we wanted to cover to make ensure that we had clean edges and boarders. Flying at 91 meters (300 feet) at 8.9 meters per second we completed our mission in 9 minutes. With no mistakes in the mission.

Discussion:

The specs for each sensors on each platform are labeled below. If you want to find more information about the sensors then just click on the link and it will take you to the companies site of each sensor for all of the more in depth specs for all uses.

Sensors:

Mavic 2 Pro Hasselblad: 20 Mega Pixel, 77 degree FOV
https://www.dji.com/mavic-2/info

M 600 Olympus: 16 Mega Pixel, 84 degree FOV
https://asia.olympus-imaging.com/product/dslr/mlens/12_20/spec.html

M 600 Stock: 16 Mega Pixel, 72 degree FOV
https://www.dji.com/zenmuse-x5/info#specs

H 520 E-90: 20 Mega Pixel, 90 +- 3 degrees FOV
https://www.yuneec.com/en_GB/accessories/cameras/e90/specs.html

Like I said earlier not all sensors are the same and not just any sensor can be best for every job and below are just a few examples of the differences that can occur in relatively identical missions.

Figure 1:

Below you will see that there are a few differences in the four orthomosaic images collected from the four missions flown. The first difference that is probably the most noticeable is the difference in area covered by the E-90 lens, you can easily see the entire west tree line and even see the west pond. This is do to the wide opening lens. This may seem like a benefit, but as I will point out later there are downsides to this. The image also is very dull and lacks true representative color because there are additional settings that were not changed that would allow for a more vibrant image.

Figure 1

Figure 2:
Looking at the map below you can see two sections that are blocked off in red and highlighted. These are the two sections we will be discussing. In the top highlighted section on the map you can see the four different layers of the four different missions flown. In the orange layer, the E-90 sensor, there is a large amount of gelling happening with the images. This causes the map to turn out blurry in that section. That is why you see large pyramid like shapes outside of the other layers of the map.

On the left hand highlighted section we can clearly see the pond that I brought up earlier. The unique part of capturing the pond in this mission that neither myself nor anyone else in my crew had ever seen before was that you could actually see the water. In every other situation I have ever encountered the water shows up as either a white or black hole. This is because it is so hard for the software to get a point that it can use over and over again sense the water is always moving.

Figure 2

Figure 3:
The Mavic 2 Pro is the smallest UAV and smallest sensor out of all the platforms and missions flown that day which made some members of other crews think that it would would under preform compared to the others, however this couldn't be farther from the truth! The Mavic flew the fastest mission and had the clearest orthomosaic out of all of them. In the images below we are comparing the detail in our vehicles. The Mavic has hands down the most clear image of the vehicles. The members of each crew were moving around throughout each flight so it is hard to be able distinguish people from other objects. With that being said it is easier to see the difference with the Mavic. In front of the Black SUV, the middle vehicle, you can clearly see a crew member sitting on the ground with his legs bent up.

Figure 3

Conclusion:
Before you fly any mission you need to make sure you have the best sensor for the job and for your UAV. Each mission will need unique setting and possible unique sensors. As you can tell every sensor has its advantages and disadvantages. Which is why you need to do your research on the mission itself so you can be best equipped to produce the best data for yourself or for your client.

Field Notes 1

Mission 1 Field Notes: 5/14/2019

Intro:

Despite common belief not all GPS systems are created equal. In this lab my team and myself preformed a mission setup that would allow us to collect three separate GPS data points from the exact same locations using photos from each team member’s phone, single solution, and fixed solution GPS signals.

Purdue Wildlife Area

Study Area:

We were at Purdue Wildlife Area,8000 IN-26, West Lafayette, IN 47906. Weather was a perfect 63 degrees Fahrenheit with winds out of the SSE at 3 MPH with no gusts with clear skies at 400 feet 
AGL. We flew the Mavic 2 pro at an altitude of 400 feet AGL and the camera set at 90 degrees down.

Methods:

Propeller GCP’s:

We deployed the plates in a scattered way that was unorganized. We took two bags filled with five plates each. Four team members left with each bag and scattered out throughout the entire field. We didn’t note what plate was turned on first or even where the plates were. This caused us problems later on for the cleanup when we had to walk the field in a line looking for the GCP plates. Even with that small oversight we were still able to collect accurate GPS data using the Measure website and the Propeller software.

Traditional GCP’s

The reach unit couldn’t connect to the cellular server. Which made it so that we had to go and make a static measurement at each of the GCP’s known as Single Solution GPS Coordinates. We made sure not to lay these down right next to the Propeller GCP’s so we could get the most data possible.

Photo Based GPS of the Traditional GCP’s

After we took the static location of the Traditional GCP’s each team member took a photo with their phones so we could find out how far off our phones GPS was from the more accurate GPS devices.

Traditional GCP

Propeller GCP

Conclusion 

Mission
My crew executed a mission that had a few errors in it that ended up being great learning experiences for me and my entire team. We initially planned to fly this mission at 121m (400 feet) over the Purdue Wildlife Area for a test of a software system that is designed to find color codes and circle them on the images to aid in search and rescue.

The mission was then flown at 300 feet and the data was collected and used for more learning experiences with arcmap and other arc software frames. I went and expanded out into more areas that would be beneficial to whomever will be using this data. Below I have the images of the data that I have made.

Map of Purdue Wildlife Area
As you can see in Figure 1 there is such good detail in the images that you can clearly see the GCP's that are less than 2'x2', and also a person (Evan Brueggemann) laying in the grass. You can also see how much error there is between a phone GPS and the actual GCP's that we placed out in the field. In some locations the phone GPS was within a few meters of, and in other locations it was no where near the GCP. This just proves that our GCP's are more accurate and reliable then phone GPS this is why we use them. It also is easier to get the images and data approved by professional surveyors which is a critical part of data being reliable.

Figure 1

3D Flood Diagram

I decided that it would be critical to know the critical flood areas, normal flood area, and obstacle over 50 feet. The reason I chose 50 feet or more is because that is something that is used for all air crafts. Distance for takeoff over a 50 foot obstacle is something that is discussed in every pre-flight for manned aircraft, and is something that will also be helpful to drone pilots and crews.

All areas that are red are obstacles over 50 feet, the dark blue areas are your critical flood areas, and the lighter blue is the area that will flood but is less likely to. The green area is the ground above the flood zones and below 50 feet. Most of the critical flood area is across the road in another field that is not the Purdue Wildlife Area. There is one small section that is prone to heavy flooding, but it is not near any parts of the field where anything is planted so the workers don't have to worry about their crops getting ruined from a normal amount of rain.

Figure 2

Figure 3

Welcome to my Portfolio!

My Background

I am a super senior at Purdue University majoring in Unmanned Areal Systems. Once I graduate from Purdue I will be heading off to pilot training for the United States Air Force! So if you can't tell already from my future job and my major I love flying and anything involving the aviation industry. I remember when I came on my visit here I was getting recruited by the football team to play for them and was on a totally different life path, then a few injuries changed all of that! I ended up flying with a friend  few times and fell in love with it, so I decided I wanted do something aviation and I wanted to push myself in an industry that was growing so I could make a big dent in it, and UAS is that industry. 

Critiques

Hans-Olof Gustafsson, AT 31900, Spring 2019

I love this portfolio! it is my temple and somewhat guide for how I want mine to turn out. I definitely would have changed a few of the color schemes with the font ant texts,but other than that it is basically perfect.

Travis Haas, Geo 336, Spring 2014 

This portfolio is very clean with very few mistakes. However, I find it to be very bland and unattractive. It seems like he didn't put a lot of effort into making the portfolio and just wanted to get a decent grade for it. He did a very good job at taking pictures of everything that he used in the projects.

Lori Bogstad, Geog 390, Fall 2015
Her organization of everything on the page is good. The content in her blog is weak and unsupportable. Her photos are blurry, she doesn't really explain what they are doing and they seem like they are just thrown in there. I like how she kept an accurate flight log.

Kayla Coonen, Geog 336, Spring 2017
I really like her color scheme and the way that it makes the information pop while still remaining professional and clean. Her photos are high quality and are all placed in meaningful and thought out locations. I wish she would have used more header tabs because it seems somewhat empty up top with only three. You can see an improvement from her first blog all the way along to her last.

Issac Flath, Geog 341, Spring 2018
A very clean and professional look as soon as you open up his portfolio. It isn't over dramatic, but still adds more visual reference than just a black background. Some of his photos were either blurry or they were taken unprofessionally outside of what looked to be his college apartment. His portfolio is very easy to manage and to look through I can find exactly what I need to find easily.