Comparing Different UAS Platforms
Introduction
Last week our AT 319 class dedicated a gathering data on the Purdue Wildlife Area. For this mission, our class was divided into groups of two to form different crews. Each crew was given a platform to use with a unique sensor. With multiple datasets being collected, our goal was to compare and contrast between the different sensors.
Study Area
Our study area was relatively similar to our last mission. We are again operating over the Purdue Wildlife Area located Northwest of Purdue airport. The missions took place on 05/22/19. The weather at the time was partly cloudy with a wind of 10 knots from the North. The outside temperature was 72 degrees.
Methods
Prior to the flights, we laid out five Propeller aeropoint GPS markers clockwise around our mission area. The fifth marker was positioned in the center for optimal coverage. Laying out these markers are important for ensuring higher quality end results for the field images. The image to the right is the Propeller GPS marker. Below is a list of sensors we used in order to gather this information. There are also links attached with much more information regarding the sensors.
Hasselblad: 1" CMOS, FOV 77degrees, 20 Mega Pixels
M600 Olympus : 16 Mega pix, 84 FOV
M600 Stock: 16 Mega pix, 72 FOV
H520 E-90: FOV 90 degrees +- 3 degrees, 20 Mega https://www.yuneec.com/en_GB/accessories/cameras/e90/specs.html
My flight crew consisted of myself and Austin Sullins. I was the designated pilot for the Mavic 2 Pro mission and Austin was the visual observer. There were no problems during the mission. My only complaint was my landing being subpar because I landed slightly off to the side of the landing pad.
Discussion
The maps and data presented below were made possible by Pix4d and ArcMap.
Above is the different orthomosaics created by the four different sensors. There were 3 platforms total and the M600 was used twice. The two different M600 sensors were the olympus and stock sensor. You are able to see that the sensors with the wider field of view created larger mosaics than the others seen.
The map seen above compares the different DSMs taken by the 4 sensors. Inset A shows how the different sensors were able to capture the trees. The higher quality images appear to have fewer smooth flat surfaces were the trees should be. It is seen in the Hasselblad sensor that the trees are more defined from above while the others sensors have trouble. This could be attributed to the smaller pixel amount in the other sensors and the higher field of view. The higher field of view tends to try and stretch the pixels in order to complete the images
The only platform to feature the pond to the West was the E-90 sensor on the H520. This was either due to the the wider field of view or the fact that the mission was made with farther boundaries to the West.
It is interesting to note the differences in image quality. Out of all the platforms, the Mavic 2 Pro had the most clear images of the cars. The other sensors created somewhat blurry or smudged images while the Mavic had the most crisp clear image.
Conclusion
The differences seen are excellent examples for when and where different sensors can be favorable. If a mission requires a more crisp image then the Mavic 2 Pro could be the desired platform to use. Although, if the mission requires a larger area to be surveyed than a wider field of view would be the recommended sensor to use. This could mean that the M600 Olympus or the H520 E-90 sensor would be ideal. The key takeaway from this lab is that choosing your sensor is integral for good data collection.
No comments:
Post a Comment