Is not possible fly at lower height?

Hi. I processed an orthomosaic I flew at 20 meters but in spite of it was a low height, I couldn’t see plants of pineapple cuz the orthomosaic lost quality when I processed. More resolution is taken flying lower I think. Is not possible flying lower with dronedeploy?

@empire099
20 meters is very low… Read the reply by @RemotelyPossible at Capturing Long Canals

I typically fly my sites at 100+ meter so that I can get a better stich of images. I have found that many times the quality of the ortho degrades as I go much lower on my sites. Most of my sites are 200+ acres.

1 Like

What speed did you have the drone set to fly and what were your camera settings?

Hey @empire099,

So you are half correct, in a single photo you are getting better resolution at a lower altitude. however since you are going to be sticthing these photos together the software needs more data to accurately process and determine tie points. As @dragonflyAS alluded to i have hit on this topic in another post but ill gladly help clarify here. It might seem weird but we have to separate the detail in one photo from the overall quality of orthomosaics. The lower you fly the more detail in a smaller area you begin to capture (this is great for a single photo) but, it means that the photo itself actually contains less data over all (since it is capturing a smaller area). For high quality orthos flying higher means you are covering more area per a photo and therefore each photo contains more raw data. This is essential for the software to be able to accurately determine the tie points in each photo to be able to render a high quality ortho.

In general the higher you fly the better the overall ortho will come out. This is a balance game, i always suggest flying several test flights at various altitudes to find what works best for your needs. It is also important to remember not every job requires a hammer. Meaning while an ortho is great and cool it is not always the best tool for the job in some cases using the individual photos at lower altitudes may be what you need; in others flying higher with a great ortho will get you the results you want. The key is to be “product oriented” do what you need to do to be able to get the results you need for a client.

So in short… ironic i suppose… you can fly lower but that may not get the result you are looking for.

Scott Lashmit

@remotelyPossible thank you for your feedback, you were so clear in your response.

Well, I appreciate your point of view, however I have not been able to work with dronedeploy for my work. I need to count plants of pineapple with a software which needs a lot of pixels in a pic, so the way I have got better results is flying with Pix4D cuz with it I can fly at 10 meters high and take a good photo that makes me able (and the software too) to make a difference between each plant (results below)

I need this quality for my counting software, a flight in 10meters of high. So, this is the problem.

By the way I congratulate to DroneDeploy because the way it generates orthomosaics is the fastest way I know in my conditions, better than Pix4D.

Hey @empire099,

Thanks for the kind words. Pix4D nice software set. Yeah i try to remind people that sometimes for a particular result you are going to have to be flexible with how you get your data processed. For certain things Using Pix4D will get you where you need to be, other things Drone Deploy is a better option (cloud computing is my favorite part as when i use pix4D on a huge data set my computer is used for several hours). I haven’t tried it personally but Drone Deploy does have an app specifically for what you are doing, but again i haven’t used it so am unclear as to the abilities and how detailed the picture needs to be. I would say you could get a balance between height and detail by experimentation.

Most of my flights result in 2cm per pixel or less resolution which should be sufficient to identify individual plants (assuming its done by computer vision based on RGB datasets). Let us know how it works out!

Best

Scott