leegreen, post: 402618, member: 2332 wrote: The two shadows in your ortho show that you had a significant amount of time between the photos. Looks like two separate flights. This would actually make it easier to remove the shadows with Pix4d. Since you can mask the shadow out of some images, and the actual ground is still visible on the other images. With the image I show below, that is not possible as I have no original ground visible under the shadows. Due to short flight time. One nice part of the shadows is for signs, you can clearly identify a stop sign (2 stop signs in this photo).
BTW: The orange painted control point is not mine. I set the photo target in bottom right. Whom ever painted that tried show direction to other control points, and the nail is not centered within the painted triangle.
The ortho I created was from 2 separate flights about 30 minutes apart and flown perpendicular as you suggested earlier in this thread. It came out much better than my previous attempts and I can't wait to try again after your rolling shutter setting recommendation. So it sounds like you edit the photos before stiching? That makes sense.
Gregg
GMPLS, post: 402712, member: 8404 wrote: The ortho I created was from 2 separate flights about 30 minutes apart and flown perpendicular as you suggested earlier in this thread. It came out much better than my previous attempts and I can't wait to try again after your rolling shutter setting recommendation. So it sounds like you edit the photos before stiching? That makes sense.
Gregg
Gregg,
Yes, process step one. Then you can make edits.
If I remember correctly the ortho is created in step three.
leegreen, post: 402618, member: 2332 wrote: Whom ever painted that tried show direction to other control points, and the nail is not centered within the painted triangle.
Unbelievable stuff like that happens, it makes a difference, basically sabotaging accuracy right out of the gate.
SHG
It appears agisoft has added rolling shutter support, though I am starting to question its value.
http://www.agisoft.com/forum/index.php?topic=5144.0
I have looked at the pix4d documentation and they describe the benefits in a worst case scenario. I think realistically this software feature adds very little accuracy to a project results, especially in a modern slow moving multi rotor below a few hundred feet.
The pix4d description, for example, claims an accuracy on GCP of 0.2 mm, claiming to have amazingly measuring the point to within 3/100 of a pixel. How that is even possible is frankly mind boggling.
I'm very interested to see real world results.. I suspect the benefits are grossly exaggerated.
Would either of you be willing to share datasets?
That's a lot of data to share.
My smallest data set of aerial photos is 900Mb.
leegreen, post: 402804, member: 2332 wrote: That's a lot of data to share.
My smallest data set of aerial photos is 900Mb.
We regularly share data that exceeds a gigabyte, I just put it on Dropbox in a zip, let it sync and then send a link.
Wait... you are able to use these $1500 drones (phantom 4 pro) to create orthophotos, topographic maps?
Sorry to hijack I just would love to know as I only have tinkered with the idea of learning this but every 'survey grade' uav I see is like 15k.
I would assume you use some auto flight program as you obviously cannot get any good results flying manually.
What flight control program do you use with this drone to do the flight path/control to do the photogrammetry?
Rich, there are several free smartphone apps that automate the flight for you. I use the Pix4D Capture app for my flights. You just window around the site using a Google Earth style interface and then select the flight speed, photo overlap and height above ground. After that you essentially press a button and the quadcopter takes off, shoots the photos and lands without any other user input. It really is slick.
Gregg
GMPLS, post: 406332, member: 8404 wrote: Rich, there are several free smartphone apps that automate the flight for you. I use the Pix4D Capture app for my flights. You just window around the site using a Google Earth style interface and then select the flight speed, photo overlap and height above ground. After that you essentially press a button and the quadcopter takes off, shoots the photos and lands without any other user input. It really is slick.
Gregg
Yes I was looking this up last night. Real neat!!!
Pricing makes me realize that other drones like the ebee are really the price they are due to the software they come with, not the drone itself.
Which version of pix4d do you have? What is the difference with the non commercial version. Would this still allow measurements and contour creation?
Right now I don't need the full priced pro version as I would need a ton of time on my own to learn this stuff before I can implement it to my business.
Rich., post: 406344, member: 10450 wrote:
Yes I was looking this up last night. Real neat!!!
Pricing makes me realize that other drones like the ebee are really the price they are due to the software they come with, not the drone itself.
Which version of pix4d do you have? What is the difference with the non commercial version. Would this still allow measurements and contour creation?
Right now I don't need the full priced pro version as I would need a ton of time on my own to learn this stuff before I can implement it to my business.
Rich,
I use Pix4dPro Mapper and Phantom 4 Pro. In my opinion the high dollar drones are not necessary for accurate ortho photos and aerial mapping topographic mapping. The most important part of the drone is the sensor quality, not number of pixels. But the larger pixel cameras, often have better sensors.
I have compared results from the
With the Phantom 4 (P4) and 12MP camera the flight AGL needed to be much lower than all the others, to yield the same accuracy. For example to get 0.8" pixel or GSP, I found the P4 had fly at about 160ft Above Ground Level (AGL) With the new Phantom 4 Pro (P4P), I can get the same accuracy fllying at 300ft AGL.
The Phantom's and Albris do use RTK GNSS, with Ground control points it is no needed. The Phantoms both outlast the Albris in battery life. Most of the other copter drones do have 6 or 8 blades. I'm told this will allow the copter to return if 1 or two motors fail.
The Ebee is a fixed wing plane. It can stay in the air 2x or 3x longer than any copter, but it requires RTK GNSS and a large land area. There is no gimbel on the fixed wing, because it would no hold the punishing landings. These fixed wings have no landing gear, they literally crash land other their belly.
The Ebee is a fixed wing plane. It can stay in the air 2x or 3x longer than any copter, but it requires RTK GNSS and a large land area. There is no gimbel on the fixed wing, because it would no hold the punishing landings. These fixed wings have no landing gear, they literally crash land other their belly.
I just saw the new ebee 2 from Chuck Snow at Earl Dudley in Biringham last week. It can fly over an hour. It has a great sensor and they have dropped the price. It comes with rtk that can be turned on later and doesn't have to be purchased in the beginning. They have also made some good upgrades to the airframe.
Flying a $1500 drone made a little nervous the first few flights. Now I can to land the copter in the palm of my hand, every time.
Landing a $18,000 fixed wing drone with RTK on it's belly (aka crash landing), would be nerve racking every time.
leegreen, post: 406410, member: 2332 wrote: Rich,
I use Pix4dPro Mapper and Phantom 4 Pro. In my opinion the high dollar drones are not necessary for accurate ortho photos and aerial mapping topographic mapping. The most important part of the drone is the sensor quality, not number of pixels. But the larger pixel cameras, often have better sensors.
I have compared results from the
With the Phantom 4 (P4) and 12MP camera the flight AGL needed to be much lower than all the others, to yield the same accuracy. For example to get 0.8" pixel or GSP, I found the P4 had fly at about 160ft Above Ground Level (AGL) With the new Phantom 4 Pro (P4P), I can get the same accuracy fllying at 300ft AGL.
The Phantom's and Albris do use RTK GNSS, with Ground control points it is no needed. The Phantoms both outlast the Albris in battery life. Most of the other copter drones do have 6 or 8 blades. I'm told this will allow the copter to return if 1 or two motors fail.
The Ebee is a fixed wing plane. It can stay in the air 2x or 3x longer than any copter, but it requires RTK GNSS and a large land area. There is no gimbel on the fixed wing, because it would no hold the punishing landings. These fixed wings have no landing gear, they literally crash land other their belly.
Yes, i took a course on the drones and the ebee was the main focus. Very neat. The ebee also can be purchased no RTK and then you need ground control, which is fine for me.
You said the P4p HAS RTK and doesnt need ground control? that seems impossible for a $1,500 drone. Even though I would rather fly lower, I probably would still prefer the pro. Maybe fly somewhere in between those altitudes as any site I fly would be much smaller in size that most people are flying. I think last year I had a 10 acre parcel priced to be flown for me. A few times a year this happens where I need to add contours to an existing larger survey. This would make more sense if I can figure out how to use it, and how to use it correctly!!! I love technology and am very tech savvy so I dont see it being an issue.
I do wrestle the thought as our county GIS has downloadable contours (DXF) that are remarkably accurate. They must have flown the county. Although aligning them correctly to the map is a crap shoot, but for rough subdivision approvals they do the job. However, I wouldn't use them for design work.
So this mostly would be a project for me to satisfy my personal need to learn and keep my tech brain happy. If I am able to use it and implement it a time or two a year then I would be ecstatic.
Sorry, I meant to say the Phantoms and Albris do NOT use RTK GNSS, you must use GCP's.
Trying to type while doing drafting, should have proofed read that.
The problem with aerial topo from imagery, is you get elevations on the top of vegetation. This works well for dirt projects. but tall grass in the middle of summer can relay be a problem.
The specs on the P4 Pro say it can fly for 30 minutes. Actual flight time collecting photos is 20 minutes, then allow 10 minutes for safe return landing.
How much land can your P4p cover in 20 minutes of collecting? Would this be able to do a good detailed job over 10-20 acres in one flight at a decent height? (200-300)
At those AGLs you can easily fly 40 acres with 75ãÉ overlapping photos.
One other advantage to the Phantom drones, there are many inexpensive and easy to use software with apps available for both Android and iOS.
The high dollar drones use costly proprietary software.
Here is a 20 minute mission plan of a 38.5ac site at 256ft AGL with 75% overlap, 386 images using the P4P.
This software will reduce the flight speed if it detects excessive wind speed. If the runs low on battery, it will return to home, allowing you to insert new battery, then it will resume the mission.
leegreen, post: 406481, member: 2332 wrote: Here is a 20 minute mission plan of a 38.5ac site at 256ft AGL with 75% overlap, 386 images using the P4P.
This software will reduce the flight speed if it detects excessive wind speed. If the runs low on battery, it will return to home, allowing you to insert new battery, then it will resume the mission.
That's awesome.
Seems to me in my area I can use something like this to do a survey of something like a golf course or beach club. Especially for a topographic survey. But even to locate features such as tree lines, buildings and parking lots.
I would still need to run the boundary and locate anything along the boundaries such as fences.
With 0.8" pixel size, I can see the top of a 4" fence, if it is not under a tree.
Does the pix4d capture app support the Phantom 4 pro?
Sent from my SAMSUNG-SM-N920A using Tapatalk
