The above picture was done with Skycatch online processing.
With the different cloud processing out there for maps I have tried several of them with different results and wanted to write what I have experienced. I ran a data set through mapsmadeeasy.com with and without GCPs. I have had great success with dronemapper, dronedeploy, without GCPs. When I ran the same set of photos through mapsmadeeasy.com without GCPs the whole map was offset. At first it didnt bother me because I wasnt using them for surveying (just a complete aerial overview) but the more I thought about it I started to question why the offset. Eample map without GCPs Why did the data set work on the other programs but not theirs? I had flown the flight with GCPs so I decided to go back and run it again and this time input the GCP data. I had 5 GCPs during this flight. This time the results were very different. Same map with GCPs
Same data set ran through Skycatch's online processing with GCPs Tylertown-road with GCPs.pdf notice the difference.
From my experience with cloud based processing
If you don't need to share a example map online and just need the basic files dronemapper comes in at the top for accuracy without a subscription (pay as you go)
Ease of use Dronedeploy comes in as number one but it adds extra hardware to the UAV and it has a monthly fee. with that fee you get excellent results without GCPs but still have the option to use them. You also have the availability to share example maps to potential customers online.
For mapsmadeeasy it comes in in between the two for ease of use. You can pay as you go small maps are free and you have a lot more options for file outputs. If you are wanting it to look professional be sure to process it with GCPs and have at least 5 to reference.
Comments
We have had a lot of feedback from users that have found that our calculated results are generally within 1% of RTK measured values and are almost always within 2%. We can't speak to the accuracy of other services but the key to getting accurate results using Maps Made Easy's volume measurement tool is to use a lot of points around the stockpile that are on the base surface on which the pile rests. 4 points is not sufficient except for with the flattest base surfaces. 20-30 will generally yield the best and most repeatable results.
Can anyone comment on comparing the accuracy of volumetric calculations between mme and dronedeploy (or other cloud based software). I'm currently evaluating the two by comparing the same data set and getting wildly different results (mme aprox 20% lower volume - 6483m3 vs 7845m3 with dd). I was hoping the pile would be surveyed but plans changed and it will not be so I can't figure out which one is closer to the correct value.
Great, I will use this workflow as soon as the weather improves. good job and you explanation videos are very useful, I joined yesterday
For comparison's sake and with Aaron's permission, we did a sample using a basemap to do the georeferencing. If you are going to judge a map's accuracy by how it lines up with a basemap, you might as well use that basemap as the input. It only takes a few minutes and gives really well aligned results.
Basemap Georefereced Version of Aaron's map
Excellent! This is a win-win. It lets me know the areas I can work on to improve my input. I'll try a little more waypoint averaging and check and compare elevation settings. I am also testing a 2 axis gimbal to keep the camera orientation more parallel with the ground.
I will be using your service more as soon as I can get a break in the weather.
Aaron,
Yes, we certainly made some adjustments yesterday that resulted in a flatter models in our Geotagged Image and Manual Georeferencing workflows. These modifications are permanent. We had been accepting our outputs as being pretty good, but after getting such a comparison, we knew we had room to improve.
Also, we manually adjusted some of your GCPs. Something wasn't adding up so we checked the points that had been entered. They were off by quite a bit in elevation (80m or more) and that was causing issues by tilting the map. The tilt was slight but was enough that the edges weren't lining up very well. Altitude measurements are always really hard to take with handheld devices and even small errors can greatly affect the outcome. This is why we usually suggest using a pretty high-end receiver and multiple readings when manually recording GCP locations.
We are glad to hear that you approve!
Maybe the Team over there can give some insight to what adjustments they did to make this happen and are they permanent?
I was going to upload some log file last night but got busy with other projects.
We would be glad to take a look at some of the Pixhawk data files if anyone is interested in passing them along. We have an online log file viewer that doesn't require the upload of any files to be used. It provides a quick way to view your basic flight stats like altitude and speed.
We can add any known file format if there is enough interest.
Maps Made Easy Log File Viewer
Glad it was of some use. I used the Pixhawk FC and trigger on distance with a Canon S100 for this flight then use the log to geotag the pictures. The data logger sounds promising for the other flight controllers like DJI that don't allow you to get into the logs and use the GPS. I know some people that will be interested in this.
Aaron,
Thanks for the writeup! We will look into the cause of that offset. Clearly that is not supposed to be there! It looks like there was a bit of curvature to the created model that was pulling the edges in. But, if the other services can do it, we can. This is actually the first comparison writeup we have seen so it gives us something to work with.
We generally recommend the use of GCPs either with our basemap georeferencing method or by entering the recorded coordinates when you need maximal accuracy or when you are using a camera that doesn’t have a GPS receiver in it (like a GoPro). You would be amazed how bad geotags can be sometimes.
17 Acre Excavation Site made with a GoPro 3+ Black and a RageCam 5....
Like Dronedeploy, we will soon be offering an inexpensive positional data logger that will be capable of recording the drone’s location so users don’t have to rely on occasionally sketchy EXIF tag accuracy. It won’t be 3/4G based so there won’t be a monthly fee associated with it, but it will provide a log file that can be uploaded along side the images with no location synchronization step needed.
We have requested that the admins here at DIYdrones allow our maps to be embedded here like a YouTube video with iFrame code. It would be nice for users to be able to post their maps inline instead of having links off the site.
As you mentioned, we currently offer free processing for smaller jobs and are evaluating the possibility of a pretty steep price cut in the future.
Maps Made Easy
Cheers!
Again, thanks for the comparison.
-
1
-
2
of 2 Next