I am a hobbyist wanting to just experience quickly what DroneDeploy can do for me. Yes, I realize that I should read all the info and watch the Academy videos and I WILL do that once I figure out that DroneDeploy is the tool I need to master. =)
So, I am looking for quick “shake and bake” workflow before my trail period runs out. My goal would be to :
- Open a Project
- Import my existing drone photographs of a simple shed structure (taken with overlap, different heights, different angles, etc.)
- Create viewable model from my photos
- Export maybe in LAZ or other point cloud format that I could import into Unreal 5.1
I am seeking a few “do this and do that” steps or a reference to the correct link that covers that type of scope.
2 Likes
From what I can see you have the workflow. Upload your images, they create everything and export the LAS.
1 Like
The DroneDeploy fly apps do not support DJI Mini2. Can I still manually fly and take pictures and submit those shots to utilize the rendering softeware?
2 Likes
Yes. DroneDeploy works with any images which have the required GPS data embedded. The Mini 2 does.
1 Like
OK, I am stuck. =)
I used Litchi control app ( No option for Mini 2 with DroneDeploy) I captured 120 images at different heights and different gimbal angles. Etc.
I opened DroneDeploy on my desktop and created a project and at that view, I clicked on upload and uploaded my jpg images (took about 10 minutes).
The upload process ended but now I am stuck as how to see that 3d model or point cloud or whatever is the next step in 3D model building.
OK, a little quick on the trigger. (no instant pudding) I guess I figured that uploading was one process and working DD rendering magic would take a while. So I sauntered over to docs and found this one to be what I needed to be reading. =) https://help.dronedeploy.com/hc/en-us/articles/1500004861881-How-To-Process-Datasets-into-Maps-and-Models
I noticed that an email would be sent when complete and sure enough an email was delivered about 14 minutes later. I am now off to see what layer of the onion I have arrived at.
This might help. There are 3 icons bottom center of the interface. The model is the middle button. To see the point cloud you go to the model and then turn on the toggle for point cloud at the top of the left navigation pane.
https://help.dronedeploy.com/hc/en-us/sections/1500000794022-Analyzing-Data
And yes indeed you are correct! In fact the model icon is now solid and complete. The email or that icon is the indicator needed to know that rendering is complete. I am off to check out what has been delivered! =)
By the way, I love this forum as no one has ranted at me about “just read the manual, you idiot!” or other troll language. Just good patience and good advice.
Well, Well - I am very I M P R E S S E D! The rendering is complete and by clicking on that middle icon at the bottom I get to see a wonderful 3D model. Now there are holes in the wheelbarrow. And for some reason I only have a legs and no torso but it is clear to me that the if I optimized my picture taking technique the model could only get better. Jeez, I can even see the stitching/sewing in my blue jeans and the grass blade details. Amazing.
I have attached a few screenshots.
Now, I am off to explore exporting and how to enable others to view. Also read the documentation (finally) and improve my picture taking.
And the various tools like measuring distances are spot on accurate!
By toggling on the Map photos Icon in the 3D model view, it is very easy to see why the holes are in the wheelbarrow as there are too may spaces in the location of the camera when the pictures were taken.
The 3D rendering can only get M U C H better with the correct photography.
Yeah there’s no time for that here. It’s supposed to be challenging and fun. You get an occasional rant but compared to some other sites it’s pretty chill. I just answered a few questions on Reddit and feel like I need a tylenol and nap.
That’s a good start! It really is amazing right? You’ll see some things here that are typical issues with photogrammetry. The holes in the wheelbarrow is because of the lack of features and because of it being a microscene it missed the mesh because whatever settings it used did allow it to span that length.
As you mentioned with the torso DroneDeploy has always had an issue with meshing vertical object that are much taller than the things around them on this scale. I bet if you look at the point cloud you’ll see it so the data is there, they just didn’t mesh it.
MichaelL - Please advise he best way for another person to view a model and manipulate it outside of the DD environment. Is that possible. Just and XYZ fly around my model.
What do you mean by manipulate? Just navigating or actual modifying/analysing it?
Blender to clean up and render the OBJ
CloudCompare to clean up and analyze the LAS points
OK, A little background. I am a geocacher that likes to build epic “hide” experiences. Way beyond what is known as “park and grab.” =) I have caches that utilize interactive messaging, etc. I want to go to the next level with clues to the real cache found in 3D models. Think MYST from “back in the day.”
For example maybe a clue(s) is hidden on the side of a shed or barn. By flying around the model, the player can seek out and stumble upon clues that will lead to lat/lon coordinates to the real thing.
So I have a website that I place text based interactive “games” for the above purpose. Now I want to add a 3D model that the site visitor can “fly around” and change viewpoints to virtually “bushwhack” a location.
I do not want them to just interact with dronedeploy site as a client would be referred to. The model is stand alone. I see that there are sites that will host and can be linked to via a window in my website. Even Facebook now accepts 3D Models and I think by using the [FBX2glTF] format.
So, I figure that there must conversion techniques that are similar to converting videos in .mov to .mp4 format for sound and conversion of .tiif files to .jpg format. for pictures And similar file conversion techniques could be applied and ideally would take a dronedeploy output and convert. etc.
Thoughts, ideas links, suggestions? Attached is link for starters: R&D - Delivering 3D Scenes to the Web
Gotcha. So if you are going to be doing smaller scenes Sketchfab is usually the goto but once the models reach a certain size or quantity you have to start paying for it. I believe it’s 10 models per month and 100MB max on the OBJ. What’s the size of yours?
I would recommended going on the Reddit Photogrammetry sub. It deals a lot more with small object modeling so there are probably some pretty cool package we have no idea of.
How do I obtain file size of my DD created model?
I peeked at SketchFab. Price is good. =) Can SketchFab Use Drone images?
I have seen a reference that a DD OBJ file can be exported and then inputed to SketchFab. Where do I find the export as OBJ “button?”
Found it! =)
Yes, but it is very prohibitive according to size. Even a small 5 acre site is 300MB and our average site is 50 acres.
I just kept moving forward with warts and holes just to figure out a work flow. I did get SketchFab to upload the DD .OBJ file. And explored a little. the final model is not near as sharp as the DD version but prolly good enough for what I am trying to do. I have not embedded the animation which eliminates the SF boilerplate yet but …
1 Like
With your help, I have figured out a work flow for my project discussed in this thread.
But, as a hobbyist, I can’t afford the $500 month for the DroneDeploy option. I have written DD to ask about a “Hobby Rate” but I doubt that will be an available for me.
Can you suggest an alternative service for converting my drone generated jpg images into a .OBJ file that I can input into SketchFab?
