Virtual Litchi Mission

[Bussty]

Hello Namirda

Thank you so much for trying!

Google Studio is pretty awesome lets you make smoother movies with lots of controls I just don't think it allows you to follow a path....yet?

Thanks again

Bussty

 

[Bussty]

I have another question... I have been doing some test flyover in the mountains of New Zealand and it seems I might have a bug as shown in this test Litchi File Mission Hub - Litchi

It seems if I'm using POI's and I change from one POI to another in the transition from the waypoint looking at POI 1 (Waypoint 3) to the waypoint looking at POI 2 (Waypoint 4) the gimbal seems to rise up from looking down at the POI 1 (Waypoint 3) then rise to a horizontal height then fall back down to look at POI 2 at Waypoint 4.

Is this normal? I would have thought the gimbal would transition seamlessly between the two POI's.

Is this an issue because I'm looking down on my POI's? That would seem weird as most drone missions would look down on any POI's

Any help gratefully received, thank you

 

[namirda]

I have another question... I have been doing some test flyover in the mountains of New Zealand and it seems I might have a bug as shown in this test Litchi File Mission Hub - Litchi

It seems if I'm using POI's and I change from one POI to another in the transition from the waypoint looking at POI 1 (Waypoint 3) to the waypoint looking at POI 2 (Waypoint 4) the gimbal seems to rise up from looking down at the POI 1 (Waypoint 3) then rise to a horizontal height then fall back down to look at POI 2 at Waypoint 4.

Is this normal? I would have thought the gimbal would transition seamlessly between the two POI's.

Is this an issue because I'm looking down on my POI's? That would seem weird as most drone missions would look down on any POI's

Any help gratefully received, thank you

Hi,

Unfortunately DJI do not publish information about exactly how the camera heading and tilt change between waypoints during a mission and so I have had a bit of a guessing game with VLM to try to come up with something that matches what DJI appears to be doing. It's not an exact science!!

In general, the camera heading and tilt between waypoints are chosen so as to give the smoothest possible video track which can sometimes be a bit different from that expected from linear interpolation. In most cases this fits quite well with the test flights I have done over the years.

In your case where WP3 and 4 have rather similar gimbal angles but very different headings this results in the gimbal tilt swinging upwards as you have observed.

If you want to force the gimbal angle to follow the valley floor then the simplest way is to have VLM insert interpolated WPs between your existing WPs. To do this, go to VLM/Setup/Change Mission Settings/Smoothing and set the "Infill Distance between Smoothed WPs" to be 50m.

As a general rule, its better to avoid having large changes in heading or gimbal angle between wps - better to add extra wps to force the mission to behave as you expect.

Hope this helps

N

 

[Bussty]

Hi,

Unfortunately DJI do not publish information about exactly how the camera heading and tilt change between waypoints during a mission and so I have had a bit of a guessing game with VLM to try to come up with something that matches what DJI appears to be doing. It's not an exact science!!

In general, the camera heading and tilt between waypoints are chosen so as to give the smoothest possible video track which can sometimes be a bit different from that expected from linear interpolation. In most cases this fits quite well with the test flights I have done over the years.

In your case where WP3 and 4 have rather similar gimbal angles but very different headings this results in the gimbal tilt swinging upwards as you have observed.

If you want to force the gimbal angle to follow the valley floor then the simplest way is to have VLM insert interpolated WPs between your existing WPs. To do this, go to VLM/Setup/Change Mission Settings/Smoothing and set the "Infill Distance between Smoothed WPs" to be 50m.

As a general rule, its better to avoid having large changes in heading or gimbal angle between wps - better to add extra wps to force the mission to behave as you expect.

Hope this helps

N

Hello again Namirda

Thank you so much for your reply that was helpful but I fear I may have been wasting your time. I was out hiking today and I got to thinking have I got my head around the Chrome Extension Version of this software? For some reason I was thinking by downloading the 2.7.1 VLM version it somehow interfaced with the Chrome Extension but of course I was loading and editing my missions directly into Litchi Mission Hub then exporting using the Extension effectively bypassing the new 2.7.1 VLM version! So when I reran things through VLM 2.71 (after turning off the Chrome Extension) it worked perfectly using strung POI's in POI mode! I am very happy

I also had some issues previously combining Normal Waypoints with POI Waypoints but by turning off POI and Using Interpolation with some manual adjustment to the Pitch angles I now have a very smooth flight in VLM 2.71. It was still pretty good having POI as my Gimbal Pitch control but there was just a little jarring as I transitioned from initial waypoints at the start to POI's from WP 3 as attached...

Mission Hub - Litchi

Plan waypoint missions from the comfort of your chair!

flylitchi.com

So there has been some awesome learns for me today...

• For best results launch Litchi Mission Hub from within VLM 2.7.1 and use the world icon to launch in Google Earth Pro
• Do not use Litchi by way of the Chrome VLM Extension to export to Google Earth Pro
• If you have a mission with all POI's VLM 2.7.1 works really well even for extreme changes.
• If you have a mission with small to moderate changes from Normal WP's to POI way points just use Interpolate and POI for Gimbal movements, should work fine and as expected.
• If however you are combining normal WP's and POI's with extreme changes between the two I will...
  • Firstly in mission settings set Heading Mode to Custom(WD) and Gimbal Pitch to Focus POI - The focus POI will setup your Heading Angles and if you change to interpolate for affected POI WP's these headings won't change which is what you want if you are going to manually set them up.
  • Add all your POI's and set your POI WP's to Focus POI plus select the POI you are pointing to. This will set up all your POI waypoints to point at your POI
  • Make sure all WP's are on Gimbal Pitch / Interpolate regardless of whether a standard WP or POI WP
  • Now go back to mission settings and change Default Gimbal Pitch Mode to Interpolate
  • Have a trial run of the flight using the World Icon in VLM 2.7.1 which will launch in Google Earth Pro
  • Review the flight and edit any gimbal pitches particularly the WP's that were originally POI WP's
  • Review any changes in Google Earth Pro until happy.
  • Done!
• So basically by having Gimbal Pitch to POI then removing it you set up the headings for the POI's but mucked up the gimble pitches so need to enter those manually (A little trial and error) but because all your waypoints are now on Interpolate mode the gimbal should smoothly transition through any WP change.

Hope I have that right?

Thanks again

Bussty

 

[namirda]

Hello again Namirda

So there has been some awesome learns for me today.......

Hi Bussty,

That's roughly what I do as well - use Focus POI to set the heading/gimbal and then switch to Interpolate to make minor adjustments.

But a few comments:

1) "Default Gimbal Pitch Mode" only sets the initial Pitch Mode for newly created WPs - so perhaps not relevant here.

2) Your curve size for WP3 is zero which maybe adds to the jarring - see the abrupt path change in GE. Increasing curve size will help.

3) I cannot see what smoothing parameters you are using in VLM but in such rugged terrain I suggest an infill distance of 50m and smoothing points per WP to be 21

4) You might want to take a look at post 768 in this thread for more information about how DJI handle POI in hotpoint mode

Your mission is very instructive because it has such extreme changes in ground elevation - I have never tried a mission where the POIs are about 1000m lower than the WPs!

These unusually large differences in elevation between WPs and POIs highlight a potential problem:

1) When I use Litchi's Focus POI to obtain initial gimbal tilt and headings for WPs 3,4, and 5 in the most recent mission you posted above I find value values of -15, -18 and -7. I guess you see the same? This is clearly nonsense - they should be -28,-41,-33 according to VLM calculations. I think is a bug in Litchi which I will investigate further.

2) If you switch on verbose mode in VLM you can see both the initial gimbal angles as calculated by Litchi and those after smoothing/interpolation as calculated by VLM which are much more in line with those you have estimated by trial and error.

3) This error is a bit concerning because Litchi transmits gimbal angles to the drone during the mission (even for drones using onboard wps) and it looks as if those might be wrong. Have you actually flown this mission? Did the footage match that calculated by VLM or was the camera pointing much too high?

4) Take a look at the flight path in GE as it passes your WP2. It's a good example of how you can easily crash your drone. The curve smoothing results in the drone flying about 100m lower than the altitude set at your WP - beware! I have had a crash due to exactly this mistake!

5) Assuming you take off from somewhere near WP1, you will very likely lose signal as you drop down to WPs 7 and 8 at the end of the mission. This will result in loss of gimbal control (if your drone uses onboard wps) or total loss of control (if you are using virtual sticks). Simple RTH is not a good idea in terrain like this! And what happens at the end of the mission if you don't have control anymore??

Good Luck with the mission!

N

Edit... I have just noticed that the elevation labelling of your two POIs in the Mission Hub is [1(-200)] which is incorrect. It looks as if Litchi have introduced a hard limit of -200m on the elevations of POIs wrt WP1 for some reason and this is leading to errors in the calculation of the gimbal angles in Focus POI mode.

I have already sent a bug report to Litchi

 

[Bussty]

Hi Bussty,

That's roughly what I do as well - use Focus POI to set the heading/gimbal and then switch to Interpolate to make minor adjustments.

But a few comments:

1) "Default Gimbal Pitch Mode" only sets the initial Pitch Mode for newly created WPs - so perhaps not relevant here.

2) Your curve size for WP3 is zero which maybe adds to the jarring - see the abrupt path change in GE. Increasing curve size will help.

3) I cannot see what smoothing parameters you are using in VLM but in such rugged terrain I suggest an infill distance of 50m and smoothing points per WP to be 21

4) You might want to take a look at post 768 in this thread for more information about how DJI handle POI in hotpoint mode

Your mission is very instructive because it has such extreme changes in ground elevation - I have never tried a mission where the POIs are about 1000m lower than the WPs!

These unusually large differences in elevation between WPs and POIs highlight a potential problem:

1) When I use Litchi's Focus POI to obtain initial gimbal tilt and headings for WPs 3,4, and 5 in the most recent mission you posted above I find value values of -15, -18 and -7. I guess you see the same? This is clearly nonsense - they should be -28,-41,-33 according to VLM calculations. I think is a bug in Litchi which I will investigate further.

2) If you switch on verbose mode in VLM you can see both the initial gimbal angles as calculated by Litchi and those after smoothing/interpolation as calculated by VLM which are much more in line with those you have estimated by trial and error.

3) This error is a bit concerning because Litchi transmits gimbal angles to the drone during the mission (even for drones using onboard wps) and it looks as if those might be wrong. Have you actually flown this mission? Did the footage match that calculated by VLM or was the camera pointing much too high?

4) Take a look at the flight path in GE as it passes your WP2. It's a good example of how you can easily crash your drone. The curve smoothing results in the drone flying about 100m lower than the altitude set at your WP - beware! I have had a crash due to exactly this mistake!

5) Assuming you take off from somewhere near WP1, you will very likely lose signal as you drop down to WPs 7 and 8 at the end of the mission. This will result in loss of gimbal control (if your drone uses onboard wps) or total loss of control (if you are using virtual sticks). Simple RTH is not a good idea in terrain like this! And what happens at the end of the mission if you don't have control anymore??

Good Luck with the mission!

N

Thanks so much for your reply!

Firstly that glitch I mentioned in an early post where altitude fell then rose is completely gone when using Standalone VLM 2.7.1 they that is very cool.

Thanks for all that info here is some feedback...

1) Yes re default modes was wondering the same thing so thank you for confirming.

2) Thanks re curve normally do that but must have slipped through. Amended and seems very smooth... thanks!

3) Yep thanks got both those locked in

4) Cheers will check out post 768


Additional...

1) Yes I did notice those gimbal discrepancies though once you took it off POI the gimbal pitch would just remain at what was necessary for the POI's but are way too small.

2) Yes see what you mean re verbose readout and gimbal pitch

3) Too be honest this whole dive back into VLM was to make a movie of my teenage sons hiking trip and Litchi and VLM makes it incredibly easy to do (more so than Google Earth Studio) I have also used VLM more flying my MIni and has always performed really well. I have ordered a Mini 3 Pro so hoping will be Litchi compatible eventually... So no haven't flown this mission, would love to though feels pretty epic but in a National Park

4) Lol if I had a multi $000 drone flying this yes I would be having an extra close look at those clearances

5) as per 4

Thanks again Namirda looks like I'm going to be able to make some stunningly smooth movies using the Stand Alone VLM 2.7.1 of my kids trips, the other one hiked into Lake Lochnagar so will do that one too. Also really looking forward to using with the Mini 3 Pro eventually.

I see you have found a bug so happy to have been able to help

Cheers

Bussty

 

[DJ Wes]

Sorry for the late reply - for some reason the mavicpilots forum software does not always notify me when a new post is added here.

And I'm sorry but I cannot replicate this problem - have tried a few different csv files but have not experienced the problem you describe.

I would be very surprised if your problem is anything to do with VLM because the latest versions no longer open or read the csv file at all - the data is now taken directly from the Litchi DOM.

Anybody else having this problem?

N

I have installed VLM 2.7.1 on a different computer ran a test and got the same locked file result. Here are the steps to reproduce this problem:

1) Open VLM.
2) Import a .csv file.
3) Use a text editor to make an insignificant change to the .csv file, then save and exit the text editor.
4) Import the same .csv file into VLM (choose "import" then click on "Import to new mission".

You will receive the "file permission" error.

This does not happen in Litchi or older versions of VLM.

 

[rexster1505]

I have installed VLM 2.7.1 on a different computer ran a test and got the same locked file result. Here are the steps to reproduce this problem:

1) Open VLM.
2) Import a .csv file.
3) Use a text editor to make an insignificant change to the .csv file, then save and exit the text editor.
4) Import the same .csv file into VLM (choose "import" then click on "Import to new mission".

You will receive the "file permission" error.

This does not happen in Litchi or older versions of VLM.

Not sure about that. I saved one of my missions as a csv file with a couple of numbers changed. Imported into VLM and it opened just fine.

 

[namirda]

I have installed VLM 2.7.1 on a different computer ran a test and got the same locked file result. Here are the steps to reproduce this problem:

1) Open VLM.
2) Import a .csv file.
3) Use a text editor to make an insignificant change to the .csv file, then save and exit the text editor.
4) Import the same .csv file into VLM (choose "import" then click on "Import to new mission".

You will receive the "file permission" error.

This does not happen in Litchi or older versions of VLM.

Hi,

Now I see your problem!

Your first description of it was not precise enough - I was tweeking the mission in Litchi but I now see that you meant to tweek the input csv file with an external editor.

I see the file permission error you describe but I get precisely the same error when I follow your steps directly in the Mission Hub - as I said previously, I don't believe this has anything to do with VLM because the latest versions do not use csv files at all.

This problem occurs because you have changed a file to which the browser has already made an internal reference.

It may not be ideal but one workaround is to force the browser to refresh that file reference by first selecting any other csv file (no need to import it - just select it in the 'Choose File' dialog) and then select your initial file again. This will force the browser to refresh the internal reference.

Thanks

N

PS I suddenly realised why we are seeing different things:

The Mission Hub window within VLM runs the Google Chrome browser internally . It could be that other browsers automatically refresh file references without the workaround above. Which browser are you using? I think you will see the same problem if you try with Google Chrome

 

[Bussty]

I reported earlier that when making a Google Earth Pro movie after creating a path in VLM I had a "bump" in my movie, I thought this was due to me making the Litchi path in the Chrome Extension of VLM and not using the standalone Version 2.7.2 as soon as I used the Standalone VLM it worked fine...

BUT...

I have made anther flight/movie using VLM 2.7.2 standalone and I have a similar issue (different part of the map)

Here is a link to the Youtube Video Youtube Link the "bump" happens at about 1.30 and is really noticeable. There is also some slight twitching at the very start.

Here is the Litchi Flight Plan Litchi Flight Link

Note: When you fly the path in Google Earth Pro it doesn't jump you only get it when you make the movie.

It happens if you use GEP to either generate a movie or just a series of jpegs then sequence them in something like Davinci Resolve. NOTE: Generating a series of jpegs then sequencing gives much better results in terms of final quality.

If anybody could have a look and confirm they get the same behaviour would be much appreciated. If you have a fix would also love to know.

I have searched the internet and there is some talk that GEP Movie Maker isn't the best even if saving jpegs to sequence but it is all we have. Screen grabbing isn't really an option as the run through on GEP isn't as smooth as the final video.

Thanks heaps

Bussty

 

[Jack747]

Hello
I don't know if the subject has been covered here, but I have a strange problem with Chrome Litchi Virtual Mission plugin in Firefox 101 the plugin does not work unless I am in private browsing, if I want to export in VLM format i have this message ,

 

[namirda]

Hello
I don't know if the subject has been covered here, but I have a strange problem with Chrome Litchi Virtual Mission plugin in Firefox 101 the plugin does not work unless I am in private browsing, if I want to export in VLM format i have this message , View attachment 149766

Hi Jack,

Virtual Litchi Mission (VLM) is a windows only application which I wrote a number of years ago and have been trying to keep up to date ever since. The Chrome plugin was developed some time later by user @bazuchan to mimic some of the functionality of VLM within a Chrome plugin in order to make it accessible to non windows users.

I am not able to provide support for the plugin - maybe bazuchan will look into it for you.

N

 

[Bussty]

I reported earlier that when making a Google Earth Pro movie after creating a path in VLM I had a "bump" in my movie, I thought this was due to me making the Litchi path in the Chrome Extension of VLM and not using the standalone Version 2.7.2 as soon as I used the Standalone VLM it worked fine...

BUT...

I have made anther flight/movie using VLM 2.7.2 standalone and I have a similar issue (different part of the map)

Here is a link to the Youtube Video Youtube Link the "bump" happens at about 1.30 and is really noticeable. There is also some slight twitching at the very start.

Here is the Litchi Flight Plan Litchi Flight Link

Note: When you fly the path in Google Earth Pro it doesn't jump you only get it when you make the movie.

It happens if you use GEP to either generate a movie or just a series of jpegs then sequence them in something like Davinci Resolve. NOTE: Generating a series of jpegs then sequencing gives much better results in terms of final quality.

If anybody could have a look and confirm they get the same behaviour would be much appreciated. If you have a fix would also love to know.

I have searched the internet and there is some talk that GEP Movie Maker isn't the best even if saving jpegs to sequence but it is all we have. Screen grabbing isn't really an option as the run through on GEP isn't as smooth as the final video.

Thanks heaps

Bussty

I have been trying to solve this puzzle of the "bump" during footage shown here Bump.m4v

from this Litchi mission Mission Hub - Litchi

So far I have tried...

• This happens just after WP 12 so have been focusing on WP 12 and 13
• I removed above ground option and just typed in the total altitude, thought that had a high chance of success but no...
• Changed the curve size, was a 1000 but reduced down to 300ish but no...
• Changed the smoothing points down to 15 but no...
• Tried changing the height of the waypoint but no....
• WP 13 was on focus on POI but has no POI loaded so I thought here we go so put on interpolate but no.... (sigh)

So I was almost ready to give up when I tried one last thing...

You will be there forever waiting for your movie to render if you rely on the drone flying at max speed so in Google Earth Pro so I have been recording tours played back at 3x/4x speed THIS was the issue!! So the way to overcome this issue is to record at whatever speed your tour comes out of VLM/Litchi then if you want to speed up do so in your video editor. If using jpeg sequences you should achieve max quality!! DO NOT SPEED UP IN GEP it will cause issues. It is still weird and something to do with GEP I think. That only occurred in that one spot on quite a long tour and when I started a brand new mission with just a start and an end consisting of two way points travelling a line across where the "bump" happened sure enough if I sped up the playback an recorded that as a tour that movie also had the bump in the exact same place. Recorded at standard speed and it was fine. One really useful function of the increased speed PLAYBACK ability of GEP is to test how your flight will look if primarily using to make moves through Google Earth.

So yay! Sorted

Note: If you ever want to do top down spiraling movies to establish your location on the globe or in space use Google Earth Studio as your initial software then fade that into movies generated using VLM and Litchi with Google Earth Pro. I find it a lot easier to do complex camera runs using that system rather than trying to key frame in Google Earth Studio.

 

[Jack747]

Hi Jack,

Virtual Litchi Mission (VLM) is a windows only application which I wrote a number of years ago and have been trying to keep up to date ever since. The Chrome plugin was developed some time later by user @bazuchan to mimic some of the functionality of VLM within a Chrome plugin in order to make it accessible to non windows users.

I am not able to provide support for the plugin - maybe bazuchan will look into it for you.

N

Thanks, but this bug is only annoying, I can live with it working mostly with linux this VLM plugin is perfect for me, I haven't seen any thread specific to the plugin. If @bazuchan comes here...

 

[OzoneVibe]

I haven't seen any thread specific to the plugin

Not hard to find. Took me one attempt ..

Yet another Virtual Litchi Mission (Chrome/Firefox Extension)

I made Google Chrome and Firefox extension that acts almost exactly as VLM and can be used on Mac/Linux/Win. Virtual Litchi Mission itself is cool software, but it is windows only. I'm a linux user and running google earth in virtual machine was pain, so I spent a day learning/reversing how VLM...

mavicpilots.com

 

[DJ Wes]

Hi,

Now I see your problem!

Your first description of it was not precise enough - I was tweeking the mission in Litchi but I now see that you meant to tweek the input csv file with an external editor.

I see the file permission error you describe but I get precisely the same error when I follow your steps directly in the Mission Hub - as I said previously, I don't believe this has anything to do with VLM because the latest versions do not use csv files at all.

This problem occurs because you have changed a file to which the browser has already made an internal reference.

It may not be ideal but one workaround is to force the browser to refresh that file reference by first selecting any other csv file (no need to import it - just select it in the 'Choose File' dialog) and then select your initial file again. This will force the browser to refresh the internal reference.

Thanks

N

PS I suddenly realised why we are seeing different things:

The Mission Hub window within VLM runs the Google Chrome browser internally . It could be that other browsers automatically refresh file references without the workaround above. Which browser are you using? I think you will see the same problem if you try with Google Chrome

I use Firefox. I just tested the Mission Hub using Chrome and yes, the problem occurs there too.

Something must have changed in the way files are imported from VLM version 2.5 to 2.7 as the problem does not occur n version 2.5.

I develop software tools that generate Litchi missions in the form of csv files. My workflow requires repeated testing in VLM and Google Earth using the steps I outlines previously. The change in file handling breaks that workflow. Your work-around works. However, it adds way to many mouse clicks making the testing process much less efficient. If nothing can be done at your end I will just continue using VLM 2.5 as that version doesn't exhibit this behavior.

 

[Jack747]

Not hard to find. Took me one attempt ..

Yet another Virtual Litchi Mission (Chrome/Firefox Extension)

I made Google Chrome and Firefox extension that acts almost exactly as VLM and can be used on Mac/Linux/Win. Virtual Litchi Mission itself is cool software, but it is windows only. I'm a linux user and running google earth in virtual machine was pain, so I spent a day learning/reversing how VLM...

mavicpilots.com

Thanks !

 

[namirda]

I use Firefox. I just tested the Mission Hub using Chrome and yes, the problem occurs there too.

Something must have changed in the way files are imported from VLM version 2.5 to 2.7 as the problem does not occur n version 2.5.

I develop software tools that generate Litchi missions in the form of csv files. My workflow requires repeated testing in VLM and Google Earth using the steps I outlines previously. The change in file handling breaks that workflow. Your work-around works. However, it adds way to many mouse clicks making the testing process much less efficient. If nothing can be done at your end I will just continue using VLM 2.5 as that version doesn't exhibit this behavior.

Yes - a lot changed between 2.5 and 2.7 - csv files are no longer imported at all.

I completely gave up using the litchi csv files because they do not provide all the information required for Google Earth to fly the mission. In 2.7, the mission data is taken directly from the Litchi's Document Object Model.

Unfortunately there is probably not much either of can do to get Google to fix the file reference problem in Chrome!

N

 

[Memor86]

Hello @namirda , i just learned that litchi has removed the 100 waypoint limit, will you update VLM soon?

 

[namirda]

Hello @namirda , i just learned that litchi has removed the 100 waypoint limit, will you update VLM soon?

VLM can already process missions of more than 99 waypoints - try it and see!

However, the helix and grid options cannot CREATE missions more than 99 waypoints - perhaps that is what you are referring to.

I can easily increase that limit if required but I would like to keep some limit in place to prevent users unintentionally generating enormous missions with thousands of waypoints.

I have never needed more than 99 waypoints in a mission - I always run out of battery before I run out of waypoints!

What would you like the limit to be?

N