Waypoints 2.0 Glitch - Mavic 2 Zoom

[nfarruggio]

Hello,

I ran a waypoint mission 2 times in a row but unfortunately both missions were not identical. I manually positioned the angle of the drone to keep an object in the center of the frame. It accomplished this the initial mission, but the the drone did not keep the object in the center of the frame during the subsequent mission. Has anyone run into this

-Nick

 

[brett8883]

Hello,

I ran a waypoint mission 2 times in a row but unfortunately both missions were not identical. I manually positioned the angle of the drone to keep an object in the center of the frame. It accomplished this the initial mission, but the the drone did not keep the object in the center of the frame during the subsequent mission. Has anyone run into this

-Nick

This seems to be a common issue that maybe updated in a future release but for the time being the code in waypoints 2.0 is something like this in English

Fly to waypoint 1: GPS point x, at altitude y, yaw to compass direction z.

Fly to waypoint 2: GPS point x, at altitude y, yaw to compass direction z.

Fly to waypoint 3: etc etc etc.

What it is missing is the ability to set a POI as a fixed focus point and interpolate movement between waypoints. You can sometimes get lucky with it if there isn’t much yaw direction movement between waypoints but the drone isn’t recording your movement between waypoints only the orientation at each waypoint.

Again might be fixed in the future but for now your best bet is Litchi. In the Litchi app you can set the drone to continuously track a GPS point so the drone will always be looking in that direction and at the specified height. No matter if the drone is blown off course or whatever will always be looking at that point.

The pro trick with poi waypoints is the transition from looking at one thing to the next. I’ve attched a screen shot a Litchi mission I made to show you. Hopefully you can see this well enough.


My Mission starts out with my POI #1, the lake, and points 1-3 are Fixed on the lake looking at a height of 3feet so the drone is almost looking down into the lake. Points 5-8 are looking at POI #2 but that POI is set to be 93 ft in the air, almost the same height as the waypoint itself so the drone will be gazing into the distance in that direction.

Waypoint #4 is the traditional waypoint which I have set not to look in any particular direction per se but it set to “interpolate” meaning it will smoothly act as the half way transition from point 3 to point 5. If I didn’t have this transition waypoint the drone would quickly pivot from the one direction to the next when the POI or “focus point” changed. This is what DJI Waypoints 2.0 is missing and is critical to smooth fluid movements and keeping your subject in view

 

[Michael Gilbert]

Waypoints 2.0 does have the ability to track a POI as a fixed focus point. It sounds like nfarruggio was not using POI but instead framing the shot (position lat and log, yaw, and most importantly pitch of the gimbal) separately at each waypoint.
I've used this method many times and each subsequent flight behaved as the initial. Framing the shot like this gives you a bit more accuracy since you don't have to rely on the POI sitting perfectly at the desired location and height. nfarruggio did not mention any pitch settings in his example which might explain the variance. brett8883, with using your example, I believe Waypoints 2.0 can achieve the same results. At waypoint #4, instead of interpolate, which waypoints 2.0 does not have, you can manually frame the shot to the midpoint position. The results should be the same.

 

[Steady Eddie]

Thanks for that brett8883. I must admit I never understood the "interpolate" property of a Litchi waypoint. Will give it a go next time. Unfortunately my Mavic Pro does not do Waypoints 2.0, so I cant try Michael Gilbert's suggestion.

 

[nfarruggio]

Waypoints 2.0 does have the ability to track a POI as a fixed focus point. It sounds like nfarruggio was not using POI but instead framing the shot (position lat and log, yaw, and most importantly pitch of the gimbal) separately at each waypoint.
I've used this method many times and each subsequent flight behaved as the initial. Framing the shot like this gives you a bit more accuracy since you don't have to rely on the POI sitting perfectly at the desired location and height. nfarruggio did not mention any pitch settings in his example which might explain the variance. brett8883, with using your example, I believe Waypoints 2.0 can achieve the same results. At waypoint #4, instead of interpolate, which waypoints 2.0 does not have, you can manually frame the shot to the midpoint position. The results should be the same.

I have utilized the POI option in Waypoints 2.0. However, I found that when I manually framed the shot in waypoints mode produced a smoother drone motion and thus a smoother video. Along with many other, I experienced jerky/choppy video due to the drone's attempt at keeping the POI in the center of the frame.

 

[JosephAir]

Try centering camera up and down before staring a waypoint and you will find it makes a difference. Enjoy.

 

[DJ Wes]

This seems to be a common issue that maybe updated in a future release but for the time being the code in waypoints 2.0 is something like this in English

Fly to waypoint 1: GPS point x, at altitude y, yaw to compass direction z.

Fly to waypoint 2: GPS point x, at altitude y, yaw to compass direction z.

Fly to waypoint 3: etc etc etc.

What it is missing is the ability to set a POI as a fixed focus point and interpolate movement between waypoints. You can sometimes get lucky with it if there isn’t much yaw direction movement between waypoints but the drone isn’t recording your movement between waypoints only the orientation at each waypoint.

Again might be fixed in the future but for now your best bet is Litchi. In the Litchi app you can set the drone to continuously track a GPS point so the drone will always be looking in that direction and at the specified height. No matter if the drone is blown off course or whatever will always be looking at that point.

The pro trick with poi waypoints is the transition from looking at one thing to the next. I’ve attched a screen shot a Litchi mission I made to show you. Hopefully you can see this well enough.
View attachment 64611

My Mission starts out with my POI #1, the lake, and points 1-3 are Fixed on the lake looking at a height of 3feet so the drone is almost looking down into the lake. Points 5-8 are looking at POI #2 but that POI is set to be 93 ft in the air, almost the same height as the waypoint itself so the drone will be gazing into the distance in that direction.

Waypoint #4 is the traditional waypoint which I have set not to look in any particular direction per se but it set to “interpolate” meaning it will smoothly act as the half way transition from point 3 to point 5. If I didn’t have this transition waypoint the drone would quickly pivot from the one direction to the next when the POI or “focus point” changed. This is what DJI Waypoints 2.0 is missing and is critical to smooth fluid movements and keeping your subject in view

Hi brett8883,

I have used Litchi to plan and fly waypoint missions for about two years and I have to say that I have never experienced the problem you describe with “quickly transitioning” between waypoints. I’ve created waypoint missions with multiple POIs and have never found the need to create intermediate, interpolated waypoint between those that are focused on POIs in order to smooth out the transitions. Perhaps I’m not understanding the issue you describe. So, to illustrate what my experience is, I created a very simple example with:

• Two waypoints both at 49 feet
• Two POIs, one at 3 feet and one at 50 feet

I then flew this mission and made a video. As in your example, I have two POIs, one at 3 feet elevation and one at 50 feet elevation. The heading of the first waypoint is set to point at POI 1 and the heading of the second waypoint is set to point at POI 2. The two waypoints are both at approximately 50 feet while the first POI is at 3 feet and the second POI is at 50 feet. Please let me know if I’m not grasping the issue with this simple mission that you are trying to describe with yours.

What you say is true about not being able to “track” a POI. Instead, what people should understand, is that orientation information (heading, pitch, etc.) is stored for each waypoint and the drone will simply smoothly transition between the orientation of one waypoint to the orientation of the next. Depending on how you set up your POIs and waypoints, the subject may not always remain in the center of your view, but your drone will smoothly transition between the two orientations. However, this is not what you appear to be describing with your statement: “If I didn’t have this transition waypoint the drone would quickly pivot from the one direction to the next when the POI or “focus point” changed.”

What I would expect (and what the drone does) is smoothly transition from pointing at the first POI to pointing at the second POI. I have included both a snapshot of the mission from Litchi, and the resulting video. What am I missing in the problem that you describe?

Snapshot from Litchi:

And, the resulting video:

 

[brett8883]

Hi brett8883,

I have used Litchi to plan and fly waypoint missions for about two years and I have to say that I have never experienced the problem you describe with “quickly transitioning” between waypoints. I’ve created waypoint missions with multiple POIs and have never found the need to create intermediate, interpolated waypoint between those that are focused on POIs in order to smooth out the transitions. Perhaps I’m not understanding the issue you describe. So, to illustrate what my experience is, I created a very simple example with:

• Two waypoints both at 49 feet
• Two POIs, one at 3 feet and one at 50 feet

I then flew this mission and made a video. As in your example, I have two POIs, one at 3 feet elevation and one at 50 feet elevation. The heading of the first waypoint is set to point at POI 1 and the heading of the second waypoint is set to point at POI 2. The two waypoints are both at approximately 50 feet while the first POI is at 3 feet and the second POI is at 50 feet. Please let me know if I’m not grasping the issue with this simple mission that you are trying to describe with yours.

What you say is true about not being able to “track” a POI. Instead, what people should understand, is that orientation information (heading, pitch, etc.) is stored for each waypoint and the drone will simply smoothly transition between the orientation of one waypoint to the orientation of the next. Depending on how you set up your POIs and waypoints, the subject may not always remain in the center of your view, but your drone will smoothly transition between the two orientations. However, this is not what you appear to be describing with your statement: “If I didn’t have this transition waypoint the drone would quickly pivot from the one direction to the next when the POI or “focus point” changed.”

What I would expect (and what the drone does) is smoothly transition from pointing at the first POI to pointing at the second POI. I have included both a snapshot of the mission from Litchi, and the resulting video. What am I missing in the problem that you describe?

Snapshot from Litchi:
View attachment 65916
And, the resulting video:

The OP was having an issue with waypoints 2.0 in the DJI Go 4 app and so I suggested to use Litchi instead because it IS better.

When I was talking about how it looks at a certain compass direction I was talking about DJI Go 4 not Litchi.

Stop

Switching the discussion to Litchi, the example you give is on the right track but isn’t quite what I am talking about. In your example the drone is always looking “south.” Though it dies turn a bit what I am talking about is when the drone turns a COMPLETELY different direction. See below.

In this new example, which I have modeled after yours, we see the drone fly down the road looking at the houses just like in your example. However, in my new example we have the drone looking at the “south” side of the street and then the next waypoint is almost 150 feet higher than the first waypoint.

Now this is where things get tricky. At what point will the drone ascend to the new height and at what point will it yaw to look at the house that is on the North side of the street?

Will it yaw clockwise or counter clockwise to the new waypoint?

POI #2 is at a higher altitude than POI #1 as well so when will the gimbal pitch to look at the new POI?

By adding an “interpolate” waypoint to this mission I can control those things and when they happen.

So now with the addition of the “interpolate” waypoint I am telling it to rise half the altitude between waypoint 1 and 3 by the time it gets to waypoint 2. I also want by this time for it to have yawed enough to focus on POI #2 therefore it will continue looking at POi #2 the rest of the mission.

So I know that from waypoint #2 to #3 it will be focusing on the POI #2 the whole time but will ascend much faster the second half of the altitude from that was from waypoint #1 to #3. This way I know I get plenty of shots of POI #2 and this cool effect of it rising quickly while still focusing on POI #2.

No me I could have decided I wanted a continuously gradual transition but I think this is going to be more interesting. The important thing is that I got to chose instead of just letting the drone figure it out.

I’m sorry I know that was long

 

[DJ Wes]

The OP was having an issue with waypoints 2.0 in the DJI Go 4 app and so I suggested to use Litchi instead because it IS better.

When I was talking about how it looks at a certain compass direction I was talking about DJI Go 4 not Litchi.

Stop

Switching the discussion to Litchi, the example you give is on the right track but isn’t quite what I am talking about. In your example the drone is always looking “south.” Though it dies turn a bit what I am talking about is when the drone turns a COMPLETELY different direction. See below.View attachment 65925

In this new example, which I have modeled after yours, we see the drone fly down the road looking at the houses just like in your example. However, in my new example we have the drone looking at the “south” side of the street and then the next waypoint is almost 150 feet higher than the first waypoint.

Now this is where things get tricky. At what point will the drone ascend to the new height and at what point will it yaw to look at the house that is on the North side of the street?

Will it yaw clockwise or counter clockwise to the new waypoint?

POI #2 is at a higher altitude than POI #1 as well so when will the gimbal pitch to look at the new POI?

By adding an “interpolate” waypoint to this mission I can control those things and when they happen.
View attachment 65927
So now with the addition of the “interpolate” waypoint I am telling it to rise half the altitude between waypoint 1 and 3 by the time it gets to waypoint 2. I also want by this time for it to have yawed enough to focus on POI #2 therefore it will continue looking at POi #2 the rest of the mission.

So I know that from waypoint #2 to #3 it will be focusing on the POI #2 the whole time but will ascend much faster the second half of the altitude from that was from waypoint #1 to #3. This way I know I get plenty of shots of POI #2 and this cool effect of it rising quickly while still focusing on POI #2.

No me I could have decided I wanted a continuously gradual transition but I think this is going to be more interesting. The important thing is that I got to chose instead of just letting the drone figure it out.

I’m sorry I know that was long

Hi brett8883,

Yes, I understand that the OP was asking about DJI’s waypoints 2. And yes, Litchi waypoints is better than DJI waypoints 2 in numerous ways. My comment was in response to your statement: “If I didn’t have this transition waypoint the drone would quickly pivot from the one direction to the next when the POI or “focus point” changed.” To me that statement appeared to suggest that the transition between waypoints was something other than a smooth, controlled rotation. I was attempting to describe how Litchi transitions between waypoints when using POIs (and focus POI).

In your new example, the drone will smoothly rotate counter-clockwise as the gimbal pitch smoothly transitions from POI 1 to POI 2. To answer your questions:

• It will yaw counter-clockwise because that is the smaller of the two possible rotations required to change focus from POI 1 to POI 2.
• The gimbal pitch will also smoothly transition from the POI 1 height to the POI 2 height as the drone travels from waypoint 1 to waypoint 2.

While I agree that you can introduce further control by inserting an intermediate waypoint (as you suggest), with your example, an intermediate waypoint may slow down the transition between 2 and 3 while increasing the rotational speed between 1 and 2. I would go as far as saying that your example is somewhat extreme having two close waypoints each pointing at POIs which would cause close to a 180 degree rotation. One even might argue that such a set of waypoints and POIs constitutes a poorly designed mission.

I can tell from your descriptions that you understand how waypoints work. It was only that one statement that I was trying to clarify. I think we agree pretty much on what you can do with waypoints and POIs. I just didn’t want anyone to be left with the feeling that one had no control on how the drone transitions as you go from one waypoint/POI combo to the next.

Happy flying.