In this video, we test the precision landing feature of the DJI Spark DJI Mavic Pro and DJI Inspire 2. How close is the precision features? Well, I test it thoroughly 
Precision Landing of test DJI Spark DJI Mavic Pro and DJI Inspire 2
- Thread starter LaunchHSV
- Start date
- Joined
- May 1, 2015
- Messages
- 1,512
- Reaction score
- 684
- Location
- Los Angeles, CA
- Website
- www.advexure.com
Joseph – Awesome test and thanks for sharing! Before watching I would have said that the Spark has the most accurate RTH, surprised and happy to see Mavic be as spot on as it was. Not all that surprised by the results of the I2. Have always seen it in the +/- 1-3 feet range. A little bit disappointing to not see an improvement with the GPS Module over the aircraft's GPS but in the end, how often are I2 owners solely relying on RTH? We always say it's a feature there if you need it, do not rely on it.
Safe flying!
Safe flying!
- Joined
- Oct 12, 2016
- Messages
- 1,090
- Reaction score
- 569
Glad the I2 didn't land at the centre of the circle where you were standing when activating RTH with the GPS module, or it might have given you a close shave
The smaller drones do have good precision - but when you think that you want the spark to launch and land on your hand, then it needs to be able to get good accuracy. Similarly with the Mavic - people will probably launch and return from things like tables and walls, so it's good that it can manage to set back down pretty accurately.
The I2, well, so long as it gets within a couple of feet of where it launched from, thats good enough - I don't think may of us will launch from tabletops or walls or hands with the I2 A semi precision landing would be nice with it though so it could at least make it back on to a 3ft or 5ft launch-pad. Most times I've let it land itself, it's managed to get at least a couple of legs on the pad, but a bulls-eye would be nice!
The smaller drones do have good precision - but when you think that you want the spark to launch and land on your hand, then it needs to be able to get good accuracy. Similarly with the Mavic - people will probably launch and return from things like tables and walls, so it's good that it can manage to set back down pretty accurately.
The I2, well, so long as it gets within a couple of feet of where it launched from, thats good enough - I don't think may of us will launch from tabletops or walls or hands with the I2
A semi precision landing would be nice with it though so it could at least make it back on to a 3ft or 5ft launch-pad. Most times I've let it land itself, it's managed to get at least a couple of legs on the pad, but a bulls-eye would be nice!
Last edited:
I am a huge fan of the Mavic lol. I am disappointed that I spent $30USD on the GPS Module only to see that it does not work as advertised....
- Joined
- Oct 12, 2016
- Messages
- 1,090
- Reaction score
- 569
The smaller ones are "more accurate" because DJI have implemented some code to make use of the downwad vision sensors to take a location reference image after take-off - it's why they need to rise vertically to a minimum height. That straight up movement and height give the vision system a decent resolution image which it can utilise for comparision on the return leg. The GPS will get it in the ball-park area, but has drift. The vision reference image is the bit that gets it the accuracy - it's also why they say that the take off point shouldn't have lot of things moving around in it... or the reference image simply won't be comparable.
There's probably no reason they can't do the same on the I2 (they've implemented it on the older P4 after all!), just they need to allocate engineering resources to it - and I guess they're busy enough with other i2 problems at the moment, like keeping the firmware stable!!!!
Possibly in time we might also see forward terrain-follow added, again similar to the P4 - the sensors are there, just they need to create the code to do it.
There's probably no reason they can't do the same on the I2 (they've implemented it on the older P4 after all!), just they need to allocate engineering resources to it - and I guess they're busy enough with other i2 problems at the moment, like keeping the firmware stable!!!!
Possibly in time we might also see forward terrain-follow added, again similar to the P4 - the sensors are there, just they need to create the code to do it.
Wonder if the I2 inaccuracy has something to do with that odd glass plate used for VPS with just the I2 model? Seems you wave it around while the I2 is stationary, or I did. Very odd tilting required to get that blue thermometer scale to hit top and bottom to advance. Might have something to do with the illumination on it too as i did it indoors and maybe the light wasn't enough, or there were shadows (I could see shadows or vignetting on the two VPS camera screens in Assistant 2. Their instructions aren't real good with their glass plate method.
I doubt if it has anything to do with hand catching of the smaller ones or accuracy of them. No doubt DJI discourages it in the instructions due to safety (I have neither the Mavic or Spark to check.).
I doubt if it has anything to do with hand catching of the smaller ones or accuracy of them. No doubt DJI discourages it in the instructions due to safety (I have neither the Mavic or Spark to check.).
- Joined
- Oct 12, 2016
- Messages
- 1,090
- Reaction score
- 569
One of Spark's main sales points is that you can launch n recover from the palm of your hand - remember they're aiming for the ultimate selfie drone... to make sure they kill completion like karma, Lilly etc
The main reason will simply be engineering time to implement the code on the I2, that and whether they think it will be something that the I2 customer wants...
The main reason will simply be engineering time to implement the code on the I2, that and whether they think it will be something that the I2 customer wants...
Similar threads
