How to access Depth map SDK?

[lpic]

Hi All,

I am keen to exploit the AI of EVO2 as shown with the tracking and obstacle avoidance videos.
I was expecting to access the depth map from the onboard processing, but there seems to be no mention in the SDK.
Does anyone have knowledge of the if/how/when such an SDK will be made available?

 

[Jagerbomb52]

I would think since the EVO 2 is not even released yet and now is delayed by a month or two any SDK will have to follow after that if ever. Autel is not very fast at releasing anything.

 

[Rorysee]

Sorry, are you guys talking about depth maps as in 3D displacement maps? IF SO...will the EVO 2 be able to generate these maps and will these maps come with a texture map.

I know that the tech exists because I have seen some amazing images from military-based software engineers. IF Evo 2 has this capability will it be seq based or single image-based?

 

[Ansia]

Sorry, are you guys talking about depth maps as in 3D displacement maps? IF SO...will the EVO 2 be able to generate these maps and will these maps come with a texture map.

I know that the tech exists because I have seen some amazing images from military-based software engineers. IF Evo 2 has this capability will it be seq based or single image-based?

If you are talking about photogrammetry, theEvo already does it.

Explorer App's Mission feature and My First Photogrammetry with the Evo

As soon as I got home, I tried the new mission feature in the beta version of the Explorer app. I still need to play with it, since I made some mistakes and it shows on the end result, but I'll get there. Anyways, here are the results. The 3D rendering ended up with a weird black circle around...

autelpilots.com

 

[lpic]

Let me clarify, The 3D depth map is typically used by the vehicle to map the environment in *real time*, thus allow for the autonomous navigation of obstacles. I am only guessing that the EVO2 does this, after i saw the impressive obstacle avoidance videos. Creation of the depth map in real time using only camera sensors is the key technology VSLAM.
Once the 3D depth map is available... then the real fun can begin, identifying shapes as trees. building, people..etc. You get the idea.

 

[Ansia]

Let me clarify, The 3D depth map is typically used by the vehicle to map the environment in *real time*, thus allow for the autonomous navigation of obstacles. I am only guessing that the EVO2 does this, after i saw the impressive obstacle avoidance videos. Creation of the depth map in real time using only camera sensors is the key technology VSLAM.
Once the 3D depth map is available... then the real fun can begin, identifying shapes as trees. building, people..etc. You get the idea.

What you are reffering to is the stereoscopic cameras drones use to create a 3D rendering of it's surroundings, thus avoiding obstacles in real time. To my knowledge, you don't get access to these cameras, other than for OA. At least from all the drones I've seen and/or owned.