Search Results
7/23/2025, 6:57:21 PM
>>16732031
>the highly expensive FLIR
1. you don't even understand why those gimbals cameras are so expensive
2. the resolution of that FLIR is 640x512 and even in a correctly encoded video you're still seeing a heavily post processed to make it easier to interpret and remove noise
> near perfectly
I see a lot of visual glitches in that object and ground lights.
>t and is miles away from the object.
angular size =/= size
>can clearly see with your eyes the camera zoomed out
Zoom affects all the objects in the FoV in the same way, it doesn't alters the angular size that is the only thing you're seeing. So you can't estimate size by the background, your perception of movement and zoom.
That camera of the video is controlled by moving a virtual point on the ground, that works fine for objects far away and/or moving on the ground but if you add a object way closer then it creates illusions, a better tracking method would be simple visual tracker that ignores the geo coordinates. Tip: if you were able to remove the background then you (and the researcher) would interpret the video and object in a very different way.
>the highly expensive FLIR
1. you don't even understand why those gimbals cameras are so expensive
2. the resolution of that FLIR is 640x512 and even in a correctly encoded video you're still seeing a heavily post processed to make it easier to interpret and remove noise
> near perfectly
I see a lot of visual glitches in that object and ground lights.
>t and is miles away from the object.
angular size =/= size
>can clearly see with your eyes the camera zoomed out
Zoom affects all the objects in the FoV in the same way, it doesn't alters the angular size that is the only thing you're seeing. So you can't estimate size by the background, your perception of movement and zoom.
That camera of the video is controlled by moving a virtual point on the ground, that works fine for objects far away and/or moving on the ground but if you add a object way closer then it creates illusions, a better tracking method would be simple visual tracker that ignores the geo coordinates. Tip: if you were able to remove the background then you (and the researcher) would interpret the video and object in a very different way.
Page 1