image processing - player size with different Zoom -


i have got many videos youtube soccer players, need estimate player size different zoom in various videos.

i know player height around 185cm, if assume height of far player (zoom out) 185 , make stander case depend on it, can estimate player height @ different zoom, if zoom in more stander case height larger 185cm, don't have information focal length or distance player.

how can make proper estimation @ different zoom, assuming stander case?

  1. i not think can measure player height , using presumptions on height same time.

    because in way presume height avg 185 cm measurement gets same value. instead can isolate more 1 players in standing pose (the more better), create average z distance camera map (assuming players 185cm in height). interpolate / smooth between measured grid points, filter out inconsistencies , compute player height again.

  2. more precise lock onto known object size

    for example think rail unified size. if captured slight angle can estimate camera frustrum , compute z-map or camera parameters more precisely.

    another superb object ball (on ground)... can obtain zoom @ 2 different distances camera , know ground flat. can try compute ground plane equation parameters + camera distortion whole image ...

  3. if have hi-resolution video stream

    then may better detect steps (during run near predefined speed) while walking average step distance people (almost regardless on height) 70-75 cm. detect player want measure , can obtain zoom of him directly ... height measurement should more accurate

    also if know race of player can lock on more racial average features head size,arm/leg size ratio, etc... there studies that...


Comments

Popular posts from this blog

qt - Using float or double for own QML classes -

Create Outlook appointment via C# .Net -

ios - Swift Array Resetting Itself -