Data collection by a dynamic area of interest camera technique
A method of data collection for cameras applicable to any sport where a greater understanding of fluid movement can be obtained by having a faster frame per second or greater resolution for certain parts of a scene. The method coupling a computer to a first stage camera and a second stage camera, focusing said first stage camera on a defined dynamic area of interest, directing said second stage camera on a reduced area of interest, replacing certain data collected from the first stage camera with data collected from the second stage camera, and outputting a compilation onto a display.
1 . A method for improving data collection for a dynamic area of interest comprising the steps of:
coupling a computer having a microprocessor to a first stage camera and a second stage camera, said second stage camera having a greater frame rate and/or greater resolution than said first stage camera;
focusing said first stage camera on a defined dynamic area of interest defining a first set of coordinates and capturing frames into a video feed connected to said computer having a high speed interface, wherein a real-time computer vision algorithm coupled to said computer processes said video feed to observe a player's motion;
directing said second stage camera on a reduced area of interest defining a second set of coordinates;
replacing said coordinates from said first camera with said coordinates from said second camera forming a compilation having the same time line;
estimating linear velocity and the rotational axis of a ball using an object detector and a template matching on subsequent frames in conjunction with the coupling said first stage camera and said second stage camera and directing of said second stage camera onto the ball through said coordinates coordinated by said first stage camera and the combination of imaging from both cameras onto an imaged scene;
outputting said compilation provides faster frame per second and/or greater resolution for certain parts of said imaged scene.
2 . The method according to claim 1 including the step of processing said reduced area of interest to encompass the same or a similar area of interest as said first stage camera.
3 . The method according to claim 1 including the step of attaining a model that predicts a path curve to have less granulation between frames.
4 . The method according to claim 1 including the step of attaining a model that predicts a path c copied from a previous movement that is then n peated for the data taken.
5 . The method according to claim 1 including the step of using polarization to determine features, said dynamic area of interest enhancing said polarization by determining which angle light bounces off the surface of an object, collecting data, and splitting said data into four sub pixels each finding different polarization angles.
6 . The method according to claim 1 including the step of improving tracking by placing said first stage camera and said second stage camera 90 degrees with respect to each other, wherein said tracking measures redefined location.
7 . The method according to claim 1 including the step of processing said dynamic area of interest and said reduced area of interest using Gigabit Multimedia Serial Link (GMSL) cameras placed at calculated locations.
8 . The method according to claim 1 herein the state of compiling includes a ball flight algorithm modeling the hitting of diff su aces and how much grass got in the way.
9 . The method according to claim 1 including the step of tracking a name on a golf ball utilizing said first stage car and said second stage camera.
10 . The method according to claim 1 including the step of tracking dimples on a golf ball utilizing said second stage camera.
11 . The method according to claim 1 including the step of utilizing multiple cameras separated by a wide baseline to enable accurate 3D estimation.