Devices and methods for tracking moving objects
The present invention is directed to methods and devices for tracking moving characteristics of at least two moving objects. The present invention identifies an occlusive condition(s) of the moving objects on a video frame(s), analyzes information of a physical framework(s) of the moving objects, finds and records a position(s) of the occluded moving object(s) on the video frames and tracks a moving character(s) of the moving objects. The present invention further provides devices and methods for tracking the moving characteristics of behaviors of living bodies.
1. A method for tracking moving characteristics of at least two moving objects, comprising:
(a) providing a video picture of the at least two moving objects;
(b) transforming the video picture into a plurality of video frames;
(c) identifying if there is an occlusion of the at least two moving objects on the plurality of video frames;
(d) if there is no occlusion occurred, identifying and recording each object's position respectively of the at least two moving objects on the video frames;
(e) if there is an occlusion occurred, locating non-occluded endpoints of physical frameworks of the moving objects in both an occluded situation and a non-occluded situation on the video frames, and then performing the following steps:
(e1) identifying and recording position of non-occluded moving objects on the video frames before and during occlusion;
(e2) creating a distance map of the endpoints of the physical frameworks of occluded moving objects based on a Euclidean distance and a turn direction of the endpoints of the physical frameworks of the occluded moving objects during occlusion on the video frames;
(e3) comparing between the endpoints of the physical frameworks of the occluded moving objects on occlusive video frames and the endpoints of the physical frameworks of the occluded moving objects on previous video frames by using the distant map;
(e4) selecting best matching endpoints with the smallest Euclidean distance and/or a similar turn direction on the previous video frames according to the distance map;
(e5) extrapolating an information of the physical frameworks of the occluded moving objects on the video frames;
(e6) interpolating the position of the occluded moving objects on the video frames; and
(e7) recording the position of the occluded moving objects; and
(f) tracking the moving characteristics of the at least two moving objects on the video frames.
2. The method of claim 1 , wherein the position comprises a centroid and/or an endpoint.
3. The method of claim 1 , wherein each of the non-occluded and occluded endpoints further comprises a head point and/or a tail point.
4. The method of claim 1 , wherein the information of the physical frameworks comprises a vector, centroid, endpoint, turning angle, turning direction, speed, length and/or location.
5. The method of claim 1 , wherein the moving objects further comprises living bodies.
6. The method of claim 5 , wherein the living bodies comprise a fly, fish, mammals, birds, and/or human.
7. The method of claim 1 , wherein the moving characteristics comprise paths, staying time, poses, centroids, endpoints, turning angles, turning frequency, rolling frequency, turning directions, speed, length and/or locations.