Optical Flow Method for Object Tracking-Brief Introduction to Optical Flow Method

The concept of optical flow was first proposed by Gibson in 1950. It is the instantaneous speed of Pixel Motion of a spatial moving object on the observation imaging plane, the relationship between the previous frame and the current frame is determined by the change of pixels in the time domain and the correlation between adjacent frames in the image sequence, so as to calculate the motion information of objects between adjacent frames. In general, optical flow is produced by the movement of the foreground target, the motion of the camera, or the joint motion of the two. The calculation method can be divided into three types:

(1) Regional or feature-based matching methods;

(2) frequency-based method;

(3) gradient-based method;

Simply put, optical flow is the "instantaneous speed" of Pixel Motion of spatial moving objects on the observation imaging plane ". The Research of optical flow is to determine the "motion" of the respective pixel positions by using the temporal variation and correlation of pixel intensity data in the image sequence ". The purpose of studying the optical flow field is to obtain a sports field that cannot be obtained directly from the image sequence.

 

Assumptions about the optical flow method:

(1) The brightness between adjacent frames is constant;

(2) The frame acquisition time of adjacent video frames is consecutive, or the motion of objects between adjacent frames is relatively "tiny ";

(3) maintain spatial consistency. That is, the pixels of the same sub-image share the same motion.

 

There are two concepts to be explained:

Sports Ground is actually the movement of objects in the 3D real world;

The optical flow field is the projection of a sports field on a two-dimensional image plane.

 

As shown in, the pixel (x, y) in H moves to the position (x + u, Y + V) in I, and the offset is (u, v ).

 

The principle of Optical Flow Method for Target Detection: assign a velocity vector to each pixel in the image, thus forming a motion vector field. At a specific time point, the points on the image correspond one to one with those on the 3D object. This correspondence can be calculated through projection. The image can be dynamically analyzed based on the velocity vector features of each pixel. If there is no moving target in the image, the optical flow vector continuously changes throughout the image area. When an image contains a moving object, the target and the background are relatively moving. The velocity vector formed by a moving object must be different from the background velocity vector, so that the position of the moving object can be calculated. It should be noted that, when using the Optical Flow Method for moving object detection, the calculation amount is large, and the real-time and practicality cannot be guaranteed.

Principle of Optical Flow Method Used for Target Tracking:

(1) processing a continuous video frame sequence;

(2) Use certain target detection methods to detect potential targets for each video sequence;

(3) If a frame has a foreground target, find its representative key feature points (which can be randomly generated or used as feature points );

(4) For any two adjacent video frames, find the optimal position of the key feature points in the previous frame in the current frame to obtain the coordinates of the foreground target position in the current frame;

(5) Such iteration can achieve target tracking;

Http://blog.csdn.net/carson2005/article/details/7581642

Http://blog.csdn.net/gxf1027/article/details/8650878 (cvpr)

Http://blog.csdn.net/chenhongc/article/details/5755878 gaussian background

Optical Flow Method for Object Tracking-Brief Introduction to Optical Flow Method