Simple introduction of optical flow method

The concept of optical flow was first proposed by Gibson in 1950. It is the instantaneous velocity of the pixel motion of the space moving object in observing the imaging plane, it is to find out the correspondence between the previous frame and the current frame by using the change of the pixel in the image sequence and the correlation between the neighboring frames, so as to calculate the motion information of the object between the adjacent frames. In general, the light flow is caused by the movement of the foreground target itself, the motion of the camera, or the joint motion of the two in the scene. The calculation method can be divided into three categories:

(1) Region-based or feature-based matching method;

(2) The method based on frequency domain;

(3) Gradient-based method;

In simple terms, the optical flow is the "instantaneous velocity" of the pixel motion of a space moving object in the observation of the imaging plane. The study of optical flow is to use the time domain variation and correlation of pixel intensity data in the image sequence to determine the "motion" of the respective pixel position. The purpose of studying the optical flow field is to approximate the field which can not be obtained directly from the image sequence.

The premise hypothesis of optical flow method:

(1) constant brightness between adjacent frames;

(2) The frame time of adjacent video frames is continuous, or the motion of the object between adjacent frames is "tiny";

(3) Maintain spatial consistency; that is, the pixel points of the same sub-image have the same motion

Here are two concepts to explain:

The sports ground is actually the movement of the object in the three-dimensional real world;

The Optical flow field is the projection of the playground on the two-dimensional image plane.

As shown in the figure above, the pixels in H (x, y) move to the position (X+U,Y+V) in I and the offset is (u,v).

The principle of optical flow for target detection is to assign a velocity vector to each pixel in the image, thus forming a motion vector field. At a certain moment, the point on the image corresponds to the point one by one on the three-dimensional object, which can be computed by projection. According to the speed vector feature of each pixel, the image can be dynamically analyzed. If there is no moving target in the image, the amount of light vectoring is continuously changing throughout the image area. When there are moving objects in the image, there is a relative movement between the target and the background. The velocity vectors formed by moving objects must be different from the velocity vectors of the background, so that the position of moving objects can be calculated. It is necessary to remind that the use of optical flow method for moving object detection, the computational capacity is large, can not guarantee real-time and practical.

The principle of optical flow method for target tracking:

(1) The processing of a continuous video frame sequence;

(2) For each video sequence, a certain target detection method is used to detect the potential target;

(3) If a frame appears a foreground target, find its representative key feature points (can be randomly generated, you can also use corner points to do feature points);

(4) for any of the following two adjacent video frames, look for the key feature points appearing in the previous frame in the current frame of the best position, so that the foreground target in the current frame position coordinates;

(5) The tracking of the target can be achieved by this iteration;

Citation: http://blog.csdn.net/carson2005/article/details/7581642