The Research Of Handwriting Tracking Method Based On Camshift

As a very important area of research in computer vision technology, moving object tracking has a wide range of practical applications and important significance in the military, intelligent transportation systems, industrial control, defense security, and biomedical science research. It roles to the daily livelihood of the people and brings convenience for the life. The result of tracking algorithm has a direct impact on the stability and accuracy of the object tracking, a semi-automatic algorithm design is used in the traditional moving target tracking, that limits the practicality of the algorithm to manually initialize the position of the tracking target. Tracking templates are mostly based on the target feature or motion information, although the tracking reliability can be achieved, the templates need to deal with a large amount of data and its computation is complex, which limits the real-time tracking.This paper firstly discusses the image pre-processing method and characteristics matching algorithm, then studies and analyzes the advantages and disadvantages on a variety of target detection and classic tracking algorithms. This paper applies the meanshift tracking algorithm based on color information, analyzes two aspects of feature extraction and iterative research, uses a improved optimization on this basis of the algorithm. The feasible scheme is put forward in order to achieve the requirements of the tracking system. Mainly studied the differential frame processing in the visual system, to determine the initial target state, dynamic detection and location of the nib target. Tracking process conducted sub-block selection, to determine the direction of motion, shrink search coverage, and further optimized algorithm. The method in this paper is that, to dynamically positioning the interested target from a continuous frame sequence, predict the movement rules of the target, identify the position information of the target area in the next frame. The method has realized the goal of accurate tracking and increased the robustness and real-time performance in the tracking process, satisfactory results are obtained.