视觉目标的识别及跟踪技术的研究

STUDY ON VISUAL OBJECT RECOGNISING AND TRACKING TECHNIQUE

结合应用的需要,通过对一些传统算法的修改优化,提出完整的视觉目标的识别算法。该算法利用图像处理方法,对外框和其中特征点进行计算。并在目标跟踪技术上,提出了一种基于可预测的"搜索窗口"的方法,实验证明,这加快了目标跟踪算法的处理速度,算法能提供优秀的跟踪检测并实时处理。

Combining the application requirements,a complete visual object recognition algorithm was proposed based on modifying and optimising some traditional algorithms.It calculates the outer frame and its feature point with image processing method.For object tracking technique,the paper proposed an approach based on predictable ＂search window＂.Experiment attests that it expedites the processing speed of the object tracking algorithm and the algorithm can provide good tracking detection and processes real-timely.