图象目标识别图形的自适应步级检测

建立图象目标识别模型，用形状、灰度和运动特征描述图象目标。基于目标建模，把目标识别、门限及目标图形区域步级检测、虚漏警调节和目标空域条件有机地结合起来，给出牵引式跟踪系统中图象目标识别图形的自适应步级检测算法。该算法用于检测具有识别特征的图象目标图形，并成功地应用到实际的实时跟踪系统中的图象目标识别和跟踪。对实际图象的处理结果和在实时跟踪系统中的实验说明本文研究的技术的有效性和适用性。

Adaptive and Progressive Detection of Image Object Recognition Figure

The image object recognition model, which describe image object using shap, grey level and motion characters, is built. Based on the modelling ohect, a algorithm of the adaptive and progressive detection of the image object recognition figure in the led tracking system is presented by combining organically the object recognition, the progressive detection of the threshold and the object tigure region, the modification of the false alarm and false dismissal with object spatial condition. The algorithm is used to detect the image object figure with recognition characters and successtlilly appIied to the recognition and tracking of the image object in the real-time tracking system.Some experimental results for rela images and the experiments in the real-time tracking system show that the technology in this paper is etlbctive and applicable.