论文部分内容阅读
针对基于单特征不能准确识别复杂环境下空域目标问题,利用偏最小二乘法(PLS)预处理飞机图像特征,通过变量投影重要性分析法(VIP)选取有效图像特征,并将有效特征融合到粒子滤波算法中,实现了颜色与形状特征融合的粒子滤波跟踪算法。该方法在雾霾和遮挡情况下的跟踪结果准确,且跟踪速度得到提高。
Aiming at the problem that the single feature can not accurately identify the target of airspace in complex environment, the PLS is used to preprocess the image features of the aircraft, the effective image features are selected by the variable projection importance analysis (VIP), and the valid features are fused to the particles In the filtering algorithm, the particle filter tracking algorithm of color and shape features is realized. The method has the advantages of accurate tracking results under haze and occlusion conditions and improved tracking speed.