基于数据融合的红外目标识别

为了获得良好的红外目标识别性能,综合应用了图像处理、模式识别和数据融合领域内的新技术.采用了神经网络和证据理论集成的数据融合方法进行目标识别的数据融合.根据LVQ神经网络在目标识别领域内应用特点,构造了基于证据理论的基本概率赋值函数.对此目标识别技术进行了测试,结果表明,采用此技术后的识别的可信度得到了较大提高.

IR target recognition based on data fusion

In order to get good capability of IR target recognition, new techniques such as image processing, pattern identification and data fusion are applied, The method integrating neural network with evidence theory is used to data fusion of target recognition, According to the characters of LVQ neural network application in the field of target recognition, the function of evidence theoretical basic probability evaluation is constructed. The limited samples are used to test the target recognition technique. The result of target recognition test shows that the recognition reliability with this data fusion technique is higher than that without the technique.