基于统计模型的图像融合

摘要

As a very important component of the electric vehicle ,magneto electric machine has as a signifi‐cant impact on the performance of electric vehicle power train system .This research is based on detection of magneto electric machine in Chery Company .An image fusion algorithm based on Statistical model is proposed for infrared and visible image .This approach is based on fusion rules with proposed weighted average S.imulation experiments show that fusion image with GGD model has a better fusion result .%永磁电机是电动汽车的主要部件之一，其可靠性关系着汽车的整体性能指标。以奇瑞集团永磁电机检测生产线为工况背景，通过大量理论分析和现场试验，提出了一种新的红外与可见光图像融合算法，对源图像小波系数进行广义高斯统计建模的多模态图像融合，采用新的加权平均融合规则来进行小波系数的融合，通过数据估计特征指数和形状参数，判断融合数据是否在广义高斯或对称α稳定的吸引域内；最后，将融合后的系数重构为一幅图像，比较基于这两种统计模型下融合图像的效果。实验结果证明，基于广义高斯小波系数建模的加权平均法融合的效果比较好。