针对夜间汽车晕光现象引起的交通安全问题, 从规避碰撞物的角度出发, 设计了一种红外与可见光图像融合的视频抗晕光系统。系统通过对可见光图像和红外图像做MSR图像增强, 解决了夜间可见光图像亮度低, 暗处信息不易获取的问题, 并提高了红外图像对比度, 提升了融合图像的清晰度; 通过YUV与小波变换结合的方式对增强后的可见光图像和红外图像进行融合, 消除了晕光现象。实验结果的主客观分析表明: 该融合算法比YUV与小波融合算法在熵、均值、平均梯度、标准差上分别提高了1.6%、13.5%、25.3%、0.6%, 该系统不仅能有效消除晕光, 还对融合后图像的亮度和暗处细节信息有较大提升, 提高了夜间驾驶安全性。

To avoid collisions caused by night vehicles halation, a video anti-halation system of infrared and visible images fusion was designed. Visible light image and infrared image were enhanced by MSR enhancement algorithm to solve the difficulty in achieving the dark place information caused by low-light level of night visible light image, and the contrast of infrared image was improved, which consequently improved the definition of the fusion image; And the halation was eliminated by the method of infrared and visible images fusion based on YUV and wavelet transformation. Compared with the YUV and wavelet fusion algorithm, the fusion algorithm proposed in this paper can increase the entropy, mean value, average gradient and standard deviation by 1.6%, 13.5%, 25.3% and 0.6%, respectively. Experiment results show that the designed system can effectively eliminate the halation, and improves the image brightness and details in the picture, which improves the safety of night driving.