为了解决夜间会车滥用远光灯造成驾驶员晕光的问题, 提出一种在IHS色彩空间下改进Curvelet变换融合可见光与红外图像的抗晕光方法。该方法通过改进Curvelet变换实现图像二维细节信息的有效表达, 提高图像清晰度, 其中提出的低频系数权值自动调节融合策略能够将晕光信息剔除, 避免其参与融合过程; 与IHS变换相结合能够保留原图中的色彩信息, 避免色彩失真。对实验结果的主客观分析表明, 该方法消除晕光比较彻底, 与IHS-小波融合相比, 融合图像的标准差、平均梯度、边缘强度、信息熵分别提高了47.15%、53.10%、52.46%、4.45%, 对比度和清晰度显著提升, 细节信息也更加丰富, 人眼视觉效果更好, 有利于驾驶员观察前方路况, 提前做出预判, 消除安全隐患, 提高夜间行车的安全性。

In order to solve the problem of driver halation caused by the abuse of high beam lights at night, the anti-halation method of visible and infrared image fusion was proposed based on improved Curvelet transform in IHS color space. The method could effectively express the two-dimensional detail information of the image by improving the Curvelet transform, and improved the image clarity. The proposed fusion strategy of self-adjusting low-frequency coefficient weights could remove the halation information and avoid its participation in the fusion process. The color information in the original image was preserved by combining with the IHS transform to avoid color distortion. Subjective and objective analysis of the experiment results show that the method can eliminate the halation more completely. Compared with IHS-wavelet fusion, standard deviation, average gradient, edge intensity and entropy of the fusion image increase by 47.15%, 53.10%, 52.46% and 4.45%, respectively, its contrast and clarity are significantly enhanced, the details are also more abundant, and the visual effect of the human eye is better. The method is helpful for the driver to observe the road condition ahead, make judgment in advance, eliminate safe hidden trouble and improve the safety of night driving.