Abstract: Based on the digital camera imaging characteristic, this paper has mainly introduced the process of solving the inverse camera response function by the nonlinearity distribution of the image patch’s color along the image edge, and proposed a calculation method of Inverse Camera Response Function（ICRF） of a single image based on Kernel estimation. The experimental results show that this method in this paper can get more accurate ICRF and higher recognition rate in image identification.

Key words: nonlinearity distribution, Kernel estimation, image identification

摘要： 从数码相机的成像特性出发，详细介绍了通过图像边缘图块色彩的非线性分布来求解其相机响应逆函数的过程，提出了一种基于核密度估计的相机响应逆曲线求解方法。实验证明，该方法能够获得更加准确的相机响应逆函数，用于图像真伪鉴别可以得到较高的识别正确率。

关键词: 非线性分布, 核密度估计, 图像鉴别