基于特征点混沌算法的农产品图像数字水印实现

发布时间：2018-04-29 22:06

  本文选题：农产品仿真图像 + 特征点　； 参考：《江苏农业科学》2017年04期

【摘要】：针对当前农产品图像数字水印易受到几何变形等攻击破坏、不可感知性差、鲁棒性不高等问题,提出特征点混沌算法。通过特征窗内部的最小像素差的平方和确定特征点,且过滤掉低对比度侯选极值点;利用特征点邻域像素的梯度方向分布特性,为每个特征点指定方向参数,利用Tent混沌序列对图像进行置乱,通过相关性函数进行水印检测。仿真试验结果显示,本研究算法在提取猕猴桃和葡萄的数字水印过程中,注重特征点信息以及周围的信息,提取的水印效果清晰,且原始图像没有被破坏;在透明性测试指标中本研究算法的信噪比最大,嵌入后具有较好的不可见性;本研究算法的鲁棒性评价指标中最小,噪声攻击测试后本研究算法的相关系数下降速度最慢。
[Abstract]:Aiming at the problems that the digital watermarking of agricultural product image is easily damaged by geometric deformation and other attacks, the inperceptibility is poor, and the robustness is not high, a feature point chaos algorithm is proposed. The feature points are determined by the square sum of the minimum pixel difference inside the feature window, and the low contrast candidate extremum points are filtered out, and the directional parameters are specified for each feature point by using the gradient direction distribution characteristics of the pixels adjacent to the feature points. The Tent chaotic sequence is used to scramble the image and the correlation function is used to detect the watermark. The simulation results show that in the process of extracting the digital watermark of kiwifruit and grape, the algorithm pays attention to the information of feature points and surrounding information, the watermark effect is clear, and the original image has not been destroyed. In the transparency testing index, the SNR of the algorithm is the largest and the invisibility is better after embedding, and the robustness evaluation index of the algorithm is the smallest, and the correlation coefficient of the algorithm is the slowest after the noise attack test.
【作者单位】： 河南牧业经济学院;
【基金】：河南省政府决策研究招标课题(编号:2013B156)
【分类号】：S126;TP309.7
，

本文编号：1821764