论文部分内容阅读
为了改善噪声图像的质量,提出了一种基于KALMAN滤波的降噪方法,该算法采用递推性算法,因此,可以适用平稳与非平稳过程,这就解决了其他估计方法的限制性困难。该方法分析了噪声图像的特征,并且在一价高斯噪声的基础上改写了噪声图像的观测方程,同时,使用NSHP模型来构造图像的过程方程,大大的降低了卡尔曼滤波更新中的计算量。仿真结果表明,卡尔曼滤波方法可以明显的减弱了原始图像上噪声,并且有效的解决了图像滤波必然伴随的模糊细节问题,和其他传统噪声降噪方法比较,更好的保持了原图像中的一些线条,点和边缘的细节信息,体现了自己的自适应优点。
In order to improve the quality of the noise image, a noise reduction method based on KALMAN filter is proposed, which uses recursive algorithm. Therefore, the stationary and non-stationary processes can be applied, which solves the restrictive difficulties of other estimation methods. The method analyzes the characteristics of the noise image and rewrites the observation equation of the noise image based on the monovalent Gaussian noise. At the same time, the NSHP model is used to construct the process equation of the image, which greatly reduces the computational cost of the Kalman filter update . The simulation results show that the Kalman filtering method can significantly reduce the noise on the original image and effectively solve the fuzzy detail problem which is inevitably accompanied by the image filtering. Compared with other traditional noise reduction methods, the Kalman filtering method can better preserve the original image noise Some line, point and edge detail information reflects its own adaptive advantages.