小波在雷达图像去噪中的应用研究
发布时间：2018-12-18 04:27
【摘要】：随着时代的进步，知识量的扩充，小波被越来越多的学者所重视，并在图像去噪方面得到广泛应用。雷达图像不仅仅运用在军事领域，现如今也被运用在气象、人们外出导航等方面。雷达图像的质量影响人们对信息量准确度的把握。因此本文对小波在雷达图像去噪方面进行研究。 本文首先对雷达图像进行分析，研究雷达图像特有的噪声特点、小波基本理论、小波的特性。运用MATLAB软件模拟雷达含噪声图像，在不含噪声的原始雷达图像上，加入高频噪声，分别为椒盐噪声、随机高频噪声。模拟含有高频噪声的雷达图像，找到小波变换与雷达图像噪声中的关联性，做好小波对雷达图像去噪的前期工作。 依据雷达图像中信息变量的奇异性，在常用的小波阈值公式系数中，引入奇异因子p，进行相关小波阈值公式改进，再利用理论与实际等价公式，运算得出雷达图像阈值。首先将含有随机高斯噪声的雷达图像进行小波一级分解，其次对图像进行提取，运用得到阈值对雷达图像进行小波阈值去噪，最后小波重组得到去除噪声后的雷达图像。同时与中值滤波去噪法进行对比，验证新的阈值公式的合理性、有效性。 另外运用小波相关性去噪，处理雷达图像中高斯噪声。在雷达通过每个周期变换的时间内，，可得到一个帧的最初原始雷达信号图像。每帧之间都具有很强的帧内相关特性，特别是两两相邻的帧间相关特性表现更强。依据相关性强这一特点，再结合小波系数间相关性特征，改进了一种相关性去噪算法。首先将含有高斯椒盐噪声雷达图像进行小波分解，其次对不同层次中的信息量进行小波阈值去噪，最后进行小波重组。其中与传统的Donoho阈值进行对比。并在分解次数中进行了一级小波与二级小波进行的比较。评估其对雷达图像去噪效果，证明小波相关性的合理性、有效性。
[Abstract]:With the progress of the times and the expansion of knowledge wavelet has been paid more and more attention by more and more scholars and has been widely used in image denoising. Radar images are used not only in military, but also in weather and navigation. The quality of radar images affects the accuracy of information. Therefore, this paper studies the radar image denoising based on wavelet transform. In this paper, first of all, we analyze the radar image, study the characteristic of the noise, the basic theory of wavelet and the characteristic of wavelet. The MATLAB software is used to simulate the radar image with noise, and the high frequency noise is added to the original radar image without noise, which are salt and pepper noise and random high frequency noise respectively. The radar image with high frequency noise is simulated and the correlation between wavelet transform and radar image noise is found. According to the singularity of information variables in radar image, the singularity factor p is introduced into the coefficients of the commonly used wavelet threshold formula, and then the correlation wavelet threshold formula is improved, and then the radar image threshold is obtained by using the equivalent formula of theory and practice. Firstly, the radar image with random Gao Si noise is decomposed by wavelet first order, then the image is extracted, then the radar image is de-noised by wavelet threshold. Finally, the radar image after removing noise is obtained by wavelet reorganization. At the same time, it is compared with the median filter denoising method to verify the rationality and validity of the new threshold formula. In addition, wavelet correlation denoising is used to deal with Gao Si noise in radar image. The original radar signal image of a frame can be obtained within the time of radar passing through each period. Each frame has strong intra-frame correlation, especially between two adjacent frames. According to the characteristics of strong correlation and correlation between wavelet coefficients, a correlation denoising algorithm is improved. Firstly, the radar image with Gao Si's salt and pepper noise is decomposed by wavelet, then the information in different levels is de-noised by wavelet threshold, and then the wavelet is reorganized. It is compared with the traditional Donoho threshold. In the decomposition times, the comparison between the first wavelet and the second wavelet is carried out. The denoising effect on radar image is evaluated, and the rationality and validity of wavelet correlation are proved.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
本文编号：2385338
[Abstract]:With the progress of the times and the expansion of knowledge wavelet has been paid more and more attention by more and more scholars and has been widely used in image denoising. Radar images are used not only in military, but also in weather and navigation. The quality of radar images affects the accuracy of information. Therefore, this paper studies the radar image denoising based on wavelet transform. In this paper, first of all, we analyze the radar image, study the characteristic of the noise, the basic theory of wavelet and the characteristic of wavelet. The MATLAB software is used to simulate the radar image with noise, and the high frequency noise is added to the original radar image without noise, which are salt and pepper noise and random high frequency noise respectively. The radar image with high frequency noise is simulated and the correlation between wavelet transform and radar image noise is found. According to the singularity of information variables in radar image, the singularity factor p is introduced into the coefficients of the commonly used wavelet threshold formula, and then the correlation wavelet threshold formula is improved, and then the radar image threshold is obtained by using the equivalent formula of theory and practice. Firstly, the radar image with random Gao Si noise is decomposed by wavelet first order, then the image is extracted, then the radar image is de-noised by wavelet threshold. Finally, the radar image after removing noise is obtained by wavelet reorganization. At the same time, it is compared with the median filter denoising method to verify the rationality and validity of the new threshold formula. In addition, wavelet correlation denoising is used to deal with Gao Si noise in radar image. The original radar signal image of a frame can be obtained within the time of radar passing through each period. Each frame has strong intra-frame correlation, especially between two adjacent frames. According to the characteristics of strong correlation and correlation between wavelet coefficients, a correlation denoising algorithm is improved. Firstly, the radar image with Gao Si's salt and pepper noise is decomposed by wavelet, then the information in different levels is de-noised by wavelet threshold, and then the wavelet is reorganized. It is compared with the traditional Donoho threshold. In the decomposition times, the comparison between the first wavelet and the second wavelet is carried out. The denoising effect on radar image is evaluated, and the rationality and validity of wavelet correlation are proved.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
【引证文献】
相关期刊论文 前1条
1 荣霞;薛伟;朱继超;;一种新的小波阈值函数在图像去噪中的应用[J];电子测量技术;2016年05期
本文编号：2385338
本文链接：https://www.wllwen.com/kejilunwen/wltx/2385338.html
