【摘要】：隨著時代的進步，知識量的擴充，小波被越來越多的學者所重視，并在圖像去噪方面得到廣泛應(yīng)用。雷達圖像不僅僅運用在軍事領(lǐng)域，現(xiàn)如今也被運用在氣象、人們外出導(dǎo)航等方面。雷達圖像的質(zhì)量影響人們對信息量準確度的把握。因此本文對小波在雷達圖像去噪方面進行研究。 本文首先對雷達圖像進行分析，研究雷達圖像特有的噪聲特點、小波基本理論、小波的特性。運用MATLAB軟件模擬雷達含噪聲圖像，在不含噪聲的原始雷達圖像上，加入高頻噪聲，分別為椒鹽噪聲、隨機高頻噪聲。模擬含有高頻噪聲的雷達圖像，找到小波變換與雷達圖像噪聲中的關(guān)聯(lián)性，做好小波對雷達圖像去噪的前期工作。 依據(jù)雷達圖像中信息變量的奇異性，在常用的小波閾值公式系數(shù)中，引入奇異因子p，進行相關(guān)小波閾值公式改進，再利用理論與實際等價公式，運算得出雷達圖像閾值。首先將含有隨機高斯噪聲的雷達圖像進行小波一級分解，其次對圖像進行提取，運用得到閾值對雷達圖像進行小波閾值去噪，最后小波重組得到去除噪聲后的雷達圖像。同時與中值濾波去噪法進行對比，驗證新的閾值公式的合理性、有效性。 另外運用小波相關(guān)性去噪，處理雷達圖像中高斯噪聲。在雷達通過每個周期變換的時間內(nèi)，，可得到一個幀的最初原始雷達信號圖像。每幀之間都具有很強的幀內(nèi)相關(guān)特性，特別是兩兩相鄰的幀間相關(guān)特性表現(xiàn)更強。依據(jù)相關(guān)性強這一特點，再結(jié)合小波系數(shù)間相關(guān)性特征，改進了一種相關(guān)性去噪算法。首先將含有高斯椒鹽噪聲雷達圖像進行小波分解，其次對不同層次中的信息量進行小波閾值去噪，最后進行小波重組。其中與傳統(tǒng)的Donoho閾值進行對比。并在分解次數(shù)中進行了一級小波與二級小波進行的比較。評估其對雷達圖像去噪效果，證明小波相關(guān)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.
【學位授予單位】：哈爾濱理工大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TN957.52

【引證文獻】

相關(guān)期刊論文 前1條

1 榮霞;薛偉;朱繼超;;一種新的小波閾值函數(shù)在圖像去噪中的應(yīng)用[J];電子測量技術(shù);2016年05期

本文編號：2385338