基于整數(shù)小波變換的無人機(jī)偵查圖像的壓縮

摘 要:基于傳統(tǒng)小波變換過程復(fù)雜的特點(diǎn)和SPIHT編碼算法過程重復(fù)運(yùn)算、存儲(chǔ)量大，并且偵查圖像的目標(biāo)像素小且目標(biāo)數(shù)量大，相關(guān)性低，提出了一種基于差分脈沖編碼調(diào)制和整數(shù)小波變換相結(jié)合的改進(jìn)混合編碼算法。首先以壓縮效果為準(zhǔn)則，以偵查圖像為標(biāo)準(zhǔn)訓(xùn)練圖像，以壓縮比、峰值信噪比為參數(shù)來確定最優(yōu)的小波基。然后以(9，7)整數(shù)提升小波對圖像進(jìn)行分解，對低頻子帶的重要系數(shù)先進(jìn)行DPCM編碼。最后對整個(gè)變換后的子帶進(jìn)行SPIHT編碼。實(shí)驗(yàn)結(jié)果表明該算法在相同的比特率下得到的重構(gòu)圖像的PSNR值高于原算法，而且偵查圖像的低頻近似效果得到了改善。

關(guān)鍵詞:圖像壓縮; 偵查圖像; 整數(shù)提升小波; 分層樹集合分裂算法

中圖分類號(hào):TP919.8 文獻(xiàn)標(biāo)識(shí)碼:A

文章編號(hào):1004-373X(2010)14-0049-04

UAV Reconnaissance Image Compression Based on Integer Wavelet

HUANG Jin-jie，LU Ren-hu

(Automatic College， Harbin University of Science and Technology， Harbin 150080， China)

Abstract:An improved omnibus coder arithmetic based on differential pulse code modulation (DPCM) and IWT is proposed， in view of the problems of complicated convolution process of wavelet transform， repeated calculations and a large number of memories of SPIHT algorithm， as well as the object pixels of spy image are very small and have big quantity， the relativity between the pixels in the frame is little. At first， the compression result is given as the standard， and the spy image as the standard training image， the compression ratio， PSNR and reserved energy percentage after compression are taken as parameters to look for the best wavelet filter. Secondly， the image is decomposed by using(9，7) tap integer lifting wavelet， the DPCM is used to code in the lowest frequency sub-band. Finally， the SPIHT are used to code all the sub-band after transformed. Experiments demonstrate that the PSNR value is higher than that of the original SPIHT at the same bit rate.

Keywords:image compression; spy image; integer lifting wavelet; SPIHT

隨著無人機(jī)偵查技術(shù)的迅速發(fā)展，偵查圖像分辨率的提高，偵查數(shù)據(jù)量日益龐大，數(shù)據(jù)的存儲(chǔ)和傳輸逐步成為一個(gè)迫切需要解決的問題，數(shù)據(jù)壓縮技術(shù)就是解決這個(gè)問題的有效途徑。基于小波變換的嵌入式零樹小波編碼算法EZW、分層樹集合分裂算法SPIHT具有優(yōu)良的壓縮性能，系統(tǒng)實(shí)現(xiàn)難度低，尤其是分層集合分裂算法是當(dāng)前最成熟的幾種編碼算法之一。但是由于偵查圖像的空間冗余較小，容量大，細(xì)節(jié)豐富，采用傳統(tǒng)的壓縮方法存在一定程度的局限性。

本文針對SPIHT算法中存在的不足，結(jié)合偵查圖像的特點(diǎn)，對SPIHT編碼算法進(jìn)行了改進(jìn)，提出了基于差分脈沖編碼調(diào)制(DPCM)和整數(shù)小波變換相結(jié)合的適合偵查圖像的改進(jìn)混合編碼算法。……