现代电子技术
現代電子技術
현대전자기술
MODERN ELECTRONICS TECHNIQUE
2013年
14期
93-95
,共3页
图像去噪%小波阀值萎缩法%混合模型%中值滤波
圖像去譟%小波閥值萎縮法%混閤模型%中值濾波
도상거조%소파벌치위축법%혼합모형%중치려파
image denoising%wavelet threshold atrophy method%mixture model%median filtering
图像去噪问题是一个古老的难题,也是当前研究的热点问题,而图像小波去噪算法在图像去噪方面虽然已取得了一定进步,但在这一领域仍然有许多问题需要研究,为了进一步提高图像去噪质量,改善图像视觉效果。在此通过在小波阀值萎缩法、基于混合模型的小波去噪法、小波去噪与其他算法相结合的三类方法中分别选用了三种典型算法即VisuSh?rink法、基于高斯混合模型小波去噪法、中值滤波与小波去噪相结合的算法,对当前基于小波变换图像去噪这三类典型问题进行了研究。研究表明对于单一的噪声,用相应某种算法,就可能取得较理想效果。而对于混合噪声,单独的一种算法取得的效果是比较差的,只有采用几种算法相结合才能取得较好的效果,因而在此也为图像去噪指明了以后的研究方向。
圖像去譟問題是一箇古老的難題,也是噹前研究的熱點問題,而圖像小波去譟算法在圖像去譟方麵雖然已取得瞭一定進步,但在這一領域仍然有許多問題需要研究,為瞭進一步提高圖像去譟質量,改善圖像視覺效果。在此通過在小波閥值萎縮法、基于混閤模型的小波去譟法、小波去譟與其他算法相結閤的三類方法中分彆選用瞭三種典型算法即VisuSh?rink法、基于高斯混閤模型小波去譟法、中值濾波與小波去譟相結閤的算法,對噹前基于小波變換圖像去譟這三類典型問題進行瞭研究。研究錶明對于單一的譟聲,用相應某種算法,就可能取得較理想效果。而對于混閤譟聲,單獨的一種算法取得的效果是比較差的,隻有採用幾種算法相結閤纔能取得較好的效果,因而在此也為圖像去譟指明瞭以後的研究方嚮。
도상거조문제시일개고로적난제,야시당전연구적열점문제,이도상소파거조산법재도상거조방면수연이취득료일정진보,단재저일영역잉연유허다문제수요연구,위료진일보제고도상거조질량,개선도상시각효과。재차통과재소파벌치위축법、기우혼합모형적소파거조법、소파거조여기타산법상결합적삼류방법중분별선용료삼충전형산법즉VisuSh?rink법、기우고사혼합모형소파거조법、중치려파여소파거조상결합적산법,대당전기우소파변환도상거조저삼류전형문제진행료연구。연구표명대우단일적조성,용상응모충산법,취가능취득교이상효과。이대우혼합조성,단독적일충산법취득적효과시비교차적,지유채용궤충산법상결합재능취득교호적효과,인이재차야위도상거조지명료이후적연구방향。
Image denoising is an old problem,but a hot topic of current research. The image wavelet denoising algorithm has achieved some progress,but many problems still need to be solved in this area. In order to further improve the denoising quality and improve the image visual effects,three typical algorithms(VisuShrink algorithm,wavelet denoising method based on Gauss mixture model,and algorithm combining median filtering with wavelet denoising)are selected respectively from three methods:wavelet threshold shrinkage denoising method,wavelet denoising method based on mixture model,and combination of wavelet denoising algorithm and other algorithms. The current typical three types of image denoising methods based on wavelet transform are studied. The study result shows that,for a single noise,the corresponding some algorithm is possible to achieve ideal effect;for the mixed noise,only by combining several algorithms,can better results be achieved,since a single algorithm’ s effect is relatively poor. Therefore,the direction of the future development of image denoising research is pointed out in this pa?per.