一種基于小波多尺度邊緣檢測(cè)的圖像融合算法
Multiscale Edge Detection Image Fusion Algorithm Using Wavelet Transform
-
摘要: 該文提出了一種新的基于多尺度邊緣檢測(cè)的小波圖像融合方法,是一種利用圖像邊緣特征的小波圖像融合方法,融合過(guò)程利用了圖像的多尺度邊緣信息。為了更好地保持圖像的邊緣,該文在圖像融合過(guò)程中將圖像去噪與邊緣檢測(cè)相結(jié)合。提出了一種物理意義明確的小波最佳分解層數(shù)的確定方法。利用統(tǒng)計(jì)分析的評(píng)判準(zhǔn)則,如熵、標(biāo)準(zhǔn)偏差等,評(píng)價(jià)二維多聚焦圖像不同小波分解層的融合效果,表明該方法提高了圖像的熵和標(biāo)準(zhǔn)偏差的值,算法效果良好。
-
關(guān)鍵詞:
- 圖像融合; 邊緣檢測(cè); 小波變換; 多尺度分析
Abstract: A novel image fusion algorithm using wavelet transform multiscale edges detection is proposed in this paper. This wavelet analysis image fusion method is feature-based. The source images are fused using their multiscale edges information. In order to preserve fusion image edges better, the denoising and edges detection combined are statistical parameters is used to evaluate multifocus image fusion performance. The technique determining optimal decomposition level has explicit physical meaning in this paper. Comparing source images entropy with fusion image entropy at the same decomposition level, if arbitrary source image entropy is larger than the fusion image entropy, it is not neccesary to make next level decomposition. Experimental results demonstrate the effectiveness of the algorithm which increases entropy and standard deviation of fusion image. -
何友,王國(guó)宏,陸大,彭應(yīng)寧.多傳感器信息融合與應(yīng)用.北京:電子工業(yè)出版社,2000,第1章.[2]Hall D L, Llinas J. A introduction to multisensor data fusion[J].Proc. IEEE.1997, 85(1):6-[3]Li H, Manjunath B S, Mitra S K. Multisensor image fusion using the wavelet transform[J].Graphical Models and Image Processing.1995, 57(3):235-[4]Zhang Z, Blum R S. A categorization of multiscaledecomposition -based image fusion schemes with a performance study for a digital camera application[J].Proc. IEEE.1999, 87(8):1315-[5]Mallat S G, Zhong S F. Characterization of signals from multiscale edges. IEEE Trans. on PAMI, 1992, 14(7): 710 - 732.[6]Mallat S G. A Wavelet Tour of Signal Processing. San Diego:Academic Press, 1998: Chapter 6. -
計(jì)量
- 文章訪問(wèn)數(shù): 2761
- HTML全文瀏覽量: 121
- PDF下載量: 932
- 被引次數(shù): 0
下載:
