對比度增強(qiáng)方法總結(jié)

1、、插值自適應(yīng)直方圖均衡化一種霧天圖像低對比度增強(qiáng)的快速算法匹1=1圖像分塊，每塊大小均為mxn，稱為固定分塊；計算每個分塊的均衡化變換函數(shù)；對于圖像中任意一點，以該點為中心，創(chuàng)建大小為mxn的當(dāng)前分塊;根據(jù)面積，計算當(dāng)前分塊范圍內(nèi)每個固定子塊的權(quán)值；對每個變換函數(shù)進(jìn)行加權(quán)平均，計算變換后的灰度值，替代當(dāng)前點（即當(dāng)前分塊中心） 灰度值二、彩色飽和度對比度增強(qiáng)彩色數(shù)字圖像處理R 1-R - mln( R, G, B max( R, G, B1.G =一一一G - mln( R, G, Bmax( R, G, B 一 min( R, G, BB B - mln( R, G, B三、Contrast

2、Enhancement of Images using Human Contrast SensitivityAlgorithm Enhancc(S, I. L. UjInput:Coiirml pararnstpr AInput Image JJnwRT And iipppp bounds L and lOutput: Enhanced Image VBoginI ind P b| & J(p)3 Vp ait mini mas/saddle pointeSort. P into list Pfv rf SmeppA ndPu. 1f,A)r-u-rFind Q = (|& = /(p) Ap

3、 M miniiuas/ssddlc pointsSort Q into List QfIr SvjcepA?tdPush(ir, Q1.r -u-rRjeturn IEndAlgorithm SwepAndPushf!. f, 6)Input:Input image Il.i&l of gray val ues SCuillrul pLuajlLtjler 5Output:Output image 1Boginr = ffor each s 5obtain boolean matrix B 5 Bjj = 1 ilT lij sIdentify 皿 of hillocks H in Bfor

4、 each Hi e IIfind pmnit m J(Pfnnz P- Pmai E 宜占ecih - m訕皿(IJ 一) 10)for each p e FL= Aeuhlookup & p the net scaling factor in the history for p ilT (Sq,叩燦十J 叩 pjj = 8 一辦尸(p)= (1 + A心)* f + 身update 辦=如十 3applvEnd1. ProcessHillocks (I, t , A)輸入：源圖像/、控制變量T、A 輸出：處理后的圖像/1=1創(chuàng)建Hillocks空棧S創(chuàng)建 hillock h用1初始化h內(nèi)的

5、像素S.push(h)while (S != empty)h = S.pop()根據(jù)h的閾值t在h內(nèi)尋找連通區(qū)域for(每個連通區(qū)域c)創(chuàng)建新的hillock h用c初始化h內(nèi)的像素for(h內(nèi)的每個像素)1(P) = (1 + s )*( 1(p) - t) + t其中t為h的閾值，s為hillock最大值threshold (h ) = min( 1 (P p.l 1(pi) = h 局部最小值(或極小值)threshold (h ) = max( threshold (h ), threshold (h) + A)S.push(h)2.增強(qiáng)算法 Enhancer( t , I, L, U)輸入：源圖像/，控制變量T、A量，上下邊界U、L 輸出：增強(qiáng)后的圖像/ TOC o 1-5 h z 1=11=ProcessHillocks(1，t，A )I=U-II=ProcessHillocks(I，t，A )I=U-I返