數(shù)字圖像處理考試知識(shí)點(diǎn)整理.doc

1. Image（圖像）：（1）Picture, Photograph （圖片）Visual data （可視數(shù)據(jù)）Usually two- or three-dimensional （常為二維或三維）；（2）書本概念：an image may be defined as a two-dimensional function f(x,y),when x and y are spatial（plane） coordinates,and the amplitude of f at any pair of coordinates(x,y)is called the intensity or gray level of the image at that point(一幅圖像可以定義為一個(gè)二維函數(shù)f（x，y）x和y均為空間（平面）坐標(biāo)，在任何一對(duì)空間坐標(biāo)中的幅值f稱為圖像在該點(diǎn)的強(qiáng)度和灰度)。2. digital image（數(shù)字圖像）：（1）An image which is“discretized,”i.e.,defined on a discrete grid(ex. scanner)（一個(gè)被離散化的圖像，即在一個(gè)離散的網(wǎng)格(例如，掃描儀)上定義）；Two-dimensional collection of light intensity values (or gray values)（二維光強(qiáng)度值(或灰度值)的集合）Matrix, two-dimensional function（矩陣，二維函數(shù)）（2）書本概念：When x，y，and the intensity values of f are all finite，discrete quantities（當(dāng)x，y和灰度值f均為有限離散數(shù)值時(shí)，稱為數(shù)字圖像）。3. digital image processing（數(shù)字圖像處理）：process digital image by means of a digital computer（借助于數(shù)字計(jì)算機(jī)來(lái)處理數(shù)字圖像）。4. pixel（像素）：denote the elements of adigital image（描述數(shù)字圖像元素）。5. example of field that use digital image processing（使用數(shù)字圖像處理領(lǐng)域的實(shí)例）見課本P29。6. fundamental steps in digital image processing（數(shù)字圖像處理的基本步驟）：Image Acquisition（圖像獲?。﹊mage enhancement and restoration（圖像增強(qiáng)與恢復(fù)）image segmentation（圖像分割）feature selection（特征選擇；此過程又稱為description（描述）Image Representation（圖像表示）image interpretation（圖像解釋；此過程也可稱為image recognition（圖像識(shí)別）。7. structure of the human eye（人眼結(jié)構(gòu)）（成像原理）見P58。8. The colors that humans perceive in an object are determined by the nature of the light reflected from the object（人體感受一個(gè)物體的顏色有物體反射光的性質(zhì)決定）。9. monochromatic（or achromatic）light（單色光或無(wú)色光）：light that is void of color（沒有顏色的光）；the only attribute of it is its intensity or amount（它唯一的屬性就是強(qiáng)度或大?。?。10. gray level（灰度級(jí)）：denote monochromatic intensity（表示單色光的強(qiáng)度）。11. gray scale（灰度級(jí)/灰階）：the range of measured values of monochromatic light from black to white（從黑到白的單色光的度量值范圍）。12. chromatic（color）light（彩色光）：spans the electromagnetic energy spectrum from approximately 0.43 to 0.79 m（彩色光的能量譜跨越0.43-0.79m的范圍）。13. Three basic quantities to describe the quality of a chromatic light source（三個(gè)基本量來(lái)描述彩色光源的質(zhì)量）：Radiance（輻射）：the total amount of energy that flows from the light source（從光源流出的能量的總量）；Luminance (measured in lumens (lm)（流明）：gives a measure of the amount of energy an observer perceives from a light source（給出觀察者從光源感知到的能量的量度）；Brightness（亮度）：Embodies the achromatic notion of intensity（體現(xiàn)了無(wú)色的強(qiáng)度概念）。14. sampling（采樣）：digitizing the coordinate values（對(duì)坐標(biāo)值進(jìn)行數(shù)字化）。15. quantization（量化）：digitizing the amplitude values（對(duì)幅值數(shù)字化）。16. image (spatial) resolution（圖像（空間）分辨率）：the largest number of discernible line pairs per unit distance（每單位距離可分辨的最大線對(duì)數(shù)量）。17. intensity resolution（灰度分辨率）：the smallest discernible change in intensity level（灰度級(jí)中可分辨的最小變化）。18. some basic relationships between pixels（像素間的基本關(guān)系）見課本P90。應(yīng)掌握P93公式2.5-1、2.5-2、2.5-3。19. spatial domain（空間域）：The aggregate of pixels composing an image（構(gòu)成圖像的像素的聚合）。Spatial processing（空間域操作）：intensity transformation and spatial filter（灰度變換和空間濾波）operate directly on pixels（直接在像素上操作）。特別提示：灰度變換是對(duì)像素進(jìn)行點(diǎn)操作（point processing），而空間濾波是進(jìn)行鄰域操作（neighborhood processing）。20. basic intensity transformation function（基本的灰度轉(zhuǎn)換函數(shù)）：image negative（圖像反轉(zhuǎn)）log transformation（對(duì)數(shù)變換）power-law（gamma）transformation（冪律（伽馬）變換）piecewise-linear transformation function（分段線性變換函數(shù)）。需要知道各個(gè)函數(shù)的作用與效果（對(duì)應(yīng)圖），P129。21. histogram（直方圖）：the histogram of a digital image with intensity levels is in the range 0，L-1 is a discrete function h（rk）=nk，where rk is the kth intensity value and nk is the number of pixels in the image with intens rk（灰度級(jí)范圍為0，L-1的數(shù)字圖像的直方圖是離散函數(shù)h（rk）=nk，其中rk是第k個(gè)灰度值，nk是圖像中灰度值為rk的像素個(gè)數(shù)）。怎么畫：h(D)= number of pixels which have the gray level equal to D（D=0,12p-1）（h（D）=灰度等于D的像素?cái)?shù)）Normalized histogram（歸一化直方圖）： H(D)=h(D)/(MxN) （歸一化直方圖需要做第二步）。均衡化、匹配（PPT）。22. smoothing spatial filter（平滑空間濾波器）：used to noise reduction（降低噪聲）（1）smoothing linear filter（平緩線性濾波）：definition（定義）：the output（response）of a smoothing，linear spatial filter is imply the average of the pixels contained in the neighborhood of the filter mask（平滑線性空間濾波器的輸出（響應(yīng)）是包含在濾波器模板鄰域內(nèi)的像素的簡(jiǎn)單平均值）。Application（應(yīng)用）：（會(huì)模糊細(xì)節(jié)）。（2）order-statistic（nolinear）filter（統(tǒng)計(jì)排序（非線性）濾波器）：definition（定義）：whose response is based on ordering the pixels contained in image area encompassed by the filter，and then replacing the value of the center pixel with the value determined by the ranking result（這種濾波器的響應(yīng)以濾波器周圍的圖像區(qū)域中所含的像素的排序?yàn)榛A(chǔ)，然后使用統(tǒng)計(jì)排序結(jié)果決定的值代替中心像素的值）。Application（應(yīng)用）：median filter（中值濾波）用于處理椒鹽噪聲（salt-and-pepper noise）（以黑白點(diǎn)形式疊加在圖像上）十分有效。23. sharping spatial filter（銳化空間濾波器）：（1）first derivative（一階微分）特征：must be zero in areas of constant intensity（在恒定灰度區(qū)域的微分值為0）；must be nonzero at the onset of an intensity step or ramp（在灰度臺(tái)階或斜坡處微分值不為0）；must be nonzero along ramps（沿著斜坡的微分值不為0）。（2）second derive（二階微分）特征：must be zero in constant area（在恒定區(qū)域微分值為0）；must be nonzero at the onset and end of an intensity step or ramp（在灰度臺(tái)階或斜坡的起點(diǎn)、終點(diǎn)處微分值不為0）；must be zero along ramps of constant slope（沿著斜坡的微分值必須為0）。（銳化圖像三種方法都要會(huì)，P182）24. combining spatial enhancement methods（混合空間增強(qiáng)法）見P19125. frequency domain（頻域）：26. Fourier series（傅里葉級(jí)數(shù)）: any periodical function can be expressed as the sum of sines and cosines multiplied by aappropriate coefficent（任何周期函數(shù)可以表示為乘以適當(dāng)系數(shù)的正弦或余弦的和）。27. Fourier transform（傅里葉變換）：28. DFT：discrete fourier transform（離散傅里葉變換）。（還有計(jì)算，P257）29. FFT：fast fourier transform（快速傅里葉變換）。30. characteristics of the frequency domain（頻率域的特性）：Impossible to make direct associations between specific components of an image and transform（不可能直接關(guān)聯(lián)圖像的特定組件和轉(zhuǎn)換）The slowest varying frequency component (u=v=0) corresponds to the average gray-level of an image（最慢的變化頻率分量（u = v = 0）對(duì)應(yīng)于圖像的平均灰度級(jí)）Move away from the origin of the transform result,the low frequencies correspond to the slowly varying components（遠(yuǎn)離變換結(jié)果的原點(diǎn)，低頻對(duì)應(yīng)于緩慢變化的分量）Move away further,the higher frequencies correspond to faster and faster gray-level changes:Edges,noise（進(jìn)一步移動(dòng)，較高的頻率對(duì)應(yīng)于更快更快的灰度級(jí)變化：邊緣，噪聲）。31. Lowpass filter（低通濾波器）：a filter H（u，v）that attenuates high frequencies while passing low frequencies（濾波器H（u，v），其在通過低頻時(shí)衰減高頻）。32. Highpass filter（高通濾波器）：a filter H（u，v）that attenuates low frequencies while passing high frequencies（濾波器H（u，v），其在通過高頻時(shí)衰減低頻）。33. steps for filtering in the frequency domain（頻域?yàn)V波步驟）：multiply fp(x,y)by（-1）x+y to center its transform（用（-1）x+y 乘以fp(x,y)移到其變換中心）；compute the DFT，F(xiàn)(u,v)，of the image from step1（計(jì)算來(lái)自第一步的圖像DFT）generate a real，symmetric filter function，H(u,v)，of size PQ with center at coordinates (P/2,Q/2),form the product G(u,v)=H(u,v)F(u,v)（生成一個(gè)實(shí)的、對(duì)稱的濾波函數(shù)H(u,v)，其大小PxQ，中心在(P/2,Q/2).用矩陣形成乘積G(u,v)=H(u,v)F(u,v)）。obtain image gp(x,y),where the real part is selected in order to ignore parasitic complex components that we are dealing with padded arrays（得到圖像gp(x,y)，其中，為忽略由于計(jì)算不準(zhǔn)確導(dǎo)致的寄生復(fù)分量，選擇了實(shí)部。Tips：下標(biāo)p指出我們處理的是填充后的陣列）obtain the final processed result，g(x,y)，by extracting the MxN region from the top,left quadrant of gp(x,y)（通過gp(x,y)的左上象限提取MxN區(qū)域，得到最終處理結(jié)果g(x,y)）。34. 三種低通濾波器的特點(diǎn)：ideal lowpass filter（理想低通濾波器）butterworth lowpass filter（巴特沃斯低通濾波器）gaussian lowpasss filter（高斯低通濾波器）35. 三種高通濾波器的特點(diǎn)：ideal highpass filter（理想高通濾波器）butterworth highpass filter（巴特沃斯高通濾波器）gaussian highpasss filter（高斯高通濾波器）36. 同態(tài)濾波（homomorphic filtering）步驟：（P311）37. Image restoration（圖像恢復(fù)）：Improve an image in some predefined sense（在某些預(yù)定義義下改善圖像）。38. Model of the image degradation/restoration process（圖像退化/恢復(fù)過程模型）39. noise model（噪聲模型）（如何獲??？）：40. 怎么處理周期噪聲：41. restoration in the presence of noise onlyspatial