封皮與摘要精

1、黑龍江工程學院本科生畢業(yè)論文本科學生畢業(yè)論文直方圖均衡化算法的研究系部名稱:測繪工程系專業(yè)班級:地理信息系統(tǒng)BW05-30王源超學生姓名:指導教師:張玉娟職 稱:黑龍江工程學院二OO九年六月The Graduati on Thesis for Bachelor's DegreeStudy on Local Area HistogramEqualizationSp ecialty:Geogra phic Information SystemClass:BW05-30Name of student:WangYuanChaoSupervisor: Lecturer.ZhangYuJuanHe

2、ilo ngjia ng In stitute of Tech no logy2009-06 - Harbin本文主要研究了局部直方圖均衡化（LAHE）及其修正算法在塵肺X線圖像增強 中的應(yīng)用。塵肺病是我國最主要的一種職業(yè)病。目前，醫(yī)學界尚無將其徹底治愈的辦法。塵肺X線檢查是確定塵肺和分期的主要診斷方法。然而，一方面塵肺X借助于數(shù)字圖像處從而有助于降低線圖像具有動態(tài)范圍寬、細節(jié)豐富和對比度差的特點：另一方面塵肺早期病變特 征并不是十分明顯。這些因素給塵肺的早期診斷帶來了難度。 理技術(shù)，對塵肺X線圖像進行對比度增強處理，突出塵肺病灶， 塵肺早期診斷的漏診率。它能夠在很大程度直方圖均衡化（HE）是

3、一種最常用的圖像對比度增強技術(shù)，上增強圖像的視覺效果，幫助人們識別圖像中隱藏的信息。當我們感興趣的信息 包含在圖像中相對較小且灰度分布比較均勻的區(qū)域時，HE并不能給出令人滿意的結(jié)果。LAHE寸HE故了一些修正，取得了比較好的結(jié)果。但是 LAH只有一個控制參 數(shù)：相關(guān)區(qū)域（矩形窗）的大小W這顯然還不夠靈活。小窗口在增強圖像中細節(jié)的 同時削弱了圖像的整體輪廓，而大窗口能使圖像的整體輪廓更加清晰， 但卻模糊 了圖像中的細節(jié)。論文結(jié)合塵肺X線圖像的特點對LAH進行了兩方面的修正。一 方面，仍然保持相關(guān)區(qū)域為大小固定的矩形窗， 通過改變相關(guān)區(qū)域內(nèi)直方圖的構(gòu) 造方法來解決這個問題。另一方面， 保持局部直方

4、圖的構(gòu)造方法不變，通過自適 應(yīng)的改變相關(guān)區(qū)域的大小來解決這個問題。最后我們給出采用LAH及其修正算法 對塵肺X線圖像進行對比度增強處理的結(jié)果，并比較了各種算法的優(yōu)劣。關(guān)鍵詞：圖像增強，直方圖，局部直方圖均衡化，塵肺ABSTRACTThis paper focuses on the study of Local Area Histogram Equalizati on( LANE) and its app licati ons in pn eumoc oni osis X-ray films enhan ceme nts.Pn eumoc oni osis is one of the major

5、 occ up ati onal diseases of which no cures have turned out effective so far. X-ray film exam in ati on is n ecessarily adop ted to diag nose and discer n pneumoconiosis. However, X-ray film features an excessively broad range, overexhaustive details and poor con trast. Also, it is not obvious as to

6、 the p athological cha nges of early pn eumoc oni osis symp toms.These factors add difficulty to the early diag no ses o f pn eumoc oni osis.With the help of digital image p rocess in g,the con trast of pneumoconiosis X-ray film can be enhanced,andsuch early symptoms will be emp hasized with better

7、clarity in the fe edback.Thereby,we are hoping to reduce the misdiag no sis rate.Histogram Equalization(HE) as an image contrast enhancement technique iswidely used. By improving the visual effect of images considerably, it assists the viewer in discovering hidden messages.H owever,satisfaction is h

8、ard to be achieved by HE when we are interested in the information presented in small er areas with relatively grayscales.Local Area Histogram,as an extension based on standard HE technique,has given pleasing rsults.However,LAHE shows a lac k of flexibility because there is only one parameter-size o

9、f the rectangle windo w. Small win dows help in crease the detail clarity of image,but weake n the ove rall outline of the image.Big windows,on the other hand,give better outline by sacrifici ng the detail clarity .In this paper, the author tries to improve LAHE fr om two aspects. On the one hand,we can modify the structure of the local hist ogram;on the other hand,we can also adjust the size of rectangle window adapt ively. At last, the results using LANE and its modifications will be given and com