ROV(Remotely Operated Vehicles)是水下作業不可或缺的重要工具，其活動係由ROV配備的攝影機，將水底的影像傳到水面支援船隻上，由操作人員下達指令的方式進行。藉由ROV所具備的各種設備，吾人不必實際潛到深海中，就可以從事海底管路檢查、海底纜線、打撈、港埠設施的調查、水下的維修、飛機失事的搜索等作業。而台灣附近海域、湖泊與河川因為受到程度不同的污染，水中含有細微的懸浮物，使得水質混濁的程度相當平均。吾人在這種水域中所看到的物體影像，類似陸地上有濃霧時的狀況，只知道有物體存在，而無法清晰看出其輪廓。透過ROV的攝影機所拍到的相片，被拍物體也往往無法清晰辨識。
　　本研究針對這個問題，首先，利用邊緣檢測方法找出汙染海域中拍攝所得之物體邊緣影像，而後根據彩色影像係由RGB三原色所組成的特性，將其由RGB彩色模型轉換至HSI彩色模型後，以未污染之HSI三值為基準，計算並判別混濁水域的像素成分；將混濁水域的像素成分由實際相片中移除後，重建混濁水域拍到相片的原始影像。然而混濁水域的像素成分，會隨著濁度的不同而有些許差異；因此預定以五個不同濁度的HSI之數值當做基準數據，透過倒傳遞網路處理，將該水域在不同濁度下的混濁像素成分計算出來，作為修正ROV攝影機拍到相片的依據。這項影像處理技術可以移除混濁的干擾，協助人員對於水底下物體的識別能力，對於ROV在混濁水域作業時的物體識別有助益。

ABSTRACT
　　Under the progress of science and technology with rapid change，in order to save amount of time and manpower cost，using automation to replace artificially one after another，and go to the development of the world to the sea，rely on experienced dive personnel early，but because the working danger is high，in order to make it convenient for the fact that avoid causing danger and using at sea，so has developed out the suitable substitutability measuring machine. For example，there is the remote control under water (ROV: Remotely Operated Vehicles).

　　ROV can come to advance handling to design in the ocean，it can utilize in lighting，taking a photograph of observing，the sonar forms images and uses the work，such as mechanical arm homework，etc.,

　　And can supply with the electricity via asking for the cable from the surface of water，convey and receive and control the signal，such as TV image signal for control ROV’s sport . And ROV can be used in sea floor production for oil development, in addition, it usually uses to do something , such as investigation or maintenance under water，search of the aviation accident of checking daylily, the cable on the pipeline of sea floor，salvaging，port facility etc., It can be used to clear away the submarine mine effectively for militarily. There is treatment about the sea floor image，it having lofty ideals for the relevant personnel engaged in research of the respect under water，there is very important practicability. This project is mainly to the thing that when ROV photographs to fetch looking like in the polluting sea area，it can use the colored image characteristic to process the polluting picture，and combine neural network to process it. And it is pollute to when ROV take a photograph of image of sea area trying to make and restore movements under water to come，it can be after this treatment to look forward .And wish that it can be clear that distinguish the image，and hope that it can be contribution to the color image processing under water.

參考文獻
1.A.Adler,R.Guardo,A neural network image reconstruction technique for electrical impedance tomography,IEEE Trans.Med.Imaging,1994
2.Andrews,H.C.Computer Techniques in Image Processing, Acadmic Press, New York,1970
3. Bates,R.H.T., and Mc Donnell,M.J. Image Restoration and Reconstruction. Oxford University Press,New York,1986.
4. Boyd,J.E.and Meloche,J.“Binary Restoration of Thin Object in Multidimensional Imagery,”IEEE Trans.Pattern Anal.Machine intell.,vol.20,no.6,pp.647-651,1998
5.C.M.Bishop,Neural Networks for Pattern Recognition,Oxford University Press,Oxford,1995
6.Geman,D.,and Reyonlds,G.“Constrained Restoration and the Recovery of disxontinuities,”IEEE Trans. Pattern Anal .Machine Intell.,vol.14,no.3,pp.367-383,1982
7.Giannakis,G.B.,and Heath,R.W.,Jr.“Blind Identification of Multichannel FIR Blurs and Perfect Image Restoration,”IEEE Trans.Image Processing,vol.9,no.11,pp.1877-1896,2000
8.Hurn,M.,and Robinson ,A.H.“An extension of Geman and Reynolds’ Approach to Constrained Restoration and the Recovery of discontinuities,”IEEE Trans.Pattern Anal.Machine Intell.,vol.18,no.6,pp.657-662,1996
9.Harikumar,G.,and Bresler,Y.“Perfect Blind Restoration of Image Blurred by Multiple Filters:Theory and Efficient Algorithms,”IEEE Trans.Image Processing,vol.8,no.2,pp.202-219,1999
10.Mesarovic,V.Z.“Iterative Linear Minimum Mean-Sequare-Error Image Restoration from Partially Known Blur,”J.Opt.Soc.Amer-A.Optics, Image Science,and Vision,vol.17,no.4,pp.711-723,2000
11.Malacara,D.Color Vision and Colorimetry:Theory and Applications,SPIE Optical Engineering Press,Bellingham,Wash,2001
12.M.Egmont-petersen,D.de Ridder,H.Handels,“Image Processing with neural networks-a review”,2001
13.R.C Gonzalez, and R.E Woods, Digital Image Processing, Addison-Wesley Publishing Company,1992
14. Rosenfeld,A.,and Kak,A.C.Ditital Picture Processing,vols.1 and 2,2nd ed.,Academic Press,New York.
15.Wolff,RS.,and Yaeger,L,Visualization of Natural Phenomena , Spring-Verlag,New-York,1993
16.Yitzhaky,Y.,Lantzman,A.,and Kopeika,N.S.“Direct Method for Restoration of Motion Blurred Image,”J.Opt.Soc.Amer.-A Optics,Image Science,and Vision,vol.15,no.6,pp.1512-1519,1998
17.王兆璋、劉金源, 無人水下遙控載具之簡介,1999
18.莫顯蕎,陳天毅,談遙控潛航器之應用,1999
19黃將修,施明昌,水下無人載具與雷射影像技術在海洋微生物研究的應用,1999