200521478 玖、發明說明： 【發明所屬之技術領域】 本發明係有關一種攜帶式頻譜取像系統，尤指一種以 一數位攝影器材前加置一多重影像稜鏡及一濾鏡，而以單 次拍攝擷取物體影像，即可獲取該物體影像的頻譜資料。 【先前技術】 按，數位影像的技術隨著近期光學影像或電子產品大 廠的重大技術突破，而使有關於數位影像之光電產品曰漸 融入我們的生活中，該些產品包括有顯示器、數位相機、 數位攝影機、掃描器…等眾多產品，而應用該些光電產品 及其配合之影像處理軟體後，亦使我們的生活上產生急遽 的轉變，如電影數位動畫或炫麗使人目不暇給的畫面呈 現，一再地告訴我們數位化的時代已經來臨了， 一般的物體在不同的燈光下所呈現的顏色不盡全然相 同，習知技藝於擷取其影像資訊時，多半只取其R(Red)、G (Green)、B (Blue)三種顏色的組合為其資訊，該取樣 顏色資訊太少，無法真實呈現物體原本之顏色，因此物體 的顏色呈現常發生「同色異譜」的現象，且其僅可做為一 景夕像初步參考數據，應用於影像測量（如檢查人體皮膚的 顏色來判斷是否患皮膚癌）或高精確度色彩儲存（如博物 館古物數位典藏）時，該三種顏色資訊就嫌不足。因此， 提供較為精確的物體色彩頻譜就顯得相當地重要，亦為本 發明所欲解決的問題。 200521478 請參閱圖一所示，係為一習知頻譜擷取系統之結構 圖，其係利用一數位相機11配合一由複數頻譜波長濾光片 121所構成之轉盤12來擷取一影像，該轉盤上12之複數頻譜 波長濾、光片121的頻譜係為可見光的頻譜 (400nm〜70〇nm)，以700nm頻譜的波長最長；4〇〇nm頻譜 的波長最紐。拍攝時，配合一燈光17的照明該被拍攝物體 13，轉盤調整一頻譜波長濾光片121對準該數位相機丨之鏡 頭，以擷取一頻譜影像，欲取得另一頻譜影像時，必須再 調整轉盤12至另一種頻譜波長濾光片121對準數位相機n 之鏡頭，再擷取另一頻譜影像，以此類推，但該複數頻譜 影像要完整擷取時，必須連續拍攝七次影像方可完成，其 拍攝時間相對地變長，該複數影像於被擷取後，必須經i 灰階景>像14之處理，再經由一色彩矩陣15之運算，便可製 作出:還原影像16,而該轉盤12之控制調整乃由一電腦（圖 ，未示）來完成，故該頻譜擷取系統所利用設備太多且太 笨重，不利於攜行，此即為該習知頻譜影像娜系統之最 主要缺失 明再參閱美國專利編號US5900942號發明專利，其係為 了種多光譜影像系統（Multi Spectral Imaging System )，其 係1在7數位攝糾材巾湘分光結構，級拍攝的影像 ϋ日!由三ΐ電荷耦合器（ccd)做-擷取，便可達到拍攝 譜波段資訊，但其特殊的數位攝影器材 構較為複雜，故其製造成本相當地高且製作相當不 ’再者’同時利用三個電荷輕合|| (CCD)來擷取影像 200521478 時·’欲同時將三組曝光參數調整至可之範圍亦不易達 成；且該特殊數位攝影器材之組成後的體積亦較—般 攝影器材魔大，而不易於做一攜行。 【發明内容】 基於解決以上所述習知技藝的缺失，本發明為一種攜 帶式頻譜取像系統，其主要的目的為數位攝影器材前 加置-多重影像稜鏡及-親，而以單次拍攝擷取物體影 像，即可獲取祕贿⑽所有_，並可影像上精 確的色彩重現度，且頻譜影像進而可_色彩管理系統處 理頻譜資訊到裝置設備相騎色彩錢，皆可獲得良好之 運用，且本發明結構之體積較小，易於做隨身攜行，利於 移動至任一地點做物體之拍攝，具有極佳之機動性。 為達到上述之目的，本發明所利用構件係包括有下列·· 一數位攝景》器材，具有鏡頭及儲存影像之裝置，用以擷 取欲做頻譜分析之影像； ° 一多重影像稜鏡，設置於該數位攝影器材之鏡頭前，可 將該被擷取影像分割成若干個；以及 一濾鏡’由若干個不同頻譜波長之濾光片所組成，置於 多重影像棱鏡及數位攝影器材之間，其係對應多重影 像棱鏡所分割若干個影像，並使該若干個影像各別對 應一不同頻譜波長濾光片。 為進一步對本發明有更深入的說明，乃藉由以下圖 示、圖號說明及發明詳細說明，冀能對貴審查委員於審 200521478 查工作有所助益。 【實施方式】 以下以文字配合圖式說明本發明為達成目的所使用的 技術手段與功效，而以下圖式所列舉之實施例僅為輔助說 明’以利貴審委瞭解’但本案之技術手段並不限於所列舉 圖式。 請參閱圖二所示，係為本發明之攜帶式頻譜系統之實 施結構圖，其係包括有下列之構件： 一數位相機2，亦可為一數位攝影機，具有鏡頭及儲存影 像之裝置，用以擷取欲做頻譜分析之影像； 一多重影像棱鏡3，設置於該數位相機2之鏡頭前，可將 該被擷取影像分割成若干個；以及 一濾鏡4’由若干個不同頻譜波長之濾光片4&所組成，置 於多重影像稜鏡3及數位相機2之間，其係對應多重影 像稜鏡3所分割若干個影像（本實施例為五個），並使 該若干個影像各別對應一不同頻譜波長濾光片4a，濾 鏡4之若干濾光片的光譜落於4〇〇nm至700nm之間，且 該若干濾光片4a之顏色係包含有紅、綠、藍、黃、紫 等之顏色。 藉由上述之構成，該稜鏡3係為一多重影像之五稜鏡或 六稜鏡，且該濾鏡4的裁切與稜鏡3形狀相同。 請參閱圖三所示，係為本發明之被擷取影像透過多重 影像稜鏡及濾鏡後獲得五個頻率影像之示意圖，其中該物 200521478 體5之影像透過了圖二所揭示的結構（數位相機2、多重影 像棱鏡3、濾'鏡4)，本實施例中藉由該可切割成五個影像 之稜鏡3配合遽鏡4之五色頻譜，於一次拍攝時取得五張影 像以進行影像頻譜之分析。 β月參閱圖四所不’係為本發明之遽鏡可藉由兩種頻譜 排列順序獲得複_譜雜之示_，其巾域光度/波長 ，例之線性圖中，該波長摘取波段於至7⑻之可見 光範圍中，且依感光度的關係可分為A、B、c、D、E…等 五種代號’械表五贼光度細，而每個舰之光譜波 長長度皆為lGGnm ’且其线㈣隔g5()nm，目此該五種 感光度的湖順序可分為兩種，—為循序排列法，其係為 將A、B、C、D、e之五種感光度直接依其順序排列；二為 間隔排列法’其係將未有重覆波長之A、c、E先行排入， #將丑、D排入’此種排列法於操取頻譜時，影像交錯部份 會較為明顯，故有胁職之分割，但無論㈣那-種方 法’皆可完成複數頻譜影像。 請參閱圖五所示，係為本發簡由—色製作以便 於切割影像且麟有效影像之示意圖，其侧_鏡4濾光 的特性，以製作一色稿6，該色稿6係以黑色為底色，並於 二上均勻分布紅、綠、藍、黃、白...等之色線，以組成前 述顏色之_’進行分鄕像有效部份，經由數位相 機2搭配稜鏡3及濾、鏡4後，便形成—被均分成五色之色稿 61 :且其被均分成五色之色稿61内，更係各別包括有一擷 取區域62，利用該被均分成五色之色稿61之製作，便可使 200521478 一物體5經由數位相機2搭配稜鏡3及濾鏡4拍攝後，可順利 完成影像之切割，並從複數頻譜影像41中分別找到複數有 效之影像（包括：第一有效影像411、第二有效影像412、 第二有效影像413、第四有效影像414、第五有效影像415)， 由於滤鏡4具有過濾光譜功能，因此妹過濾的顏色呈現黑色 線條，可穿透顏色線條較其他不可穿透線條顏色為明亮， 依此特性可輕易分割多重影像有效部分，以利於頻譜分析。 請參閱圖六所示，係為本發明之多頻譜影像的擷取至 還原影像之圖示流程示意圖，其中一數位相機2搭配稜鏡3 及濾鏡4來擷取複數影像，該具有切割成複數頻譜影像之濾 鏡41的顏色係為可見光的顏色，以紅色的波長最長；紫色 的波長最短。拍攝時，配合一燈光71的照明該被拍攝物體 5，以獲得一複數頻譜影像41，並從複數頻譜影像41中分別 找到複數有效之影像（包括：第一有效影像411、第二有效 影像412、第三有效影像413、第四有效影像414、第五有效 衫像415)，而該複數有效影像經由一色彩矩陣72之運算， 便可付還原衫像73，且該還原影像73與被拍攝物體5相 較，具有極高的還原度。 請參閱圖七所示，係為本發明之多頻譜影像的擷取至 還原景>像之較詳細流程圖，其係包括色彩校正及頻譜取像 及還原影像的實施方法，其係分別敘述如下·· 曰 (I) 一種攜帶式頻譜取像的校正方法，其係包括有·· (al)將反射頻譜81做一反射頻譜測量82，再利用主成 份分析法（PCA) 83以求得一特徵向量84 ; 200521478 (i 射頻献前—驟步所求得的特徵向量84做 、ί工間分析85 ’以求得—特徵值％參數； i =數軸機之_雜配—乡重聽棱鏡及- f鏡87 ’用以拍攝—影像而獲得-組強度值88參 數；以及 U4)將前賴特徵娜與強度細做為參數，並實施 〜線f生回歸89分析，以獲得_轉換矩陣94參數。 ㈣骑納α ’找錄仙舰過稜鏡及濾 鏡拍攝到的賊值’無體反射輯_換鱗。在求得 轉換矩陣前必須取得—組校正數據，選正目標後而 取得頻譜與數位相機·值；色塊頻譜資料取得方法為利 用-頻譜儀’取得校正色塊頻譜後，再使用一主成份分析 法（Principal Component Analysis ’ PCA)，以求得特徽向 量與特徵值。轉色塊_祕後，細细多變 回歸最小平方法树紐械_度值娜賊的轉換矩 陣。 、 (Π) —種攜帶式頻譜取像及還原影像的實施方法，其係包 括有： (M)以一數位相機之鏡頭前搭配一多重影像稜鏡及一 濾鏡92拍攝一物體5以獲得若干影像頻譜； (b2)藉由一色稿6的製作並與前述若干影像頻譜搭配， 以便於切割该複數影像頻譜且獲得有效影像； (b3)將前述有效影像經由一強度值93、換轉矩陣94及 特徵值95等參數的運算，並輔以特徵空間合成％一 200521478 影像；以及 (b4)將猶步麟獲得之縣加人反咖譜97的參數 後，即可還原該影像的真實狀態。 以上步驟（b4)之反射頻譜為一物體加上一光源之反 射頻譜，必須將其除上光源頻譜後，即可求得一物體之反 射率頻譜。而求得物體之反射率頻譜後，只需再取得各種 光源頻譜，即可模擬物體於各種不同光源下 反 藉由上述圖-至圖七電路之動作流程，可明確瞭^本 發明乃藉由-數位攝影糾前加置—多重影像稜鏡及一遽 鏡，而以單次拍攝擷取物體影像，即可獲取該物體影像的 所有㈣’並可提供輯上精確的色㈣現度且頻譜影 η可利用色彩管理系統處理頻譜資訊到裝置設備相關 s號’皆可獲得良好之運用，且本發縣構之體積 ^小，易於做隨身贿，利於移動至任—地賴物體之拍 Ξ之ΐί極佳之機動性’故提出專利案之申請以尋獲專利 ❿ _、”不上所述’本發明之結構特徵及各實施例皆已詳細揭 ”而可充刀顯示出本發明案在目的及功效上均深 =生，=產業之利用價值’且為目前市面上前所未 專利2件鱗槪之精神所述，本發_完全符合發明 唯以上所述者，僅為本發明之較佳實施例而已，當不 Ζίΐϊ本發騎實施之細，即大凡依本發明申請專 巳乍之均等變化與修飾，皆應仍屬於本發明專利涵 12 200521478 蓋之範圍内，謹請 貴審查委員明鑑，並祈惠准，是所 至禱。 【圖式簡單說明】 圖一係為一習知頻譜擷取系統之實施結構圖。 圖二係為本發明之攜帶式頻譜系統之實施結構圖。 圖三係為本發明之被擷取影像透過多重影像稜鏡及濾鏡後 獲得五個頻率影像之示意圖。 圖四係為本發明之濾鏡可藉由兩種頻譜排列順序獲得複數 頻譜影像之示意圖。 圖五係為本發明藉由一色稿的製作以便於切割影像且獲得 有效影像之示意圖。 圖六係為本發明之多頻譜影像的擷取至還原影像之圖示流 程不意圖。 圖七係為本發明之多頻譜影像的擷取至還原影像之較詳細 流程圖。 圖號說明： 11〜數位相機 12〜轉盤 121〜濾光片 13〜物體 14〜灰階影像 15〜色彩矩陣 200521478 16〜還原影像 17〜燈光 2〜數位相機 3〜棱鏡 4〜濾鏡 4a〜濾光片 41〜複數頻譜影像 411〜第一有效影像 412〜第二有效影像 413〜第三有效影像 414〜第四有效影像 415〜第五有效影像 42〜循序排列法 43〜間隔排列法 5〜物體 6〜色稿 61〜被均分五色之色稿 62〜擷取區域 71〜燈光 72〜色彩矩陣 73〜還原影像 81〜反射頻譜 82〜反射頻譜測量 83〜主成份分析法 200521478 84〜特徵向量 85〜特徵空間分析 86〜特徵值 87〜數位相機搭配稜鏡與濾鏡 88〜強度值 89〜線性回歸 91〜影像分割技術 92〜數位相機搭配稜鏡與濾鏡 93〜強度值 94〜轉換矩陣 95〜特徵值 96〜特徵空間合成 97〜反射頻譜200521478 发明 Description of the invention: [Technical field to which the invention belongs] The present invention relates to a portable spectrum acquisition system, and more particularly to a multi-camera and a filter placed in front of a digital photographic equipment, and The second shot captures the image of the object to obtain the spectrum data of the object image. [Previous technology] According to the recent digital technology breakthroughs in optical imaging or electronics manufacturers, digital imaging technology has gradually integrated into our lives optoelectronic products. These products include displays, digital Cameras, digital cameras, scanners, and many other products, and the use of these optoelectronic products and their associated image processing software, has also made a radical change in our lives, such as movie digital animation or dazzling dizzying pictures Presentation has repeatedly told us that the age of digitalization has come. The colors of ordinary objects under different lights are not all the same. When acquiring the image information, most of them only take R (Red). The combination of three colors, G (Green) and B (Blue) is its information. The sampled color information is too little to truly represent the original color of the object. Therefore, the phenomenon of "metamorphism" often occurs in the color of objects, and its It can only be used as preliminary reference data for a scene and image, and applied to image measurement (such as checking the color of human skin to determine whether it is affected by skin) Time) or a high degree of color accuracy store (such as the Museum of Antiquities Digital Collections), the three colors of information on suspected inadequate. Therefore, it is very important to provide a more accurate color spectrum of objects, which is also a problem to be solved by the present invention. 200521478 Please refer to Figure 1. This is a structural diagram of a conventional spectrum acquisition system. It uses a digital camera 11 and a turntable 12 composed of a complex spectrum wavelength filter 121 to capture an image. The complex spectrum wavelength filter of 12 on the turntable, and the spectrum of the light sheet 121 are the spectrum of visible light (400nm ~ 70nm), with the wavelength of 700nm being the longest; the wavelength of 400nm is the newest. When shooting, with a light 17 to illuminate the object 13, the turntable adjusts a spectral wavelength filter 121 to point at the lens of the digital camera 丨 to capture a spectrum image. To obtain another spectrum image, you must Adjust the turntable 12 to another spectral wavelength filter 121 to align the lens of the digital camera n, and then capture another spectrum image, and so on. However, if the complex spectrum image is to be completely captured, the image must be taken seven times in a row. It can be completed, and its shooting time is relatively long. After being captured, the complex image must be processed by i grayscale scene> image 14 and then operated by a color matrix 15 to produce: restored image 16 And the control adjustment of the turntable 12 is completed by a computer (picture, not shown), so the spectrum acquisition system uses too much equipment and is too bulky to carry, which is the conventional spectrum image. The main missing part of the system is referred to the US patent No. US5900942. It is for the Multi Spectral Imaging System. The captured image is the next day! It can be captured by the triple-coupled charge coupled device (ccd), which can reach the spectral band information, but its special digital photography equipment is more complicated, so its manufacturing cost is quite high and its production is quite 'Further' when capturing images with three charges simultaneously || (CCD) 200521478 · 'It is not easy to adjust the three groups of exposure parameters to a range that is possible at the same time; and the composition of this special digital photographic equipment The volume is also larger than that of ordinary photography equipment, which is not easy to carry around. [Summary of the Invention] Based on solving the above-mentioned shortcomings in the conventional arts, the present invention is a portable spectrum acquisition system. The main purpose of the invention is to add multiple images in front of digital photographic equipment and pro- Capture and capture the image of the object, you can get all the secrets, and the accurate color reproduction on the image, and the spectrum image can further _ color management system process the spectrum information to the device and equipment to ride the color money, and you can get good The structure of the invention is small, easy to carry around, and it is convenient to move to any place to shoot objects, and has excellent mobility. In order to achieve the above-mentioned object, the components used in the present invention include the following: a digital photography device, with a lens and a device for storing images, for capturing images to be subjected to spectrum analysis; ° a multiple image 稜鏡, Placed in front of the lens of the digital photographic equipment, the captured image can be divided into several; and a filter 'composed of several filters with different spectrum wavelengths, placed on multiple image prisms and digital photographic equipment In between, it corresponds to a plurality of images divided by a multiple image prism, and each of the plurality of images corresponds to a different spectral wavelength filter. In order to further explain the present invention in depth, the following illustrations, drawing number descriptions, and detailed descriptions of the invention are intended to help your review committee in the investigation of 200521478. [Embodiment] The following text and drawings illustrate the technical means and effects used by the present invention to achieve the purpose, and the examples listed in the following figures are only for the purpose of explaining 'for the benefit of the review committee's understanding.' It is not limited to the listed drawings. Please refer to FIG. 2, which is a structural diagram of the implementation of the portable spectrum system of the present invention, which includes the following components: a digital camera 2 or a digital camera, which has a lens and a device for storing images. To capture the image for spectrum analysis; a multiple image prism 3, which is set in front of the lens of the digital camera 2, can divide the captured image into several; and a filter 4 'consists of several different spectrums The wavelength filter 4 & is composed between the multiple images 稜鏡 3 and the digital camera 2, which corresponds to the multiple images (five in this embodiment) divided by the multiple images 稜鏡 3, and makes the several Each image corresponds to a different spectral wavelength filter 4a. The spectrum of several filters of filter 4 falls between 400nm and 700nm, and the color of the plurality of filters 4a includes red and green. , Blue, yellow, purple and so on. With the above structure, the 稜鏡 3 is a 稜鏡 or 稜鏡 of a multiple image, and the cut of the filter 4 is the same as that of the 稜鏡 3. Please refer to FIG. 3, which is a schematic diagram of five frequency images obtained by the captured image of the present invention through multiple image filters and filters. The image of the object 200521478 and the body 5 passes through the structure disclosed in FIG. 2 ( Digital camera 2, multi-image prism 3, filter 4), in this embodiment, five images of 稜鏡 3 which can be cut into five images are combined with the five-color spectrum of 遽 4 to obtain five images for one shot Analysis of image spectrum. β Refer to Figure 4 for details. The mirror of the present invention can obtain complex _ spectrum miscellaneous_ through two kinds of spectrum arrangement order, its luminosity / wavelength. For example, in the linear diagram, the wavelength is extracted. In the visible light range up to 7 °, and according to the relationship of sensitivity, it can be divided into five code names, such as A, B, c, D, E, and so on. The mechanical watch has a fine luminosity, and the spectral wavelength length of each ship is lGGnm 'And its line is separated by g5 () nm, so the lake order of the five sensitivity can be divided into two kinds, which is a sequential arrangement method, which is the five kinds of photosensitive A, B, C, D, e The degrees are directly arranged in their order; the second is the interval arrangement method, which includes A, c, and E that do not have repeated wavelengths first, and #the ugliness and D are included in this arrangement method. The interlaced part will be more obvious, so there is a threatening division, but no matter which way-you can complete the complex spectrum image. Please refer to Figure 5, which is a schematic diagram of the color image made by this color in order to cut the image and the effective image. The characteristics of the filter of the side_mirror 4 are used to make a color draft 6, which is black. Is the background color, and the red, green, blue, yellow, white, etc. color lines are evenly distributed on the two, and the effective part of the image is divided into _ 'which constitutes the aforementioned color. After the filter and mirror 4 are formed, they are evenly divided into five-colored color manuscripts 61: and they are evenly divided into five-colored manuscripts 61, and each includes a capture area 62, which is used to evenly divide the five-colored manuscripts. The production of manuscript 61 can make 200521478 an object 5 shot through digital camera 2 with 稜鏡 3 and filter 4, and can successfully cut the image, and find multiple valid images from the complex spectrum image 41 (including: The first effective image 411, the second effective image 412, the second effective image 413, the fourth effective image 414, and the fifth effective image 415). Because the filter 4 has a spectral filtering function, the color filtered by the girl appears as a black line. Penetrating color lines are more impenetrable than others The color of the bars is bright. Based on this feature, the effective parts of multiple images can be easily divided to facilitate spectrum analysis. Please refer to FIG. 6, which is a schematic flow chart of the process of capturing multi-spectral image to restore image according to the present invention. Among them, a digital camera 2 with 稜鏡 3 and filter 4 is used to capture complex images. The color of the filter 41 of the complex-spectrum image is the color of visible light, with red having the longest wavelength and purple having the shortest wavelength. When shooting, the subject 5 is illuminated with a light 71 to obtain a complex spectrum image 41, and a plurality of valid images (including a first valid image 411 and a second valid image 412) are respectively found from the complex spectrum image 41. , The third effective image 413, the fourth effective image 414, and the fifth effective shirt image 415), and the plurality of effective images can be restored to the shirt image 73 through the operation of a color matrix 72, and the restored image 73 and the captured image Compared with the object 5, it has a very high degree of reduction. Please refer to FIG. 7, which is a more detailed flowchart of the multi-spectral image capture to restore scene > image according to the present invention, which includes the implementation methods of color correction, spectrum capture and restored image, which are described separately As follows: (I) A correction method for portable spectrum acquisition, which includes: (al) Reflecting spectrum 81 as a reflection spectrum measurement 82, and then using principal component analysis (PCA) 83 to obtain A feature vector 84; 200521478 (i The feature vector 84 obtained before the radio frequency is presented in the step, and the analysis is performed 85 'to find—the eigenvalue% parameter; i = _Miscellaneous Matching of the Digital Axis-Rural Rehearsal Prism and -f-mirror 87 'for shooting-images to obtain -group intensity value 88 parameters; and U4) using the former Lai characteristic and intensity fine as parameters, and perform ~ line f regression analysis 89 to obtain _ conversion Matrix 94 parameters. ㈣ 骑 纳 α ’Find the record of the thief and the thief value captured by the filter ’s bodyless reflection series_change the scale. Before obtaining the conversion matrix, a set of calibration data must be obtained. After selecting the correct target, obtain the spectrum and digital camera · value. The method of obtaining the color block spectrum data is to use the -spectrometer to obtain the corrected color block spectrum, and then use a main component. Analysis (Principal Component Analysis' PCA) to obtain the emblem vector and eigenvalues. After the color block is changed, it is thin and changeable. The regression is the least square method. (Π) — A method for carrying out portable spectrum acquisition and restoration of images, which includes: (M) a multi-image with a digital camera in front of the lens of a digital camera and a filter 92 to capture an object 5 to Obtaining a number of image spectra; (b2) By making a color draft 6 and matching with the foregoing image spectrums, in order to cut the plurality of image spectra and obtain an effective image; (b3) Passing the effective image through an intensity value 93, conversion The calculation of parameters such as matrix 94 and eigenvalue 95, supplemented by the feature space to synthesize the% -200521478 image; and (b4) after adding the parameters of the county obtained by Jubulin to the anti-coffee spectrum 97, the trueness of the image can be restored status. The reflection spectrum of the above step (b4) is the reflection spectrum of an object plus a light source, which must be divided by the light source spectrum to obtain the reflectance spectrum of an object. After obtaining the reflectance spectrum of the object, it is only necessary to obtain the spectrum of various light sources to simulate the object's operation under the various light sources. Using the circuit flow of the above-mentioned figures to seven, it is clear that the present invention is achieved by -Add digital photography before correction-multiple images and a mirror, and capture the image of an object in a single shot, you can get all the images of the object 'and provide accurate color accuracy and spectrum The film can use the color management system to process the spectrum information to the relevant device number s. All of them can be used well, and the size of the system is small, easy to make a bribe, and it is easy to move to any place—the filming of ground-dependent objects "Ϊ́ Excellent mobility", so file a patent application to find a patent. _, "The structure and features of the present invention and the embodiments have not been described in detail", and the present invention can be shown with a knife. Both the purpose and the effect are deep = life, = industrial use value 'and it is described in the spirit of the 2 patents that have not been patented in the market. The present _ is completely in accordance with the invention. Only the above is only the invention. Only the preferred embodiment When the details of the implementation of this hair riding, that is, all the equal changes and modifications applied for in accordance with the present invention, should still fall within the scope of the present patent patent 12 200521478, please ask the reviewing committee to understand and pray Huizheng is everything we pray for. [Schematic description] Figure 1 is a structure diagram of a conventional spectrum acquisition system. FIG. 2 is an implementation structure diagram of the portable spectrum system of the present invention. FIG. 3 is a schematic diagram of five frequency images obtained by the captured image of the present invention after passing through multiple images and filters. FIG. 4 is a schematic diagram of a complex spectrum image obtained by the filter according to the present invention through two spectral arrangement orders. Fig. 5 is a schematic diagram of the invention for making a color draft to facilitate cutting images and obtaining effective images. FIG. 6 is a schematic diagram of the illustrated process of acquiring multi-spectral image to restore image according to the present invention. FIG. 7 is a more detailed flowchart of the process of capturing multi-spectral images to restore images according to the present invention. Drawing number description: 11 ~ digital camera 12 ~ turntable 121 ~ filter 13 ~ object 14 ~ grayscale image 15 ~ color matrix 200521478 16 ~ reduced image 17 ~ light 2 ~ digital camera 3 ~ prism 4 ~ filter 4a ~ filter Light sheet 41 to complex spectrum image 411 to first effective image 412 to second effective image 413 to third effective image 414 to fourth effective image 415 to fifth effective image 42 to sequential arrangement method 43 to interval arrangement method 5 to object 6 ~ color draft 61 ~ color draft equally divided into five colors 62 ~ capture area 71 ~ light 72 ~ color matrix 73 ~ reduced image 81 ~ reflection spectrum 82 ~ reflection spectrum measurement 83 ~ principal component analysis method 200521478 84 ~ feature vector 85 ~ Feature space analysis 86 ~ Eigenvalue 87 ~ Digital camera with 稜鏡 and filter 88 ~ Intensity value 89 ~ Linear regression 91 ~ Image segmentation technology 92 ~ Digital camera with 稜鏡 and filter 93 ~ Intensity value 94 ~ Transformation matrix 95 ~ Feature value 96 ~ Feature space synthesis 97 ~ Reflected spectrum
1515