A7 B7 五、發明説明（) 5-1發明領域： (請先閱讀背面之注意事項再填寫本頁) 本發明係有關於一種影像處理裝置，特别是有關 於一種利用分段加權校正以改善輸出影像灰階深度的影 像處理裝置。未發明除了使用在影像系統之外，特别適用 於影像掃描器輸出影像之灰階深度的改善上.。 5-2發明背景： 傳統之影像處理系統，係用以將影像由光學*** 的反射、聚焦，經一光電轉換器之轉換成一電氣訊號，再 由電子電路對此電氣訊號做校正、調整、儲存、顯示或進 一步處理之用。此種影像***應用非常廣泛，例如影像掃 描器（image scanner)、攝影機和傳眞機等現代化世界中 常見的設備。這些設備彼此間雖然有差異，但是影像訊號 擷取和類比數位轉換都在很多影像***中被使用。 經濟智慧財產局員工消費合作社印製 以傳統技術的影像掃描器爲例，其方塊圖如圖1 所示’由一呆變數時序產生器9、一光源1〇、一玻璃表 面11、一反射鏡12、一透鏡組合13、_電荷耦合器 (Charge Couple Device ; CCD) 1 4、一由固定直流增益 電壓放大器15所構成的前置處理裝置、—類比數位轉換 器（analogue-to-digital convert or)16 和一後置處理裝置 1 7。其中呆變數時序產生器9的輸出脈波如第2圖所示。 2 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） A7 , B7__ 一·_丨丨 ' 五、發明説明（） 而且每個時間間隔是相等的，如第2圖中時間間隔2 3和 時間間隔27的大小是相同的。傳統技術的影像掃描器運 作的方式是，光源1 〇發射光線照設在其後面具有圖形或 文字的玻璃表面11，以等速度掃描破璃表面11，並反射 到反射鏡1 2，再由此反射到透鏡組合1 3上。經由透鏡組 合1 3聚焦到電荷耦合器1 4上，並由電荷耦合器1 4轉換 成一電器訊號。其轉換方式是由呆變數時序產生器9送 出如第.2圖中的脈波，而且脈波.21的前緣（f「ontedge)20 送入電荷耦合器1 4時，電荷耦合器14傳出前一時間間 隔所累積的電荷並開始儲存電荷’直到脈波2 5的前緣 (front edge)24送入電荷耦合器14時，電荷耦合器14 開始把在位於前緣20和前緣24之間的時間間隔23所得 的訊號電荷傳出並又開始儲存電荷，而此時所獲得之電氣 訊號即可用於後續電路之處理。固定直流增益電壓放大器 15調整此電氣訊號之直流增益，並且送到類比數位轉換 器16。而類比數位轉換器上的.REF +和REF -是用來取樣 的參考準位，若經過固定直流增益電壓放大器1 5所處理 過的訊號之電壓低於REF-，則其輸出訊號一律定爲系統 的最小値，此最小値是相關於系統的最小灰階値。若高於 R E F +則輸出訊號一律定爲系統的最大値，此最大値是相 關於系統的最大灰階値。後置處理器17以一Gamma函 數所產生的G a m m a對映表調整輸出影像的對比，經此而 獲得一可用以輸出或進一步處理的影像訊號。很明顯的， 若是R E F +和R E F -之間的範圍太小’可以用來表現影像 本紙浪尺度適用中國國家標準（CNS ) Μ規格（210X297公着.) -----------/V 裝-- (請先閱讀背面之注意事項再填寫本頁) 訂 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明説明（） 的灰階範圍將會變小，所以要增加輸出影像的品質就必.須 要把REF +和REF-之間的範園變大。 如第.3A圖所示，電荷辆合器（charge coupled device)的輸出特性曲線（0ljtpUt characteristic)和運算放 大器（operation amplifier)的轉移曲線（transfer curve)是 只有在一特定區間才是線性的，其餘區間都是非線性的。 就第3A圖而言’若元件是電荷耦合器，則其水平座標是 代表光通量，而垂直座標則是輸出電壓。若元件是運算放 大器’則其水平座標是代表輸入電壓，而垂直座標則是代 表輸出電壓。在第3A圖中區段b是線性區，但是區段a 和ε都是非線性區。因爲具有低光通量的圖素(pixe|s)都 落在區段a内，所以在區段a内的圖素代表的是影像的 暗部訊號。具有高光通量的圖素（pixe|s)都夢在區段c 内’所以在區段c内的圖素代表的是影像的量部訊號。 系統的轉移曲線圖示於第3 b圖中，是將第3a 圖倒轉所獲得的。第3B圖的水平座標軸代表的是輸入灰 階深度，而垂直座標軸所代表的是輸出灰階深度。由第 3B圖中可以很明顯的發現，若輪入的影像是位於暗部或 亮部’因爲區段a和c並不是線性的，所以其輸出影像的 灰階深度會產生失眞。若是可以讓區& b的範園擴大， 或是説讓原來非線性區的圖素移到線性區的範園内，則其 輸出影像的品質將會更好。 ____—____4 本紙張尺度適用中國國家標隼（CNS > A4規格（210X297公釐） (請先閱讀背面之注意事項再填寫本頁) 訂 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明説明（) 5-3發明目的及概述： 鑒於上述之發明背景中，傳統的影像處理裝置無 法在低成本的要求下，增加系統的灰階深度與位元數。本 發明之主要目的就是在提供一種改良式的訊號處理裝 置，使用不同灰階對應不同曝光時間以及不同類比數位轉 換參考準位’用以有效提昇系統的灰階深度，減少系統非 線性特性對***所產生的失眞，並以較低的成本提高輸出 影像的對比。 所以本發明的首要目的就是提供一個影像系 統，使其轉換曲線具有更大的線性區段。爲了要擴大影像 系統的線性區段’在不更動原有的電荷耦合器和運算放大 器之规格下，有一個比較簡便的方法就是控制電荷耦合系 統的曝光時間以及依據曝光時間來改變類比數位轉換器 的參考電壓，所以本發明針對影像中具有高光通量和低光 通量的部份做處理。因爲要針對高光通量和低光通量的部 份影像做處理，所以本發明把依據曝光時間來設定類比數 位轉換器的參考電壓，故本發明中的參考電壓的個數比習 知技術中的參考電壓的個數更多。 根據以上所述之目的，本發明提供了一種新的影 像處理裝置。本發明包含一個可以產生不同時間間隔之脈 波的加極控制訊號源，以及—個被加權控制訊號源所控 5 本紙張尺度適用中國國家楯準（CNS ) A4規格（210><297公楚） (請先閱讀背面之注意事項再填寫本頁) 裝. 訂 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明説明（) 制，而且可以將一個影像訊號轉換成具有不同曝光時間的 多次加權訊號的前置處理裝置。一個多次取樣系統，係用 以將一個多次加權訊號轉換爲一個多次取樣數位訊號。還 包括複數個參考電位，係用以在類比數位轉換器將一個類 比訊號轉換成數位訊號時，依據多次加權訊號源輸出訊號 的時間間隔，而使得相關於不同時間間隔的數位訊號可以 依據不同的參考電位將其轉換成數位訊號。 在第一實施例中，前述的前置處理系統包含將影 像訊號轉換成爲一個電氣訊號的電荷耦合器（charge coupled device)，以及一個調整前述電氣訊號的直流電 壓增益的直流增益電壓放大器。而直流増益電壓放大器中 包含有一個運算放大器（operation amplifier)。多次取樣 系統則包含一個取樣系統，係用與加權控制訊錄源的輸出 訊號相關之參考輸入訊號，作爲將一個類比訊號轉變成爲 數位訊號的參考準位，除此之外亦包含一個加法器。 特别的是本發明有兩個實施的方法，也就是線分 割和點分割。在線分割的方法上是使用多個類比數位轉換 器，而且每個類比數位轉換器上都有參考輸入訊號，以作 爲取樣的參考準位，並作爲後面步驟在決定哪一個灰階値 之圖素可以通過的依據。也包含多個遮軍系統，係用以讓 部份的灰階値之圖素可以通過遮罩裝置，而其依據就是參 考輸入訊號，也就是説不同曝光時間的訊號可以通過不同 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） (請先閱讀背面之注意事項再填寫本頁) '裝· 訂 經濟部智慧財產局員工消費合作社印製A7 B7 V. Description of Invention () 5-1 Field of Invention: (Please read the notes on the back before filling out this page) The present invention relates to an image processing device, and in particular to a method that uses segmented weighted correction to improve output Image processing device for image gray-scale depth. It is not invented that in addition to being used in image systems, it is particularly suitable for improving the grayscale depth of image output from image scanners. 5-2 Background of the Invention: The traditional image processing system is used to reflect and focus the image from the optical system, convert it into an electrical signal through a photoelectric converter, and then correct, adjust and store the electrical signal by the electronic circuit. , Display or further processing. Such imaging systems are used in a wide range of applications, such as image scanners, cameras, and transmitters that are common in the modern world. Although these devices are different from each other, image signal acquisition and analog digital conversion are used in many imaging systems. Printed by the Economic and Intellectual Property Bureau's Consumer Cooperative, taking the traditional technology of the image scanner as an example, the block diagram is shown in Figure 1 'from a variable variable timing generator 9, a light source 10, a glass surface 11, a mirror 12. A lens combination 13. Charge Couple Device (CCD) 1 4. A pre-processing device composed of a fixed DC gain voltage amplifier 15-analog-to-digital convert or ) 16 and a post-processing device 17. The output pulse wave of the dead time sequence generator 9 is shown in FIG. 2. 2 This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) A7, B7__ 一 · _ 丨 丨 '5. Description of the invention () And each time interval is equal, such as time interval 2 in Figure 2 The size of 3 and time interval 27 are the same. The conventional image scanner operates in such a way that the light source 10 emits light on a glass surface 11 with graphics or text behind it, scans the broken glass surface 11 at a constant speed, and reflects it to the mirror 12 and then Reflected on the lens combination 1 3. It is focused on the charge coupler 14 through the lens assembly 13 and converted into an electrical signal by the charge coupler 14. The conversion method is as follows: the pulse wave generator 9 sends out the pulse wave as shown in Fig. 2, and the leading edge (f "ontedge) 20 of the pulse wave .21 is sent to the charge coupler 14 and the charge coupler 14 transmits The charge accumulated in the previous time interval is started and the charge is stored. Until the front edge 24 of the pulse wave 2 5 is fed into the charge coupler 14, the charge coupler 14 starts to place the charge at the leading edge 20 and the leading edge 24. The electric charge of the signal obtained at the time interval of 23 is transmitted and the electric charge starts to be stored again, and the electric signal obtained at this time can be used for the subsequent circuit processing. The fixed DC gain voltage amplifier 15 adjusts the DC gain of this electric signal and sends it To the analog-to-digital converter 16. The .REF + and REF-on the analog-to-digital converter are reference levels for sampling. If the voltage of the signal processed by the fixed DC gain voltage amplifier 15 is lower than REF-, The output signal will always be the minimum value of the system, and this minimum value is the minimum gray level of the system. If it is higher than REF +, the output signal will always be the maximum value of the system. This maximum value is the maximum value of the system. gray値. The post-processor 17 adjusts the contrast of the output image with a Gamma mapping table generated by a Gamma function, and obtains an image signal that can be used for output or further processing. Obviously, if REF + and REF- The range is too small 'can be used to express the scale of the paper. This paper applies the Chinese National Standard (CNS) M specifications (210X297). ----------- / V equipment-(Please read first Note on the back, please fill out this page again) Order the A7 B7 printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 5. The gray scale range of the invention description will be smaller, so it is necessary to increase the quality of the output image. The range between + and REF- becomes larger. As shown in Fig. 3A, the output characteristic curve (0ljtpUt characteristic) of the charge coupled device and the transfer curve of the operational amplifier (transfer amplifier) ) Is linear only in a certain interval, and the remaining intervals are non-linear. As shown in Figure 3A, 'if the component is a charge coupler, its horizontal coordinate represents the luminous flux, and its vertical coordinate is the output voltage. If the component is an operational amplifier, its horizontal coordinates represent the input voltage and vertical coordinates represent the output voltage. In Figure 3A, the segment b is the linear region, but the segments a and ε are both non-linear regions. The pixels (pixe | s) of the luminous flux all fall in the segment a, so the pixels in the segment a represent the dark signal of the image. The pixels (pixe | s) with high luminous flux all dream in the segment Within c ', so the pixels in section c represent the magnitude signal of the image. The transfer curve of the system is shown in Figure 3b, which is obtained by reversing Figure 3a. The horizontal coordinate axis of Fig. 3B represents the input grayscale depth, and the vertical coordinate axis represents the output grayscale depth. From Fig. 3B, it can be clearly found that if the turn-in image is located in the dark part or the light part ', since the segments a and c are not linear, the grayscale depth of the output image will be lost. The quality of the output image will be better if the fan garden of the region & b can be enlarged, or the pixels of the original non-linear region are moved to the fan garden of the linear region. ____—____ 4 This paper size applies to the Chinese national standard (CNS > A4 size (210X297 mm) (Please read the notes on the back before filling this page). Order printed by the Intellectual Property Bureau Staff Consumer Cooperatives of the Ministry of Economic Affairs A7 B7 5. Description of the invention () 5-3 Purpose and summary of the invention: In view of the above background of the invention, the traditional image processing device cannot increase the gray scale depth and the number of bits of the system at a low cost. The main purpose of the present invention is to An improved signal processing device is provided, which uses different gray levels corresponding to different exposure times and different analog digital conversion reference levels to effectively improve the gray level depth of the system and reduce the system's non-linear characteristics. Improve the contrast of the output image at a lower cost. Therefore, the primary purpose of the present invention is to provide an image system with a larger linear segment of the conversion curve. In order to expand the linear segment of the image system, Under the specifications of charge couplers and operational amplifiers, a relatively simple method is to control the charge coupled system. The exposure time and the reference voltage of the analog-to-digital converter are changed according to the exposure time. Therefore, the present invention deals with the parts with high and low luminous fluxes in the image. Because the parts with high and low luminous fluxes need to be processed, Therefore, the present invention sets the reference voltage of the analog-to-digital converter according to the exposure time, so the number of reference voltages in the present invention is more than the number of reference voltages in the conventional technology. According to the purpose described above, the present invention A new image processing device is provided. The present invention includes an enhanced control signal source capable of generating pulse waves with different time intervals, and a weighted control signal source controlled by the source 5 The paper size is applicable to China National Standards (CNS) A4 specifications (210 > < 297). (Please read the notes on the back before filling out this page.) Packing. Order A7 B7 printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. A preprocessing device that converts an image signal into multiple weighted signals with different exposure times. A multiple sampling system The system is used to convert a multi-weighted signal into a multi-sampled digital signal. It also includes a plurality of reference potentials, which are used to convert an analog signal into a digital signal based on the multi-weighted signal when the analog-to-digital converter converts an analog signal to a digital signal. The time interval of the source output signal, so that digital signals related to different time intervals can be converted into digital signals according to different reference potentials. In the first embodiment, the aforementioned pre-processing system includes converting the image signal into a A charge coupled device of an electrical signal, and a DC gain voltage amplifier that adjusts the DC voltage gain of the electrical signal, and the DC gain voltage amplifier includes an operation amplifier. The multi-sampling system includes a sampling system, which uses the reference input signal related to the output signal of the weighted control signal source as a reference level for converting an analog signal into a digital signal. In addition, it also includes an adder. . Particularly, the present invention has two implementation methods, that is, line division and point division. The method of online segmentation uses multiple analog-to-digital converters, and each analog-to-digital converter has a reference input signal as a reference level for sampling and as a subsequent step in determining which grayscale pixel Basis for adoption. It also contains multiple masking systems, which are used to allow some grayscale pixels to pass through the masking device, and the basis is based on the reference input signal, that is, signals with different exposure times can be applied through different paper sizes. China National Standard (CNS) A4 specification (210X297 mm) (Please read the notes on the back before filling out this page)
經濟部智慧財產局員工消费合作社印製 的遮軍裝置。 在點分割的方法 s^ 0 L 决上疋使用一個類比數位轉換 器，而且这個類比數位轉 吨骑，以祚轉換态上有—組可改變的參考輸入 訊说 作爲取樣的參考準位，並作良4工土 加#、π 卫作爲後面步驟在決定哪 一個灰階値 < 圖素可以通 ^ ^ I、逦軍装置時的依據，而且輸入 是隨考加權控制訊號的週期而改變其大小的。另 外還包卜個多卫器（mu|tip|exe「），係^順料處理一 串資料中的位元。也包含-個遮罩系統，係用以讓部份的 灰階値之圖素可以通過遮罩裝置，而其依據就是參考輸入 訊號。與前述點分割内容相似，不同曝光時間的訊號可以 通過相關於不同輸入參考訊號時的遮罩裝置。 後置處理裝置，包含一個Gamma對映表，係用 於將多重取樣數位訊號依.據特性曲線，轉換到_個轉換後 的訊號’以使得輸出影像更爲平順（s m 〇 〇 t h)。並包含一個 加法器用以將不同曝光時間的訊號加成起來，以形成一完 整的影像訊號。. 5-4圖式簡單説明： 第1圖爲一傳統之影像掃描器之系統方塊圖。 第2圖爲_傳統之影像掃描器之控制訊號源的輸出波 形。 本紙張尺度適用中國國家標準（CNS > A4規格（210X297公釐） (請先閲讀背面之注意事項再填寫本頁) -裝- 訂 --:57線 經濟部智慧財產局員工消費合作社印製 A7 〜一 B7 五、發明説明（） 第3A圖爲一電荷轉合器.（charge coupled device)的輸出 訊號波形。 第3B圖爲一傳統之影像掃描器之輸出輸入轉移特性曲 (input output transfer curve)線 〇 第4圖爲本發明裝置中，使用分段加權校正的影像掃描 器之輸出輸入特性曲線，其中的線性區比習知技術的系統 的線性區更爲加大。 第5圖説明了本發明的第一實施例的影像掃描器的功能 方塊圖，其結構是使用空間分割。因爲在電荷竊合器 (charge coupled device)之前的元件與習知技術相同，且 與習知技術並無不同，所以由電荷耦合器開始説明。 第6圖説明了本發明所用的加權控制訊號源的輸出訊 號。， 第7圖顯示本發明的第一實施例的影像裝置中，輸出輸 入轉移特性曲線圖。 第8圖説明了本發明的第二實施例的影像掃描器的功能 方塊圖，其結構是使用時間分割。因爲在電荷搞合器 (charge coupled device)之前的元件與習知技術相同，且 $ 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） ---------—-----_1T------0 (請先閲讀背面之注意事項再填寫本頁) A7 B7 五、發明説明（) 與習知技術並無不同，所以由電荷耦合器開始説明。 5-5發明詳細説明： 在一影像系統中，影像係由像素（pixel)組成，如 果每個像素具有較多的色階（shade)，則可以組成更多的 顏色，所表現出來的影像也會更接近原影像，對比也會更 好。換句話説，也就是輸出影像的品質更好。根據上述所 提及，爲了改善輸出影像的品質，增加影像掃描器的灰階 深度（density range)是很重要的，因爲遑樣可以使得每一 個像素具有更多的色階。 影響影像掃描器輸出影像品質的重要因素如下 所述：首先是系統和電荷耦合器的灰階深度。設若系統的 灰階深度是D則系統的訊號雜訊比（signa卜to-Noise Ratio) : (S/N)s，和電荷耦合器的訊號雜訊比：（S/N)c， 必須滿足下列式子：[(N/S)s+(N/S)c】dCrD。其次是影像訊號 的位元數。設若影像訊號的位元數是8，則訊號雜訊比是 28/1 =256 ’所以灰階深度D是log256 = 2.4。很清楚的， 影像位元數越多灰階深度越大，也就是説可以用更多的色 階來組成輸出的影像。然後是後置處理装置所使用的 Gamma對映表。根據所得的輸入訊號的影像特性，可以 獲得一特定Gamma値的Gamma函數所產生的Gamma 對映表’用以調整輸出影像之特性。 9 本紙張尺度適用中國國家標準（CNS ) A4規格（2I0X297公釐） (請先閱讀背面之注意事項再填寫本頁> -襄· 訂' 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明説明（) 經濟部智慧財產局員工消費合作社印製 在傳統之技術中是使用單次取樣，縱使是爲了要 增加影像訊號的位元數’而使用多次取樣，但是系統的訊 號雜訊比.和電荷竊合器的訊號雜訊比仍不一定會滿足下 式：[(N/S)s+(N/S)c】s1〇 D。所以爲了要滿足上式，電荷耦合 器的訊號雜訊比可藉由分開處理及增加暗部曝光時間而 增加之使其符合上式。而且如眾所週知的，在一張圖片上 同時具有很多種灰階，由很深到很淺都有。當曝光時間一 增長’包括所有的亮部和暗部都有更長的曝光時間，所以 亮部就反射更多的光至電荷耦合器，將導致電荷耦合器飽 和（saturation)而失眞（distoration)。如上所述若將亮部 及暗部分開，暗部增長曝光處理，將可解決以上之問題， 同時也因爲分開處理’本發明得以對暗部增長曝光以增加 其光通量，並對亮部減少曝光時間以減少其光..通量.，使得 系統之輸出輸入轉移特性曲線的非線性問題得以解決。 上述的原理可參考第3A圖，輸入影像中的暗部 之圖素有較小的光通量’例如X，所以其輸出是透過區段 a的轉移而輸出。而輸入影像中的亮_部之圖素有較大的光 通量，例如Y，所以其輸出是透過區段c的轉移而輸出。 因爲系統的輸出輸入轉移特性曲線十，區段a和c是非線 性的，所以光通量位於區段a和c的圖素在輸出時，其灰 階深度會有失眞。爲了使光通量位於區段a和c的圖素可 以透過線性區段b而轉移到輸出，本發明設法使這些圖 ___ 10 本紙張尺度適用中國國家標準（CMS ) A4規格（210X297公釐） -----------裝— (請先閱讀背面之注意事項再填寫本頁)Army concealment device printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. In the method of point division s ^ 0 L, an analog-to-digital converter is used, and this analog-to-digital converter uses a set of changeable reference input signals as the reference level for sampling. And make a good 4 Gongtu plus #, π Wei as the next step in determining which gray level 値 < pixels can pass ^ ^ I, 逦 military device basis, and the input is based on the cycle of the weighted control signal test Its size. In addition, a multi-guard device (mu | tip | exe ") is included, which is used to process the bits in a string of data. It also contains a mask system, which is used to make some grayscale images. The element can pass through the masking device, and the basis is based on the reference input signal. Similar to the aforementioned point segmentation, signals with different exposure times can pass through the masking device when it is related to different input reference signals. The post-processing device contains a Gamma pair The mapping table is used to convert the multi-sampling digital signal to _ converted signals according to the characteristic curve to make the output image smoother (sm 〇〇th). It also includes an adder for different exposure time The signals are added to form a complete image signal. 5-4 Schematic description: Figure 1 is a system block diagram of a traditional image scanner. Figure 2 is a control of a traditional image scanner The output waveform of the signal source. This paper size applies to Chinese national standards (CNS > A4 size (210X297 mm) (Please read the precautions on the back before filling out this page)-Installation-Order-: 57-line Ministry of Economic Affairs Intellectual Property Printed by employees' consumer cooperatives A7 ~ B7 V. Description of the invention () Figure 3A shows the output signal waveform of a charge coupled device. Figure 3B shows the output and input transfer characteristics of a traditional image scanner. Curve (input output transfer curve). Figure 4 shows the output input characteristic curve of the image scanner using segmented weighted correction in the device of the present invention, in which the linear region is larger than the linear region of the conventional technology system. Fig. 5 illustrates a functional block diagram of the image scanner according to the first embodiment of the present invention, and its structure uses space division. Because the components before the charge coupled device are the same as the conventional technology, and There is no difference from the conventional technology, so the description will be started with a charge coupler. Fig. 6 illustrates the output signal of the weighted control signal source used in the present invention. Fig. 7 shows the image device of the first embodiment of the present invention. Output-input transition characteristic curve. Fig. 8 illustrates a functional block diagram of an image scanner according to a second embodiment of the present invention, and its structure uses time division. Because the components before the charge coupled device are the same as the conventional technology, and the paper size is applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) ------------ ----_ 1T ------ 0 (Please read the notes on the back before filling out this page) A7 B7 V. Description of the invention () is not different from the conventional technology, so the description will be started by the charge coupler. 5 -5 Detailed description of the invention: In an image system, the image system is composed of pixels. If each pixel has more shades, more colors can be composed, and the displayed image will also be The closer the original image is, the better the contrast will be. In other words, the quality of the output image is better. According to the above, in order to improve the quality of the output image, it is important to increase the density range of the image scanner, because this can make each pixel have more color levels. The important factors that affect the image quality of the image scanner output are as follows: First, the gray scale depth of the system and the charge coupler. If the gray-scale depth of the system is D, then the system's signal-to-noise ratio: (S / N) s, and the signal-to-noise ratio of the charge coupled device: (S / N) c, must meet The following formula: [(N / S) s + (N / S) c] dCrD. The second is the number of bits in the video signal. Assuming that the number of bits of the image signal is 8, the signal-to-noise ratio is 28/1 = 256 ', so the gray-scale depth D is log256 = 2.4. It is clear that the more the number of image bits, the greater the grayscale depth, which means that more output levels can be used to compose the output image. Then there is the Gamma map used by the post-processing device. According to the obtained image characteristics of the input signal, a Gamma mapping table generated by a Gamma function of a specific Gamma 値 can be obtained to adjust the characteristics of the output image. 9 This paper size is applicable to Chinese National Standard (CNS) A4 (2I0X297mm) (Please read the notes on the back before filling in this page >-Xiang · Order 'Printed by A7 B7 of the Intellectual Property Bureau of the Ministry of Economic Affairs and Consumer Cooperatives 5 Description of the invention () Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs in the traditional technology, single sampling is used, even though multiple sampling is used to increase the number of bits in the image signal, but the system signal noise The signal-to-noise ratio of the ratio and charge stealing coupler may not necessarily satisfy the following formula: [(N / S) s + (N / S) c] s1〇D. So in order to meet the above formula, the signal of the charge coupler The noise ratio can be increased to make it conform to the above formula by separately processing and increasing the exposure time of the dark part. And as is well known, there are many gray levels on a picture at the same time, from very deep to very light. When exposure As time goes on, it includes a longer exposure time for all bright and dark parts, so the bright part reflects more light to the charge coupler, which will cause the saturation and loss of the charge coupler. As above. Said If the bright part and the dark part are opened, the dark part will be increased exposure processing, which can solve the above problems, and also because of the separate processing, the present invention can increase the exposure of the dark part to increase its luminous flux, and reduce the exposure time of the bright part to reduce its light. .Flux, so that the non-linear problem of the system's output-input-transfer characteristic curve can be solved. The above principle can be referred to Figure 3A. The dark part of the input image has a smaller luminous flux, such as X, so its output is It is output through the transition of section a. However, the bright pixel in the input image has a large luminous flux, such as Y, so its output is output through the transition of section c. Because the system's output input input transfer characteristic curve 10. Segments a and c are non-linear, so pixels with luminous flux in segments a and c will lose their grayscale depth when output. In order to make pixels with luminous flux in segments a and c pass through linearly, The segment b is transferred to the output. The present invention manages to make these drawings ___ 10 This paper size applies the Chinese National Standard (CMS) A4 specification (210X297 mm) ----------- installation-(please first read Note the surface to fill out this page)
-、1T Μ. A7 B7 五、發明説明（ 素的并、s .、量改變到線性區的光通量範園内。很明顯的若是 '4? » w 里丫减少和χ增加，使其位於ρ和Q之間，則這些 2素將有線性的輸出輸入轉移特性曲線。所以藉由分段加 校正可以使得系統之輸出輸入特性轉曲線的線性 擴士甘 ’，、改變後系統之輸出輸入特性轉移曲線如第4圖 中所示。 本發明的第一實施例中使用加權多次取樣、分開 處理、+段加權校正以及影像合成，處理影像以作爲顯示 或儲存之用，所以可以運用於攝影機、錄影機、傳眞機和 1像掃描器等。本發明應用於影像掃描器系統中的實施 =，作爲本發明裝置之較佳實施例，並做一詳細説明。如 第5圖所示之系統方塊圖包含一加權控制訊號產生器 501、—電荷耦合器（Charge Couple Device: QCD)508、 一以直泥電壓增益放大器構成的前置處理系統5彳5，一類 比數位轉換器（ana丨〇g_t〇_gjtia丨c〇nverter)525、根據電 荷輕合器508的輸出訊號之灰階値來分離訊號的遮罩 510,此遮罩510可以用類比數位轉換器525的輸入參考 訊號502和5 03產生之訊號以作爲參考的灰階値、一加 法器535、一以硬體或軟體實現的Gamma對映表、一類 比數位轉換器530、一遮罩510、一以硬體或軟體實現的 Gamma對映表550以及一.加法器555。 其中加權控制訊號產生器5 〇1之輸出脈波如第6 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） -I I —1 I — I —1,裝, 111 —訂 I I—Ife (請先閲讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印製 經濟部智慧財產局員工消費合作社印製 A 7 B7 五、發明説明（） 圖所示，而時間間隔6 6是時間間隔6 3的1 0倍，而且時 間間隔6 3和時間間隔6 0 1相同，時間間隔6 0 5和時間間 隔66相同。 本發明應用於影像掃描器系統中的運作方式如 下：光學影像被聚焦到電荷耦合器508上，並由電荷耦 合器50 8轉換成一電氣訊號5 09。其轉換方式是由時序產 生器63送出如第6圖中的脈波，而且脈波61的前緣60 送入電荷耦合器508時，電荷耦合器508將前一段時間 所累積的電荷傳出並且開始儲存電荷，直到脈波65的前 緣64送入電荷耦合器508時，電荷耦合器508開始把在 位於前緣60和前緣64之間的時間間隔63所得的訊號電 荷傳出並開始儲存電荷，直到下一個脈波抵達電荷耦合 器。所以一個光學訊號被轉換爲一個電氣訊號.，而此時所 獲得之訊號即訊號509的亮部訊號。爲了獲得訊號509 的暗部訊號，加權控制訊號產生器5 0 1之輸出脈波如第6 圖所示，而時間間隔66是時間間隔63的1 0倍，也就是 説電荷耦合器5 0 8的曝光時間在暗部是亮部的1 0倍，其 方式是電荷耦合器508將時間間隔66間所累積的電荷倒 出來。 直流增益電壓放大器5 1 5放大訊號509的直流 電壓以獲得訊號511,類比數位轉換器525利用參考輸入 502和503作爲參考準位，以將訊號511轉換成一個數 本紙張尺度適用中國國家標準（CNS ) Α4規格（210X297公釐） 1-----------ί 裝-------IT1^------1 (請先閲讀背面之注意事項再填寫本頁) A7 ___ . .:. B7 _ 五、發明説明（） 位訊號，並把這個數位訊號饋至遮罩510以獲得訊號 51 3。而訊號5 1 3即是將灰階値大於2 5.5的圖素去掉， 原因敘述於後段。然後加法器535處理訊號5 1 3以執行 多次取樣並產生暗部訊號536。暗部訊號536所包含的即 是影像中的暗部之圖素。而後置處理系統是由Gamma對 應表545和加法器555所組成，用以對暗部訊號的解析 度做處理。而且Gamma對映表中的Gamma特性曲線的 Gamma値大於1。 在亮部訊號而言，訊號511經由類比數位轉換器 530和遮罩520加法器540以執行多次取樣，而產生一 數位訊號在本實施例中稱爲亮部訊號541。其中類比數位 轉換器530使用參考輸入504和505將訊號511轉換爲 數位訊號，並將此數位訊號並饋至遮罩520以獲得訊號 5 21.。若而後置處理***是.由G a m m a對應_表5 5 0和加法 器555所組成，用以對亮部訊號的解析度做處理。而且 Gamma對映表中的Gamma特性曲線的Gamma値小於 1 ° 暗部訊號536的對Gamma對應表545輸入後所 產生的輸出訊號暗部訊號546，而亮部訊號541的對 Gamma對應表550輸入後所產生的輸出訊號亮部訊號 551。而暗部訊號546和亮部訊號541由加法器將兩者加 成以產生一個影像訊號，上面所述的加權控制訊號產生器 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） (請先閲讀背面之注意事項再填寫本頁) 裝· 訂 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明説明（） 510之輸出圖示於第6圖中，此時加權控制訊號產生器 501可以用一般的計時脈波產生器（如時序產生器9之 屬），將其輸出以正反器（Flip Flop)或計數器（c〇unter),.. 等多項技術處理，即可獲得不等時間間隔之脈波輸出。因 此項技術爲熟悉此項技藝者所熟知，故未描述其結構。 在本發明的第一實施例當中，類比數位轉換器 525利用參考輸入502和503所提供的參考準位，用以 轉換訊號51 1成爲一個數位訊號，並將此數偉訊號韻至 遮軍510以獲得訊號513。而類比數位轉換器530利用 參考輸入504和505所提供的參考準位’用以轉換訊號 511成爲一個數位訊號，並將此數位訊號饋至遮罩51〇 以獲得訊號521。遮罩51 〇和520去除灰階値在其輸入 參考準位所提供的灰階參考値範圍之外的圖素。其灰階參 考値的計算方法如下： B = 255x10_d (1) 其中B代表暗部和亮部灰階値的界限，而〇代 表灰階深度。例如當輸入灰階深度大於一個統計所得的値 1時’系統的非線性特性就會產生。也就，是説在一個訊號 的暗部輸入一個傳統的影像處理系統時，輸出訊號的灰階 深度將會失眞。本發明用一個輸入灰階深度1. 〇 D，以作 爲分别處理暗部訊號和亮部訊號的依據。根據第（1)式灰 14 本紙浪尺度適用中國國家標準（CNS ) A4規格（2i〇x297公釐〉 ——^---Γ.__：--ί 裝 (請先聞讀背面之注意事項再填寫本頁)-, 1T Μ. A7 B7 V. Description of the invention (prime union, s., Quantity changed to the luminous flux range in the linear region. Obviously, if it is' 4? »W ri decreases and χ increases so that it is located in ρ and Between Q, these 2 primes will have linear output-input-transfer characteristic curves. So by segmentation plus correction, the output-input-characteristics of the system can be linearly expanded, and the output-input characteristics of the system after the change The curve is shown in Figure 4. In the first embodiment of the present invention, weighted multiple sampling, separate processing, + segment weighting correction, and image synthesis are used to process the image for display or storage, so it can be applied to cameras, Video recorder, transmitter, 1 image scanner, etc. The application of the present invention to an image scanner system is as a preferred embodiment of the device of the present invention, and a detailed description is made. The system shown in FIG. 5 The block diagram includes a weighted control signal generator 501, a charge coupled device (QCD) 508, a preprocessing system composed of a straight voltage gain amplifier 5 彳 5, and an analog digital conversion. (Ana 丨 〇g_t〇_gjtia 丨 c〇nverter) 525. Separate the mask 510 of the signal according to the gray scale 値 of the output signal of the charge coupler 508. This mask 510 can use the input reference of the analog digital converter 525 The signals generated by signals 502 and 503 are used as a reference gray scale 一, an adder 535, a Gamma mapping table implemented in hardware or software, an analog-to-digital converter 530, a mask 510, and a hardware Or software-implemented Gamma mapping table 550 and an adder 555. Among them, the output pulse of the weighted control signal generator 5 〇 is the 6th paper size applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm)- II —1 I — I —1, installed, 111 — Order II —Ife (Please read the notes on the back before filling out this page) Printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economy A 7 B7 V. Description of the invention () As shown in the figure, the time interval 6 6 is 10 times the time interval 6 3, and the time interval 6 3 is the same as the time interval 6 0 1, the time interval 6 0 5 and the time interval 66 The same applies to the image scanner The operation of the system is as follows: the optical image is focused on the charge coupler 508 and converted into an electrical signal 5 09 by the charge coupler 508. The conversion method is the pulse wave sent by the timing generator 63 as shown in Figure 6 Moreover, when the leading edge 60 of the pulse wave 61 is sent to the charge coupler 508, the charge coupler 508 transfers the charge accumulated in the previous period and starts to store the charge until the leading edge 64 of the pulse 65 is sent to the charge coupler 508. At this time, the charge coupler 508 starts to transfer the signal charges obtained at the time interval 63 between the leading edge 60 and the leading edge 64 and starts to store the charges until the next pulse wave reaches the charge coupler. Therefore, an optical signal is converted into an electrical signal, and the signal obtained at this time is the bright signal of signal 509. In order to obtain the dark signal of signal 509, the output pulse of the weighted control signal generator 5 0 1 is shown in Figure 6, and the time interval 66 is 10 times the time interval 63, that is, the charge coupler 5 0 8 The exposure time is 10 times longer in the dark than in the light. The way is that the charge coupler 508 pours out the accumulated charge between 66 time intervals. The DC gain voltage amplifier 5 1 5 amplifies the DC voltage of the signal 509 to obtain the signal 511. The analog-to-digital converter 525 uses the reference inputs 502 and 503 as reference levels to convert the signal 511 into a digital paper. CNS) Α4 specifications (210X297 mm) 1 ----------- ί equipment -------- IT1 ^ ------ 1 (Please read the precautions on the back before filling in this (Page) A7 ___. ..: B7 _ V. Description of the invention () Digital signal, and feed this digital signal to the mask 510 to obtain signal 51 3. The signal 5 1 3 is to remove the pixels whose gray level is greater than 2 5.5. The reason is described in the later section. The adder 535 then processes the signal 5 1 3 to perform multiple samplings and generates a dark signal 536. The dark signal 536 contains the dark pixels in the image. The post-processing system is composed of Gamma correspondence table 545 and adder 555, which are used to process the resolution of the dark signal. In addition, the Gamma 値 of the Gamma characteristic curve in the Gamma mapping table is greater than 1. As far as the bright signal is concerned, the signal 511 passes through the analog-to-digital converter 530 and the mask 520 adder 540 to perform multiple samplings, and a digital signal is generated in this embodiment called a bright signal 541. The analog-to-digital converter 530 uses the reference inputs 504 and 505 to convert the signal 511 into a digital signal, and feeds the digital signal to the mask 520 to obtain the signal 5 21. If the post-processing system is composed of G a m m a corresponding _ table 5 50 and an adder 555, it is used to process the resolution of the bright signal. In addition, the Gamma 値 in the Gamma map is less than 1 °. The dark signal 536 is input to the Gamma correspondence table 545 and the dark signal 546 is output. The bright signal 541 is input to the Gamma correspondence table 550. The resulting output signal highlights signal 551. The dark part signal 546 and the bright part signal 541 are added by an adder to generate an image signal. The weighted control signal generator described above applies the Chinese National Standard (CNS) A4 specification (210X297 mm) ( Please read the precautions on the back before filling this page.) Binding and printing printed by the Intellectual Property Bureau of the Ministry of Economic Affairs, employee consumer cooperative A7 B7 V. Description of the invention () The output of 510 is shown in Figure 6, at this time the weighted control signal The generator 501 can use a general timing pulse generator (such as the timing generator 9), and its output is processed by a flip flop or a counter, etc. Pulse output at different time intervals. Because this technique is well known to those skilled in the art, its structure is not described. In the first embodiment of the present invention, the analog-to-digital converter 525 uses the reference levels provided by the reference inputs 502 and 503 to convert the signal 51 1 into a digital signal, and rhymes the signal to cover army 510. To get signal 513. The analog digital converter 530 uses the reference levels provided by the reference inputs 504 and 505 to convert the signal 511 into a digital signal, and feeds the digital signal to the mask 51 to obtain the signal 521. Masks 51 and 520 remove pixels whose gray scales are outside the range of gray scale reference levels provided by their input reference levels. The calculation method of the gray scale reference frame is as follows: B = 255x10_d (1) where B represents the limit of the gray scale frame in the dark and bright areas, and 0 represents the gray scale depth. For example, when the input gray-scale depth is greater than a statistically obtained 値 1 ', the nonlinear characteristics of the system will occur. In other words, when a traditional image processing system is input to the dark part of a signal, the grayscale depth of the output signal will be lost. The present invention uses an input gray level depth of 1.0 D as a basis for processing the dark portion signal and the bright portion signal separately. According to the formula (1), the paper size of this paper is 14 and the Chinese national standard (CNS) A4 specification (2i0x297 mm) is applied. —— ^ --- Γ .__:-ί (please read the precautions on the back first) (Fill in this page again)
II J,> 經濟部智慧財產局員工消費合作社印製 hi B7 經濟部智慧財產局員工消費合作社印製 五、發明説明（ 階値灰階深度1.0 D時，JL灰枯* c ^ 〜及1%値爲22.5以作爲分别加 權裝置的界限。 在本發明的第一實始丨士 ^兀施例中，當權値爲10時，輸 入參考5 0 2和5 0 3分别被辞仝A 士II J, > Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs B7 Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs V. Description of the invention (When the depth of the gray scale is 1.0 D, JL is grayed out * c ^ ~ and 1 % 値 is 22.5 as the limit of the respective weighting device. In the first embodiment of the present invention, when the weight is 10, the input reference 5 0 2 and 5 0 3 are respectively regarded as A 1
义爲有關於灰階深度爲1.0D 的參考準位，而且有關於灰階 ’人階'衣度等於4.0 D的是暗部參 考準位。權値爲1時，輸束 爾入參考5〇4和505分别被設定 爲有關於灰階深度爲〇_〇D的白择炎土推厂 . 7白色參考準位，而且有關於 灰階深度等於4.0D的是暗部，灰奂唯， 疋吋邵參考準位。而權値則是用來 決定電荷耦合$ 508的曝光間之用，換句話説，權値也 就等於是控制訊號源的輸出脈波之時間間隔。在第一實施 例中，參照第6圖，時間義66是時間間隔63的十倍。 因此在遮罩510和520中爾也八iifc . π ^甲用來分離暗部和亮部的灰階値 可以計算得知爲25.5。因此兔棘念邮1Λ . 此W說灰階値大於25.5即爲亮 部，而小於2 5 · 5則爲暗部。 如第7圖所示，是本發明的第—實施例之系統的 輸出輸入轉換特性曲線。其中的虛線部份是代表傳統的系 統之輸出輸入特性曲線，而實線部份則是代表第一實施例 的系統之輸出輸入轉移特性曲線。很明顯的，非線性發生 的區段之輸入灰階深度由UD # ZD系、统的線性區域增 加了。 本發明的第二個實施例所説明的是—個時間分 15 本紙張尺度適用中國國家標準（CNS ) A4規格（210X 297公釐） A , ; - 1 - )e ---------------，'裝 I-----訂------^-- ^ (請先閲讀背面之注意事項再填寫本頁) A7 B7 經濟部智慧財產局員工消費合作社印製 五、發明説明（ 割的例子。參考第8圏， m 包含一加權控制訊號產生器 801、.一電荷耦合808、一以古、士奋m , 以直泥電壓增益放大器構成的 前置處理系統81 5，—類计赵γ絲& 觸比數位轉換器825、根據電荷輕 合器808的輸出訊號之龙略 及階値來分離訊號的遮罩810,此 遮罩 81 0可以用類t卜畆ρ 顯比數仨轉換器825的輸入參考訊號 802和8Ό3產生之却妹 "就乂作爲參考的灰階値、一多工器 (multiphexer)835、一 加.让 gg 加法器835、一以硬體或軟體實現 的Gamma對映表845、以及一加珐器855。 其中加權控制訊號產生器801之輸出脈波如第6 圖所示，饋至電荷耦合器8〇8。加權控制訊號產生器8〇1、 電荷輕合器808、前置處理系統81 5以及類比數位轉換器 825的運作方式和第一實施例中的完全相同。 電氣訊號809爲直流增益電壓放大器815所處 理而產生訊號811。類比數位轉換器825參考輸入802 和803所提供的參考準位以將訊號811轉換爲一個數位 訊號817,並將此數位訊號82 0饋至多工器82 0。因爲訊 號8 1 1内含不同的權重訊號，所以類比數位轉換器825 利用隨著加權控制訊號源輸出訊號的週期而變的參考輸 入802和803，以作爲取樣時的參考準位。 其轉換方式是由時序產生器63送出如第6圖中 的脈波，而且脈波61的前緣60送入電荷耦合器808時， 16 本紙張尺度適用中國國家標準（CNS ) A4规格（210X297公釐） I !卜！！i裝-----^-I 訂---J——TW (請先聞讀背面之注意事項再填寫本頁) A7 B7 五、發明説明（) 參考輪入802和803分别被設定爲相關於灰階深度〇 〇d 的白色參考準位以及相關於灰階深度4. 〇 d的暗部參考準 位’而且其權値爲1。當時間間隔66通過電荷耦合器808 時’參考輸入802變成相關於灰階深度i 〇D的參考準 位’而參考輸入803則仍然是相關於灰階深度4 〇D的暗 部參考準位，而且其權値爲1 〇，意即其作用時間在加權 控制訊號源的輸出脈波之時間間隔爲前面之1 〇倍。 在第6圖的週期63中，多工器820選擇權重爲 1的訊號並送至遮罩810。此時遮罩810的參考準位是灰 階値2 5.5，換句話説，凡是送到遮罩81〇的圖素中灰階 値小於25.5的部份之圖素都會被去除。在第6圖的週期 605中’多工器820選擇權重爲1 〇的訊號並送至遮罩 81 0。此時遮罩81 0的參考準位仍是灰階値25 5，但是凡 是送到遮罩810的圖素中灰階値大於25 5的部份之圖素 .都會被去除。因爲多工器820變換的週期是隨著加權控 制訊號源的週期而變動，所以可以選擇不同曝光時的訊 號，所以此方法也可以用一組的元件而達到分開處理訊號 的目的。 加法器835和遮罩810 —起運用以執行多次加 權取樣的功能’而且產生一個數位訊號，在第二實施例中 稱爲亮部訊號或暗部訊號，端視其訊號的權重而定，而後 置處理系統是由Gamma對映表845和一個加法器855 17 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） (請先閣讀背面之注意事項再填寫本頁) 、言 Γ 經濟部智慧財產局員工消費合作社印製 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明説明（） 所形成。而且Gamma對映表中的Gamma特性曲線的 Gamma値大於1時，所處理的就是暗部訊號。當Gamma 對映表中的Gamma特性曲線的Gamma値小於1時，所 處理的就是凴部訊號。換句話説.，Gamma對映表中的 G a m m a値是與加權控制訊號源8〇 1的週期同步改變，然 後亮部訊號和暗部訊號再加成以獲得一個完整的影像訊 號。 若本發明使用在彩色掃描器中，可以用兩種不同 的方式來.處理三原色（紅.、黄和藍色）的訊號。第一種方法 是線分割，用本發明的第一或第二個實施例都可以實施。 另一種方法是點分割，可以用本發明第二個實施例據以實 施0 本發明的另一個實施例是用一個可改變燈光強 度的光源來取代前面第一和第二實施例中固定光強度的 光源。爲了要控制電荷耦合器上的光通量，本實施例使用 習知技術中的呆變數控制訊號源和可改變燈光強度的光 源’而且本實施例也可以對影像系統的輸出輸入轉移特性 曲線的線性區加以擴大。因爲光通量是由光強度乘以曝光 時間而得到的，所以在本實施例中是使用一個可改變燈光 強度的光源，和一個被呆變數控制訊號源所控制的電荷耦 合器來控制電荷耦合器上的光通量。而可改變燈光強度的 光源是由加權控制訊號源所控制的，而其控制電路是一個 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） -----—h!---,.裝-------訂11-^---1 W (讀先閱讀背面之注意事項再填寫本頁) A7 ^--~_—_ 扪 — 五、發明説明（） 傳统的控制電源能量的電路^所以一個影像訊號在此實施 例中仍然是被分開處理的，就如前面第一和第二實施例中 所敘述的，可以擴大此影像系統的輸出輸入特性轉移曲線 之線性區。 這個處理影像訊號的影像處理裝置，提供了一個 擴大輸出輸入轉移特性曲線之線性區的方法，爲了要擴大 ***的輸出輸入轉移特性曲線之線性區段，電荷偶合器的 曝光時間必須加權，而且類比數位轉換器的參考準位必須 根據其所處理的訊號之權重，而有不同的參考準位。囡此 本發明分别對高光通量和低光通量的區段之訊號分别處 理。也因爲不同的曝光時間，把原來位於輸出輸入轉移特 性曲線中非線性區的亮部和暗部的圖素都轉移到線性區 中’所以其輸出訊號有很良好的輸出影像品質。 以上所述僅爲本發明之故佳實施例而已，並非用 以限定本發明之申請專利範圍；凡其它未脱離本發明所揭 示之精神下所完成之等效改變或修飾，均應包含在下述之 申請專利範固内。 ----^--hi.---裝------訂------^ (讀先鬩讀背面之注意事續再填寫本頁) 經濟部智慧財產局員工消费合作社印製 9 本紙張尺度適用中國國家標準（CNS ) A4规格（210X2.97公釐）This means that there is a reference level for the gray level depth of 1.0D, and there is a dark reference level for gray level 'human level' clothing level equal to 4.0 D. When the weight is 1, the beam reference 504 and 505 are set as Bai Zeyan soil push factory with gray level depth of 0_〇D. 7 white reference level, and the gray level depth is equal to 4.0 The D is the dark part, the gray part is gray, and the reference level is 疋 inch shao. The weight is used to determine the charge-coupled exposure of $ 508. In other words, the weight is equivalent to controlling the time interval of the output pulse of the signal source. In the first embodiment, referring to Fig. 6, the time 66 is ten times the time interval 63. Therefore, in the masks 510 and 520, there are eight iifc. The gray scale 用来, which is used to separate the dark and bright parts, can be calculated to be 25.5. Therefore, the rabbit spine reads 1Λ. This W says that the gray scale 値 is greater than 25.5 is the bright part, and less than 2 5 · 5 is the dark part. As shown in Fig. 7, it is an output-input conversion characteristic curve of the system of the first embodiment of the present invention. The dotted line part represents the output-input characteristic curve of the conventional system, while the solid line part represents the output-input transfer characteristic curve of the system of the first embodiment. Obviously, the input gray-scale depth of the non-linear section is increased by the linear region of the UD # ZD system. What is explained in the second embodiment of the present invention is that a time scale of 15 paper sizes is applicable to the Chinese National Standard (CNS) A4 specification (210X 297 mm) A,;-1-) e ------- --------, 'Install I ----- Order ------ ^-^ (Please read the precautions on the back before filling this page) A7 B7 Staff Consumption of Intellectual Property Bureau, Ministry of Economic Affairs Cooperative printed five, an example of the invention (cut example. Refer to Section 8), m contains a weighted control signal generator 801, a charge coupled 808, an ancient, venerable m, and a former composed of a straight mud voltage gain amplifier. Set the processing system 81 5—the class of Zhao γ silk & touch ratio digital converter 825, the mask 810 of the signal is separated according to the output signal of the charge light coupler 808 and the step 値, this mask 8 10 can Use the input reference signals 802 and 8Ό3 of the class t 畆 ρ explicit ratio digital converter 825 to generate the &" gray scale 乂 as a reference, a multiplexer (multiphexer) 835, and one add. Let gg add Generator 835, a Gamma mapping table 845 implemented in hardware or software, and an enamelizer 855. The output pulses of the weighted control signal generator 801 are as in section 6 As shown, it is fed to the charge coupler 808. The weighted control signal generator 801, the charge coupler 808, the pre-processing system 815, and the analog-to-digital converter 825 operate in the same manner as in the first embodiment. The same. The electrical signal 809 is processed by the DC gain voltage amplifier 815 to generate a signal 811. The analog digital converter 825 refers to the reference levels provided by the inputs 802 and 803 to convert the signal 811 to a digital signal 817, and converts the digital signal 82 0 is fed to the multiplexer 82 0. Because the signal 8 1 1 contains different weighted signals, the analog-to-digital converter 825 uses the reference inputs 802 and 803 that vary with the period of the weighted control signal output signal for sampling. The reference level of the time. The conversion method is when the timing generator 63 sends out the pulse wave as shown in Figure 6, and the leading edge 60 of the pulse wave 61 is sent to the charge coupler 808. 16 This paper size applies the Chinese national standard ( CNS) A4 specification (210X297mm) I! Bu !!! i installed ----- ^-I ordered --- J-TW (please read the precautions on the back before filling this page) A7 B7 V. Description of the invention () Reference rounds 802 and 803 Do n’t set it to the white reference level related to the gray level depth of 〇d and the dark reference level related to the gray level depth of 4.0 d 'and its weight to be 1. When the time interval 66 passes through the charge coupler 808 'The reference input 802 becomes the reference level related to the gray-scale depth i 0D', while the reference input 803 is still the dark reference level related to the gray-scale depth 4 OD, and its weight is 1 〇, which means that The time interval during which the action time is at the output pulse of the weighted control signal source is 10 times the previous. In the period 63 in FIG. 6, the multiplexer 820 selects a signal with a weight of 1 and sends it to the mask 810. At this time, the reference level of the mask 810 is gray level 値 2 5.5. In other words, all pixels in the gray level 値 of the pixels sent to the mask 810 are less than 25.5. In the cycle 605 in Fig. 6, the 'multiplexer 820 selects a signal with a weight of 10 and sends it to the mask 810. At this time, the reference level of the mask 81 0 is still grayscale 値 25 5, but all pixels in the grayscale 値 greater than 25 5 of the pixels sent to the mask 810 will be removed. Because the period of the multiplexer 820 changes with the period of the weighted control signal source, signals at different exposures can be selected. Therefore, this method can also use a group of components to achieve the purpose of processing signals separately. The adder 835 and the mask 810 work together to perform multiple weighted sampling functions and generate a digital signal, which is called a bright signal or a dark signal in the second embodiment, depending on the weight of the signal, and then The processing system is composed of a Gamma mapping table 845 and an adder 855 17 This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) (please read the precautions on the back before filling out this page). Printed by the Consumers 'Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs Printed A7 B7 by the Consumers' Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 5. Description of the invention (). And when the Gamma 値 of the Gamma characteristic curve in the Gamma mapping table is greater than 1, what is processed is the dark signal. When the Gamma 値 of the Gamma characteristic curve in the Gamma mapping table is less than 1, what is processed is the crotch signal. In other words, G a m m a 中 in the Gamma mapping table changes in synchronization with the period of the weighted control signal source 801, and then the bright and dark signals are added to obtain a complete image signal. If the present invention is used in a color scanner, the signals of the three primary colors (red, yellow, and blue) can be processed in two different ways. The first method is line segmentation, which can be implemented with either the first or second embodiment of the present invention. Another method is point segmentation, which can be implemented by the second embodiment of the present invention. Another embodiment of the present invention is to replace the fixed light intensity in the first and second embodiments with a light source that can change the light intensity. Light source. In order to control the luminous flux on the charge coupler, this embodiment uses the conventional variable technology to control the signal source and the light source that can change the light intensity. In addition, this embodiment can also transfer the linear region of the characteristic curve of the input and output of the imaging system. To expand. Because the luminous flux is obtained by multiplying the light intensity by the exposure time, in this embodiment, a light source that can change the light intensity and a charge coupler controlled by a variable source control signal source are used to control the charge coupler. Luminous flux. The light source that can change the light intensity is controlled by a weighted control signal source, and its control circuit is a paper standard that applies the Chinese National Standard (CNS) A4 specification (210X297 mm) ------- h!- -,. equipment ------- order 11-^ --- 1 W (read the precautions on the back before filling out this page) A7 ^-~ _—_ 扪 — 5. Description of the invention () Traditional The circuit for controlling the power of the power source ^ Therefore, an image signal is still processed separately in this embodiment. As described in the first and second embodiments, the output-input characteristic transfer curve of the image system can be enlarged. Linear area. This image processing device for processing image signals provides a method for expanding the linear region of the output-input-transfer characteristic curve. In order to expand the linear region of the output-input-transfer characteristic curve of the system, the exposure time of the charge coupler must be weighted, and the analogy The reference level of a digitizer must have different reference levels based on the weight of the signal it is processing. Therefore, the present invention processes the signals of high luminous flux and low luminous flux segments separately. Also, because of the different exposure times, the pixels in the light and dark parts of the non-linear region in the output-input transfer characteristic curve are all transferred to the linear region ', so the output signal has very good output image quality. The above are only good examples of the present invention, and are not intended to limit the scope of patent application of the present invention; all other equivalent changes or modifications made without departing from the spirit disclosed by the present invention should be included in the following The patent application mentioned above is within the scope of the patent. ---- ^-hi .--- install ------ order ------ ^ (read first, read the notes on the back and continue on this page, then fill in this page) Employees ’Cooperatives, Intellectual Property Bureau, Ministry of Economic Affairs Printed 9 This paper is sized for China National Standard (CNS) A4 (210X2.97 mm)