M395324 卯年6月23日替換頁 五、新型說明： 【新型所屬之技術領域】 本創作係關於一種影像輪出校正裝置，特別是一種具有查閱 表(Look-Up Table ; LUT)之多通道相機。 【先前技術】 近年來，影像感測器係例如電荷耦合器（Charge_c〇upled Device ; CCD)及互補式金屬_氧化物_半導體(c〇mplementary Metal-Oxide Semiconductor ; CMOS)廣泛用於視訊設備、照相機及 工業檢測相機等，且逐步朝向多像素化發展，因此將此等大量像 素於特定時間内讀出是必須達到的。然而，對應於多像素化之電 荷耦合器所利用之高速讀取的方式之一，係將電荷耦合器的所有 像素分割成複數個通道，並且讀取分割後通道像素的多分割讀取 方式。此方法係將電荷耦合器所聚集之複數個像素分割成複數個 通遏，使得影像訊號可分割成複數組訊號分別傳輸，以達到高速 §買取的效果。 請參閱「第1A圖」及「第1B圖」，「第1A圖」係為習知技 術中之多通道相機之示意圖、「第1B圖」係為習知技術中影像資 料畫面之示意圖。如「第1A圖」所示，在習知技術中多通道相機， 其包含有一影像感測器10、複數個類比前端(Anal〇g Fr〇m ; AFE)20、以及複數個伽瑪校正||(Gamma c〇rrect〇r)3〇。其中景《像 感測器10係由複數組通道所組成且用以掃描一影像，而當影像减 測器10掃描一影像時，影像會依據多通道之通道數目分割成相同 ^ Ι_ϋϋ^23日替換頁 目之複數組類比訊號’並且經由複數個類比前端轉換成數位 唬輸出接著，數位訊號則各自傳送至一伽瑪校正器如來對複 數組訊號進行伽瑪校正後輸出為—數位影像資料。料如「第m 圖」所不，利用上述之影像處理所得之影像資料晝面4〇，會出現 如圖上複數條明線42及暗線41相間，其中明線42係由一通道輸 出之數位影像資料、而暗線41係由另一通道輸出之數位影像資 料由於不同通道係與影像感測器中之不同位置的像素對麻， 轉：通道接《彡像訊號在透稿比前端2_換成數位訊躺過 ㈣Α由各自的獨立電路來完成，因此由於影像制器1G各像素 2程中品質料穩定以及獨立電路中各元件的差碰，都可能 =影像貧料畫面4G中如「第u圖」所示之亮度不均勻的問題。 :耗^域則利用個人電腦端之校正軟體進行修改與校正， 此夕 夕的人力與時間，增加繁複的程序及成本。此外，當 晝Ϊ通業檢測用途時’需要有低雜訊、品質清晰的 色彩二理與判讀’因此’若不同通道 取迷度外，對於除了如前述般致力於提高讀 讀取結果狀-致，之像細之影像 【新型内容】h轉—崎的目標。 馨於以上的問題’本創作的主要目的在於提供一種影像輪出 L99年6月23日替換頁 校正裝置贿決習知肋巾錄紅業頓常常因為影像感測器 之像素製程品質不佳’且多通道相機之各通道由於其具有獨立線 路且其組舰路之各元件有其差紐而使得各通道影像訊號輸出 的亮度不均勻，造成影像品質不佳、檢測不易與判斷錯誤、增加 製程成本等問題。 為達到以上所述之目的，本創作揭露一種影像輸出校正褒 置，其接收一影像訊號並輸出複數組數位校正值，其包括有一多 通道相機及複數個校正難。㈣通道相機包括有一影像感測器 及-類比前端。其中影像感測H係具有複數個像素，且複數個像 素係切分為複數個通道，並逐—掃描—影像，而轉換為與通道一 一對應之複數組類比訊號。複數個類比前端(AnalGg F1Ont End; ,E) π別連接至影像感測器之各個通道，用以分別接收各個通 道輸出之複數域比喊’並分獅複數峨比訊麟換成複數 組第-數位訊號值複數個校正歡，可選擇性地設置於多通道 相機及應用執彳了於_電腦端，且各校正歡其分職各類比前端 連接並接收各類比別端之複數組第一數位訊號值，各校正模 組具有至少-查_(L⑻k_Up職；LUT)，而每—錢表係包含 有複數健位冗度項值及與複數舰位亮度讀仙對應之複數個 數位奴正值，喊數個校正模域依據複數組第—數位訊號值而 ^對與第—數健舰辦之數位亮度·並輸出與婦之數位 凡度喝值所對應之數位校正值，而其巾複數個細表係藉由該影 像感測器中各通道中各像素掃描至由複數個灰階度所組成之一灰 99年6月23曰卷挣▼ 階影像’並於各灰階中取各通道所有像素之亮度代表值，且各個 灰階度中就補通道之鱗亮度代表值而定義—數位校正值。 本創作之功效在於建立影像輸出校正裝置查閱表時決多通 道相機輸出亮度之不平衡關題。_此職於各像素之查閱 表’可以使得由於電荷麵合器之像素製程品質不佳，且多通道相 機之各通道由於其具域路且其組成線路之各树有其差異 I*生而使件各通龄彡像喊輸出的亮度不均勻的問題，在數位訊號 輸出類比前料即藉由細表將各通道巾各像素之亮度值調整至ϋ 一相同的壳度參考值，使各通道輸出的亮度均勻一致。 以上之關於本創作内容之制及以下之實施方式之說明係用 以示範與解釋本創作之原理，並績供本_之專利申請範 進一步之解釋。 【實施方式】 為使對本創作的目的、構造'特徵、及其功能有進一步的瞭 解，拉配合實施例詳細說明如下。 ” 本創作提出-種解決多通道相機輪出不平衡之裝置及其方 法。首先，請參閱「第2圖」’「第2圖」係為本創作之影像輪出 校正裳置翻於多通道減的示意®。而如「第2圖」所示， 創作之影像輸出校正裝㈣包含有—多通道相機及複數個校正^ j施。而多通道相機係包含有一影像感測器2〇1& 一複數個類比 刖端202，其中影像感測器2〇1倍可為一電荷輕合哭 (〇!·〇>’ Deviee ; CCD)或為—互補式金純化物半^ 99年6月23日替換頁 (Complementary Metal-Oxide Semiconductor; CMOS)用以接收一影 像訊號。複數個校正模組206設置於多通道相機上，而每一校正 模組206係儲存有一查閱表(L〇〇kUpTable ; LUT)2〇7及一伽瑪校 正器（Gamma Corrector)208。 由於上述之影像感測器2〇1係由複數個像素所組成，並且複 數個像素被切分成複數個通道(channel)，一般來說，被切分為2 個以上通道，使每—通道中包含有複數個像素，因此，當影像感 測器201掃描-影像時，影像會依據通道數目分割成相同數目之 複數組類比訊號，並分別傳送至與影像感測器2〇1相連接之複數 個類比前端202，其數目亦與通道數目相同，當類比前端搬接收 來自影像感測器201之複數組類比訊號時，類比前端搬會將各 組類比訊雜換成第—數位訊號值輸出’其巾包括將訊號的數值 放大。並且於每-類比_ 2〇2之後連接—校正模Μ施，而校正 模組206用以接收自各類比前端搬所輸出之第一數位訊號值， 其中各校正模組施針對各通道對應有至少—個查閱表浙，而每 -查閱表2〇7包含有複數個數位亮度讀值以及複數個數位校正 值，且數位党度讀值與數位校正值相對應。因此，校正模組施 中之查閱表207係依據自類比前端搬輸出之第一數位訊號值並 一一比對於複數個數位亮度讀值’當其中數位亮度讀值與第一數 位訊號值鱗時’而輸出與第—數位訊號值相等的數位亮度讀值 所對應之數位校正值。最後，再將查閱表2G7輸出之數位校正值 輸入-伽瑪校正ϋ 2G8來對更進—步對數位校正值進行訊號之細 99年6月23曰替換頁 部對比與亮度校錢輸出—触影像資料。 /因此本創作提供複數個校正模組施連接於類比前端搬 ^後以知技術無法解決之多通道減輸出不平衡的問 題…、而建立影像輸出校正裝置查閱表之方法將在以下詳述之。 ,請參閱「第3圖」及「第6圖」，「第3圖」係為本創作之建 τ/像輸出；置朗表的方法示意圖，❿「第6圖」係為本 J乍之建立〜像輸出校正裝置查閱表的方法流程圖。則如「第3 圖」所不，首先，先以一般傳統方法來調整多通道相機中每一組 通道為虎的增益值(Gain)及訊號的偏移量(〇ff㈣，分別用以放大訊 號及消除訊射附帶的雜訊，並盡量將輸出織之亮度調整至相 接t帛3圖」中由左到右係為由白到黑的灰階影像21〇 〇 接著’將-影像感測器201掃描至一灰階影像21〇上，而灰 1^白IV像210係由複數個灰階所構成。於本實施例中灰階影像2川 係由24灰所構成’而本實施例中所使用之影像感測器期係為 位元輸出’因此可以解析成256灰階度輸出，及灰階影像21〇 之24灰階可均勻分佈於256灰階度中24個灰階度。又如「第6 圖」所示，將多通道相機之影像制H 201之每-通道之每一像 素皆依序掃描至灰P緣像21㈣每—個灰階上(步驟_，且於每 個灰階中，各像素均可以得到一第一數位訊號值(步驟卿。而 各通逼將通這巾所有像素於同—灰階所得之複數個第—數位訊號 值取-党度代表值(步驟130)，則每—通道於灰階影像⑽上掃描 後’可於24 p旁灰階中藉由平均各灰階所有像素之第一數位訊號值 M395324 L99年6月23日替換頁 或是取所有像素之第一數位訊號值之中位數數值而得到％個亮度 代表值。因此，每一通道掃描灰階影像21〇所得到24個亮度代表 值’可以依亮度代表值與灰階度之對應關係作圖。接著，於各灰 階中’依據複數個通道在每—灰階之亮度代表值定義—數位校正 值(步驟14〇) ’則就所有灰階則定義以一組數位校正值(步驟150) ’ 則各通道巾每-亮度代表鋪應於—數倾正值，而各通道於所 有灰階之亮度代錄係為—峻位亮度讀值(步驟副），並就對應 於24灰階之24個亮度代表值及24健位校正值分別將任兩亮^ 絲值(_亮度讀值)咖及亮度代綠(賴触亮度讀切 之化點以連續整數插人亮度代表值(亦稱數位亮度讀值)並利用内 ^法及外H雜運算而得崎應之數錢正值(步驟I%)，最 糾建立各通道之查難，射查_包括有複數她位亮度讀 值及相對應之概她位校域(步驟18〇)。 下闡為了更具關露本創作之具有查_之校正模組，以 將以四通道多通道相機為例，請參閱「表1」及「第4B圖」， =」係為本創作之影像輪出校正裝置之各通道就 “像掃描而得之亮度代表值列表。 火丨白之火 M395324M395324 Replacement page on June 23 of the following year. New description: [New technical field] This creation is about an image wheel-out correction device, especially a multi-channel camera with a look-up table (LUT). . [Prior Art] In recent years, image sensors such as Charge Coupled Device (CCD) and Complementary Metal-Oxide Semiconductor (CMOS) have been widely used in video equipment. Cameras and industrial inspection cameras, etc., are gradually moving toward multi-pixel development, so it is necessary to read such a large number of pixels in a specific time. However, one of the ways of high-speed reading utilized by the multi-pixelized charge coupler is to divide all the pixels of the charge coupler into a plurality of channels and read the multi-segment read mode of the divided channel pixels. In this method, the plurality of pixels collected by the charge coupler are divided into a plurality of channels, so that the image signal can be divided into complex array signals for transmission, so as to achieve the high-speed § buy effect. Please refer to "1A" and "1B". "1A" is a schematic diagram of a multi-channel camera in the prior art, and "1B" is a schematic diagram of an image data screen in the prior art. As shown in FIG. 1A, a multi-channel camera in the prior art includes an image sensor 10, a plurality of analog front ends (Anal〇g Fr〇m; AFE) 20, and a plurality of gamma corrections. |(Gamma c〇rrect〇r) 3〇. The image sensor 10 is composed of a plurality of array channels and is used to scan an image. When the image detector 10 scans an image, the image is divided into the same according to the number of channels of the multi-channel ^ Ι _ ϋϋ ^ 23 The complex array analog signal of the page is replaced and converted into a digital output via a plurality of analog front ends. Then, the digital signals are respectively transmitted to a gamma corrector to perform gamma correction on the complex array signal and output as digital image data. For example, if the "mth image" is not used, the image data obtained by the above image processing will be 4〇, and there will be a plurality of bright lines 42 and dark lines 41 in the figure, wherein the bright line 42 is a digital output by one channel. The image data and the dark line 41 are digital image data output by another channel. Because the pixels of different channels and different positions in the image sensor are numb, the channel is connected with the image signal. The digital information is placed in (4) and is completed by its own independent circuit. Therefore, since the quality of the pixels in the 1G pixel of the image controller is stable and the components of the independent circuit are in contact with each other, the image may be in the image poor image 4G. The problem of uneven brightness is shown in u. : The consumption domain is modified and corrected by the PC software. This time, the manpower and time increase the complicated procedures and costs. In addition, when the use of 昼Ϊ 昼Ϊ 检测 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' To, the image like a fine [new content] h turn - the goal of Saki. The problem with the above is 'The main purpose of this creation is to provide an image rotation on June 23, 1999. Replacement page correction device, bribery, and ribbed tape. Red industry often suffers from poor image processing quality of image sensor.' Moreover, each channel of the multi-channel camera has uneven brightness due to its independent line and the components of its group ship having the difference of the image signal output, resulting in poor image quality, difficulty in detection and error determination, and increased process. Cost and other issues. To achieve the above, the present invention discloses an image output correction device that receives an image signal and outputs a complex array digital correction value including a multi-channel camera and a plurality of calibration difficulties. (4) The channel camera includes an image sensor and an analog front end. The image sensing H system has a plurality of pixels, and the plurality of pixels are divided into a plurality of channels, and are scanned one by one, and converted into a complex array analog signal corresponding to the channel one by one. A plurality of analog front ends (AnalGg F1Ont End; , E) π are not connected to each channel of the image sensor, respectively, for receiving the complex domain of each channel output than the shout 'and the lion complex number 峨 讯 讯 换 换 复- The digital signal value is a plurality of corrections, which can be selectively set in the multi-channel camera and the application is executed on the _ computer end, and each correction is divided into various types of front-end connections and receiving various types of complex arrays than the other end The first digital signal value, each calibration module has at least - check _ (L (8) k_Up position; LUT), and each money table includes a plurality of health time redundancy item values and a plurality of digits corresponding to the plurality of ship brightness readings The slave is positive, shouting a number of correction modes based on the complex array first-digit signal value and the digital brightness corresponding to the first-numbered ship, and outputting the digital correction value corresponding to the woman's digital value. The plurality of fine meters of the towel are scanned by each pixel in each channel of the image sensor to be composed of a plurality of gray scales, and the gray image is generated by the gray level on June 23, 1999. Take the brightness representative value of all pixels of each channel, and each gray scale It scales up the luminance representative value of the channel defined - the digital correction value. The effect of this creation is to establish an imbalance of the output brightness of the multi-channel camera when the image output correction device is inspected. _ This look-up table for each pixel can make the pixel processing quality of the charge-matching device poor, and each channel of the multi-channel camera has its difference due to its domain path and the trees of its constituent lines. In order to make the brightness of the output of each channel screaming uneven, the brightness value of each pixel of each channel towel is adjusted to the same shell reference value by a fine meter in the analog output of the digital signal. The brightness of the channel output is uniform. The above description of the present invention and the following description of the embodiments are intended to demonstrate and explain the principles of the present invention and to provide further explanation of the patent application. [Embodiment] In order to further understand the purpose, structure, and function of the present invention, the pull-fit embodiment will be described in detail below. This creation proposes a device and method for solving the imbalance of multi-channel camera wheel. First, please refer to "Fig. 2" and "Fig. 2" for the image rotation correction of this creation. Reduced indications®. As shown in "Figure 2", the created image output correction device (4) includes a multi-channel camera and a plurality of corrections. The multi-channel camera system includes an image sensor 2〇1& a plurality of analog terminals 202, wherein the image sensor 2〇 can be a charge of lightly crying (〇!·〇>' Deviee; CCD Or - Complementary Metal-Oxide Semiconductor (CMOS) is used to receive an image signal. A plurality of correction modules 206 are disposed on the multi-channel camera, and each calibration module 206 stores a look-up table (L〇〇kUpTable; LUT) 2〇7 and a gamma correctionor 208. Since the image sensor 2〇1 is composed of a plurality of pixels, and the plurality of pixels are divided into a plurality of channels, in general, it is divided into two or more channels, so that each channel is The image sensor 201 includes a plurality of pixels. Therefore, when the image sensor 201 scans the image, the image is divided into the same number of complex array analog signals according to the number of channels, and respectively transmitted to the plurality of image sensors connected to the image sensor 2〇1. The analog front end 202 has the same number of channels as the number of channels. When the analog front end receives the complex array analog signal from the image sensor 201, the analog front end shifts each group analog signal into the first digital signal value output. The towel includes amplifying the value of the signal. And after each analogy _ 2 〇 2 is connected - the correction mode is implemented, and the correction module 206 is configured to receive the first digital signal value outputted from the various types of front end movements, wherein each correction module is corresponding to each channel There are at least one look-up table, and each-view table 2〇7 includes a plurality of digital brightness readings and a plurality of digital correction values, and the digital party readings correspond to the digital correction values. Therefore, the look-up table 207 of the correction module is based on the first digital signal value output from the analog front end and the reading value for the plurality of digital brightnesses. When the digital brightness reading value and the first digital signal value are scaled And output a digital correction value corresponding to the digital luminance reading equal to the digital signal value. Finally, refer to the digital correction value input of the 2G7 output to the gamma correction ϋ 2G8 to signal the further step-by-digit correction value. June 23, 23, replace the page comparison and the brightness check output - touch video material. / Therefore, this creation provides a plurality of correction modules to be connected to the analog front-end movement, and the problem of multi-channel reduction output imbalance that cannot be solved by the known technology... and the method of establishing the image output correction device look-up table will be described in detail below. . Please refer to "Figure 3" and "Picture 6". "Grade 3" is the τ/image output of the creation of the creation; the schematic diagram of the method of setting the table, and the "Picture 6" is the Establish a flow chart of the method like the output correction device lookup table. For example, in the "3rd figure", first, the general gain method is used to adjust the gain value (Gain) of each channel in the multi-channel camera and the offset of the signal (〇ff (4), respectively, to amplify the signal. And to eliminate the noise attached to the signal, and try to adjust the brightness of the output woven to the connected t帛3 picture" from left to right as a white to black grayscale image 21〇〇 then 'will-image sensing The apparatus 201 scans to a grayscale image 21〇, and the gray 1^white IV image 210 is composed of a plurality of gray scales. In the present embodiment, the grayscale image 2 is composed of 24 grays, and this embodiment The image sensor period used in the system is the bit output 'so it can be parsed into 256 gray scale output, and the gray scale image 21〇 24 gray scale can be evenly distributed in 24 gray scales of 256 gray scale. As shown in the "figure 6", each pixel of each channel of the multi-channel camera image H 201 is sequentially scanned to the gray P edge image 21 (four) for each gray scale (step _, and each In each gray scale, each pixel can obtain a first digital signal value (steps are clear. And each pass will pass all the pixels of the towel in the same gray scale The plurality of first-digit signal values are taken as the representative values of the party (step 130), and each channel is scanned on the grayscale image (10), and the pixels of the grayscale can be averaged in the gray level of 24 p. The first digit signal value M395324 replaces the page on June 23, 1999 or takes the median value of the first digit signal value of all pixels to obtain the % brightness representative value. Therefore, each channel scans the grayscale image 21 Obtaining 24 brightness representative values ' can be plotted according to the correspondence between the brightness representative value and the gray level. Then, in each gray level, 'based on the plurality of channels in each gray scale brightness representative value definition - digital correction value (step 14〇) 'When all gray levels are defined as a set of digital correction values (step 150)' then each channel towel has a positive value for each brightness, and the brightness of each channel is recorded for all gray levels. The system is - the brightness reading value (step sub), and the 24 brightness representative values and the 24 position correction values corresponding to the 24 gray levels respectively will be any two brightness values (_ brightness reading value) coffee and brightness generation Green (resolved by the brightness of the reading point, the continuous integer inserts the brightness representative value (also Digital brightness reading) and using the internal ^ method and the external H-hybrid operation to obtain the positive value of the number of money (step I%), the most correct to establish the investigation of each channel, the survey _ including the complex number of her position brightness reading And the corresponding one is her school (step 18〇). The following explains the calibration module with the _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ And "Fig. 4B", =" is the list of the brightness representative values obtained by scanning each channel of the image rotation correction device of the creation. Fire White Fire M395324
99年6月23日替換頁 <N to (N O OJ ir·) Ά 254 i〇 240 oo tn (N CN m ΟΪ 244 232 m (N 229 Ol 艺 OJ CN 234 CN (N 224 216 O m o 口 222 210 CO 208 m (N CN On Os 210 ON Os s OO σ\ s 03 <N ON o OO <N (N OO (N OO oo C^) 〇\ OO 卜 S «ΗΗ 卜 CN vo oo r- m v〇 〇 CN ^O On Ό Ό l〇 ΙΟ ^―Η v〇 ON 寸 〇 寸 r- tn On Ο ON ΓΛ ΓΛ Ό m On m i-H m OO (N <N m CN 卜 (N ·—< r- — <Ν m CN v〇 Ό OJ GTs ON 〇^ VO ¢^ 5； 〇> m ¢^ ON oo CN OO CN On oo Ό OO CN OO jn os Ό § |N jn σν S OO σ\ v〇 00 m 00 尝 o Ό m VO m (N ΓΛ oo ΓΛ m (N Cv) OO 04 Ό o … 卜 卜 上椒1 、<_»/· ㈣s _ 第三通道之 亮度代表值 (數位亮度讀值） s ^ 1 料5 's—X 第一實施例之 數位校正值 ί 99年6月23日替換頁 而「第4Β圖」係為本創作之多通道相機就24灰階解析之亮 度代表值_灰關顧。由於本實關〇通道減為8位元輸 出’因此可將均勻之灰階影像之24灰階依256灰階度均勻解析， 即灰階影像之第-灰階係為多通道相機之第u灰階度，而灰階影 像之第二灰階係為純勒機之第21灰时，以賴推，則就亮 度代表值及灰階度之對細係作圖可以依各通道得聽灰階度增 加而產生亮度代表值的分佈，如「第4β圖」所示，其中不同標號 刀別代表不同通迢之冗度代紐的分佈。而就本創作第一實施例 來說，於每—灰階度中依據四個通道所具有的四個亮度代表值定 義數位权正值之方法係為取每一灰階度中四個亮度代表值中一 最小亮度代表值為數位校正值，如「表L及「第4A圖」中，於 第75灰階度中四通道之亮度代表值分別為八卜μ、μ及μ , 則在第75灰階度巾，取最小亮度細^為數位校正值，以此 類推。 :然而由於本實施例之多通道相機可具有攻灰階度輸出，而 所知描之灰階影像僅具有24灰階則均勻分佈於攻灰階度中，因 t如「第4Α圖」中’同—通道中任相㈣點之_亮度代表值 數位校正值之對應關係，則利用内插法來進行估計，於本實施 例中則利用線性函數來進行内插法^因此，請參閱「表：」，「表2」 係為本創作第一實施例中利用内插法而得之查閱表。 12 卯年6月23日替換頁 卯年6月23日替換頁 第—通道之 兔度ΗΛ值 ]〇 ]] n (¾位亮度讀值） 12 13 ]4Replacement page on June 23, 1999 <N to (NO OJ ir·) Ά 254 i〇240 oo tn (N CN m ΟΪ 244 232 m (N 229 Ol Art OJ CN 234 CN (N 224 216 O mo port 222 210 CO 208 m (N CN On Os 210 ON Os s OO σ\ s 03 <N ON o OO <N (N OO (N OO oo C^) 〇\ OO 卜 S «ΗΗ 卜 CN vo oo r- mv〇〇CN ^O On Ό Ό l〇ΙΟ ^―Η v〇ON inch inch r- tn On Ο ON ΓΛ ΓΛ Ό m On m iH m OO (N <N m CN 卜(N ·—< R- — <Ν CN m CN v〇Ό OJ GTs ON 〇^ VO ¢^ 5; 〇> m ¢^ ON oo CN OO CN On oo Ό OO CN OO jn os Ό § |N jn σν S OO σ\ V〇00 m 00 t t o VO m VO m (N ΓΛ oo ΓΛ m (N Cv) OO 04 Ό o ... 卜布上椒1, <_»/· (4) s _ third channel brightness representative value (digital brightness Read value) s ^ 1 material 5 's - X digital correction value of the first embodiment ί June 23, 1999 replacement page and "4th map" is the brightness of 24 grayscale analysis of the multi-channel camera of the creation Representative value _ gray care. Because the actual channel is reduced to 8-bit output 'so can be The 24 gray scale of the uniform gray scale image is uniformly analyzed according to the 256 gray scale, that is, the first gray scale of the gray scale image is the u gray scale of the multi-channel camera, and the second gray scale system of the gray scale image is pure When the 21st ash of the machine is used, the mapping of the brightness representative value and the gray level can be performed according to the increase of the gray scale of each channel, such as the "4th figure". As shown in the figure, the different labeling knives represent the distribution of the redundancy of the different tongs. For the first embodiment of the present creation, the four radiances of the four channels are represented in each gradation. The value of the positive value of the digital weight is determined by taking the minimum brightness representative value of the four brightness representative values in each gray level as the digital correction value, as shown in Table L and Figure 4A, in the 75th gray scale. The brightness representative values of the four channels in the degree are eight-bu, μ and μ, respectively, and in the 75th gray-scale towel, the minimum brightness is taken as the digital correction value, and so on. However, since the multi-channel camera of the embodiment can have a gray-scale gradation output, and the known gray-scale image has only 24 gray levels, it is evenly distributed in the grading gradation, because t is in the "4th drawing" In the same channel, the corresponding relationship between the brightness representative value and the digital correction value is used for the estimation. In the present embodiment, the linear function is used for the interpolation method. Therefore, please refer to " Table: "Table 2" is a look-up table obtained by interpolation in the first embodiment of the creation. 12 Replacement page on June 23 of the following year Replacement page on June 23 of the following year - The value of the rabbit in the channel - 〇 ]] n (3⁄4 digit brightness reading) 12 13 ] 4
11 12 12 b ，6 17 *8 19 20 21 22 13 14 15 15 表211 12 12 b , 6 17 *8 19 20 21 22 13 14 15 15 Table 2
第一通道之 ‘ 值 〇 K數位亮度讀值)『 产代ΙΓν」所Γ ’第1道於灰階_相鄰兩灰階得到之亮 別？〇單位及22單位，而所對應之數位校正值分 早立及16早位’即以連續整數插入兩亮度代表值間，再利 ^線性函數進㈣插法取得對應之數位校正值，其中由於所得之 為數4值□此僅以整數儲存，而小數點以下數值則以四捨 五入法取捨，則可得如「表2」所示之列表。此外，於「第4Α圖」 中各通逼之起始端點，如第U灰階度之亮度代表值則亦需於端 點外以連續整數插人亮度代表值，並侧線性練之外差法來進 行數減正值之估計，因此，請參閱「表3」，「表3」係為本創作 第-實施例中利用外插法而得之查閱表。如「表3」所示，第一通 道於灰階影像之第一灰階得到之亮度代表值係為10單位，而所對 13 Z數純正_為7單位， 數化之數位校正值，_的，小數點町數值取四捨五入 去取捨，則可得如「表3」所示之列表1此類推則可得到第—通 道對應於所有絲度之查閱表。的，就不_通道可以相同 之方法與料方式而得到各自的細表。_本創作並不只限定 _用線性函數來進行内插法及外插法，亦可__非線性函數 來進行_法及-線性函數來進行外插法運算，亦可彻一線性 函數來進行_法及-雜性函數來進行外插法運算，或是利用 —非線性函數來進行内插法及—非線性函數來進行外插法運算。 請參閱「第4B圖」，「第4B圖」係為本創作第—實施例:影 像輸出校正裝置的亮度代表值.灰階度關係圖。則如「第犯圖” 所不’本實關之具有四通道之多通道械，其每—通道皆具二 —亮度代表值·紐度之對應曲線，其分別為第—通道係之第二對 應曲線101、第二通道之第二對應曲線1〇2、第三通道之第三對應 曲線103、以及第四通道之第四對應曲線1〇4，此外，於「第^ 圖」中包含有一數位校正值曲線1〇〇a，即在圖示中以粗累線表示， 為所有對應曲線之最小值的邊緣輪靡，而其數值即為本創作l 實施例之數位校正值。而其中每—亮度代表卿為—數位亮度讀 值。舉例來說’當第-通道之任—晝素讀取_影像而得到一數位 亮度讀值’係例如「表2」中之數位亮度讀值為16單位，則經由 屬於第-通道之查查酿，則以12單位之數位校正值輸:， 而於第4B圖」中’即為第-對應曲線1〇1上之數位亮度讀值為 14 M395324 〇〇 Lj9年6月23^替換苜 16單位的點沿錯垂線向下調整至數位校正值曲線腿上的數值， 即為Π單位。而以上校正之步驟，可以消除因影像感測器之製程 差異以及各通道之獨立電路所造成各通道中輸出之亮度不均句的 問題。 本創作第-實施例為—最佳實施例，係、由於當取同_灰階度 依據四通道得狀四個亮度代紐巾—最小亮度代紐為數位^ 正值時’㈣實驗數據與理論的_，以上述方法取最小健统 計可以得勝!、的鮮錄(sta論仙㈣⑽，為統代表值之 分佈中一最佳數位校正值之選取方式。 —然而，本創作之第二實施例之裝置及主要方法與本創作第一 貝％例相同’其主要的差異在於每—灰階度巾就複數個亮度代表 值定義-數位校正值之方法不同。請參閱「第5圖」，「第5圖」 係為本創作第二實施例之影像輸出校正裝置的亮度絲值-灰階度 關係圖。如「第5圖」所示，—多通道相機包含有四通道，且每 -通道皆具有—亮度代表值·灰階度之職鱗，其分縣第一通 道，第-對應曲線ΗΠ、第二通道之第二對應曲線搬、第三通道 之第三對應曲線1G3、以及第四通道之第四對應曲線刚。而其中 於同-灰階度依據四個通道所得到之四個亮度代表值中定義二數 位校正值之耗係為取四個亮度代紐之—平均值紐位校正 值即在第5圖」中，於第75灰階度之平均亮度代表值μ， 即為亮度代表值A1、A2、A3及A4的平均值。因此，於「第$ 圖」中以-粗黑曲線表示其數位校正值之分佈而得到一數位校 15 M395324 L99 月23日替換頁 值曲線ι_。_ ’每—通道就所有灰階度亦可得—查閱表， 其具有複數個數位亮度讀值及相對應之數位校正值。—、 然而，本創作並不只限定於第—實酬及第二實施例的情況 中’亦可以於各灰階度中取―最大亮度代表值為數位校正值，或 者任何適當取得參考值之方㈣可祕本創作中， 所保護之細。 本創齡要在贿立影像輸岐正裝置朗表衫通道相機 :出的亮度不平衡之問題。並且本創作可以翻於單色及彩色相 依據其各通道之亮赖校正。而姻蘭應於各像素之查閱 义，可以使得因為影像制器之像素製程品f不佳，衫通道相 機之各通道由於其具痛立線路聪組成線路之各元件有里差里 性而使得各通道影像訊賴出的亮度不均勻的問題難曹 輪出類比前端時料由查職將各魏之亮度輕至數位校正U 值，使各通道輸出的亮度均勻一致。 賴本創作以前狀實施例揭露如上，然其_以限定本 =τ本創作之精神和範圍内，所為之更動與潤飾，均 二本創:之專利保護範圍。關於本創作所私之保護範圍請 所附之申請專利範圍。 【圖式簡單說明】 第1Α圖係為1知技術中之多通軸機之系統示意圖； 圖係為利用習知技術之多通道相機產生之影像資料畫 面， 16 M395324 斤λ门/ 99年6月2^日## =圖圖係為本創作之影像輪出校正裝置; 第3圖係為本緩之建立影像如校正裝置細 意圖； 忐不The first channel's ‘value 〇 K digit brightness reading value 『 产 代 ΙΓ 」 Γ ’ ’ ’ ’ ‘ 〇 unit and 22 units, and the corresponding digital correction value is divided into early and 16 early positions', that is, a continuous integer is inserted between the two brightness representative values, and then the linear function is entered into the (four) interpolation method to obtain the corresponding digital correction value, wherein The resulting value is 4, which is only stored as an integer, and the values below the decimal point are rounded off to obtain a list as shown in Table 2. In addition, in the "4th map", the starting end points of each pass, such as the brightness representative value of the U gray scale, also need to insert the brightness representative value with a continuous integer outside the endpoint, and the side linear training difference The method is used to estimate the positive value of the subtraction. Therefore, please refer to "Table 3", which is the look-up table obtained by the extrapolation method in the first embodiment of the present invention. As shown in Table 3, the first channel in the first gray level of the grayscale image has a brightness representative value of 10 units, and the 13 Z number is purely 7 units, and the digitized digital correction value, _ If the decimal point value is rounded off, you can get the list as shown in "Table 3". Such a push can obtain the reference table corresponding to all the filaments. If you don't, the channel can get its own detailed table by the same method and material method. _ This creation is not limited to _ linear function for interpolation and extrapolation, __ nonlinear function for _ method and - linear function for extrapolation, or a linear function The _method and the heterozygos function are used for extrapolation, or the interpolation function and the nonlinear function are used to perform extrapolation. Please refer to "Fig. 4B" and "Fig. 4B" for the creation of the first embodiment: the brightness representative value of the image output correction device. The gray scale relationship diagram. For example, if the "figure map" does not have a four-channel multi-channel weapon, each channel has a second-brightness representative value and a corresponding curve of the new degree, which is the second of the first channel system. Corresponding to the curve 101, the second corresponding curve 1〇 of the second channel, the third corresponding curve 103 of the third channel, and the fourth corresponding curve 1〇4 of the fourth channel, and the The digital correction value curve 1〇〇a, which is represented by a thick line in the figure, is the edge rim of all the minimum values of the corresponding curve, and its value is the digital correction value of the embodiment of the present invention. And each of them - the brightness represents the - digital brightness reading. For example, 'when the first channel is used—the 昼 读取 reading _ image to obtain a digital brightness reading value ′ for example, the digital brightness reading value in “Table 2” is 16 units, and then the In the case of brewing, the digital value of 12 units is used for the correction value: and in the 4B picture, the digital brightness reading value on the first-corresponding curve 1〇1 is 14 M395324 〇〇Lj9, June 23^ Replace 苜16 The point of the unit is adjusted downward along the wrong vertical line to the value on the leg of the digital correction value curve, which is the unit of Π. The above correction steps can eliminate the problem of uneven brightness of the output in each channel caused by the difference in the process of the image sensor and the independent circuit of each channel. The first embodiment of the present invention is a preferred embodiment, because when the same gray gradation is based on four channels, four brightness generations are used, and the minimum brightness is a digital positive value. (4) Experimental data and The theoretical _, taking the minimum health statistics by the above method can win!, the fresh record (sta narration (4) (10), is the selection method of the best digital correction value in the distribution of the representative values. - However, the second implementation of this creation The device and main method are the same as the first example of the creation. The main difference is that the method of defining the brightness representative value-digit correction value is different for each grayscale towel. Please refer to "figure 5". "Fig. 5" is a graph of the relationship between the brightness and the gray scale of the image output correction device of the second embodiment of the present invention. As shown in "Fig. 5", the multi-channel camera includes four channels, and each- The channel has the scale of the brightness representative value and the gray scale, and the first channel of the county, the first-corresponding curve ΗΠ, the second corresponding curve of the second channel, the third corresponding curve of the third channel 1G3, and the first The fourth corresponding curve of the four channels is just. The same-gray gradation is based on the four brightness values obtained by the four channels. The value of the two-digit correction value is taken as four brightness-based values. The mean value correction value is in Figure 5, The average luminance representative value of the 75th gray scale is the average value of the luminance representative values A1, A2, A3, and A4. Therefore, the distribution of the digital correction value is represented by the thick black curve in the "$ map". Get a number of schools 15 M395324 L99 on the 23rd of the month to replace the page value curve ι_. _ 'every channel can be obtained for all gray levels - look up the table, which has a plurality of digital brightness readings and corresponding digital correction values. However, the present invention is not limited to the case of the first embodiment and the second embodiment. The maximum brightness representative value may be taken as a digital correction value or any suitable reference value (4). In the creation of the secret book, the protection is fine. The age of the creation of the image is to be used in the image of the camera. The brightness of the camera is unbalanced. And the creation can be turned into monochrome and color according to its The correction of each channel depends on the correction. The pixel look-up meaning can make the channel processing products of the image-making device f poor, and the channels of the shirt channel camera are caused by the various components of the circuit line composed of the painful line. The problem of uneven brightness is difficult. The analogy of the front end is based on the fact that the brightness of each Wei is lightly adjusted to the digital U value, so that the brightness of each channel output is uniform. The previous embodiment of Lai Ben's creation reveals the above, but _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Explanation] The first diagram is a schematic diagram of the system of the multi-axis machine in the prior art; the picture is the image data generated by the multi-channel camera of the prior art, 16 M395324 kg λ door / June 2, 1999 ##=图图 is the image rotation correction device of the creation; the third picture is the intention to establish the image such as the correction device;
第4A _為本解之影物校正裝置就24灰階解析之 度代表值·灰階度關係圖；4A _ the image correction device of the present solution is a gray scale analysis representative value and gray scale relationship diagram;
第4B圖係為本創作第一 灰階度關係圖； 實施例之多通勒機的亮度代表值 第5 _為係為本創作第二實施例之多通道相 值··灰階度關係圖；以及 第6圖係為本創作之建立影像輪出校正装置查 程圖。 ~~ 【主要元件符號說明】4B is the first gray scale relationship diagram of the creation; the brightness representative value of the multi-passer machine of the embodiment is 5th_ is the multi-channel phase value·· gray scale relationship diagram of the second embodiment of the creation And Figure 6 is a map of the establishment of the image wheel rotation correction device for this creation. ~~ [Main component symbol description]
機的亮度代表 閱表的方法流 10 ' 201 影像感測器 20、202 類比前端 30、208 伽瑪校正器 40 影像資料畫面 41 暗線 42 明線 l〇〇a、l〇〇b 數位校正值曲，線 101 第一對應曲線 102 第二對應曲線 M395324 ,_ 99年6月23日替換頁 103 第三對應曲線 104 第四對應曲線 200 多通道相機 206 校正模組 207 查閱表 210 灰階影像 A 卜 A2、A3、A4 亮度代表值 A5 平均亮度代表值The brightness of the machine represents the flow of the reading method 10 ' 201 image sensor 20, 202 analog front end 30, 208 gamma corrector 40 image data screen 41 dark line 42 open line l〇〇a, l〇〇b digital correction value song Line 101 First Correspondence Curve 102 Second Corresponding Curve M395324, _June 23, 2007 Replacement Page 103 Third Corresponding Curve 104 Fourth Correspondence Curve 200 Multi-Channel Camera 206 Correction Module 207 Lookup Table 210 Grayscale Image A Bu A2, A3, A4 brightness representative value A5 average brightness representative value