l232〇56 狄、發明說明 【發明所屬之技術領域】 本發明係有關於一種影像輸出裝置，特別是與一種具影 像補償功能之影像輸出裝置有關。 【先前技術】 電荷搞合裝置（Charge Coupled Device，CCDs)和互補式 =氧半場校電晶體（CMOS )，常使用在掃瞄器的掃瞄頭或 疋數位式相機上作為受光體。其中CCDs在表面感受到光線 時，會將電荷反應在元件上，當整個CCDs上的所有感光元 件^產生的訊號，就構成了一個完整的畫面。而CMOS主 要=利=矽和鍺這兩種元素所做成的半導體，使其在CM〇s “ =页負電和帶正電級的半導體，這兩個互補效應所產 的電机即可被處理晶片紀錄並解讀成影像。因此當這些受 :體本身具有缺陷，亦極具有損壞畫素（bad pixel)時，將 會影響到影像品質。 f CCD A㈣解决方法是先將損壞畫素的位址記錄下來，在 上：記隸由0，斤記錄的影像要被讀出時， ㈣平^ 其相鄰畫素的平均記錄值所取代，亦即使用 兩點千均，方法進行損壞晝素之補償。 參閱第1圖所示為傳統上影像輪屮^ ^ 由受光體讀取時，此資料像輸出處理糸統，當—資料 訊影響過濾掉，再經過pR 枚正102，先將雜 異最後再將影像_出A 又正1 〇4，消除受光體間之差 輪出。其中晝素補償記憶體HH)係用來同牛 1232056 紀錄於畫素檢測後損壞畫素之位址。當影像資料傳入時，損 壞晝素補償區塊11 0會將所輸入之影像資料，不論是壞的查 素或是好的畫素資料均進行補償，補償後之影像資料再傳^ 到輸出選擇區塊120，而另一方面未經補償之影像資料亦會 經由一延遲模組130傳送到輸出選擇區塊12〇。此時輪出^ 擇區塊120即會根據畫素補償記憶體1〇〇中所記錄之損壞^ 素之位址’選擇輸出由延遲模組13()傳送之未經補償影像; 料或是由損壞畫素補償區塊丨丨〇補償後之影像資料。在上_ ::::處理中，需要一額外之畫素補償記憶體ι〇〇來記錄 置’如第1圖所示傳統方式之缺點是需要複雜 轾序及額外記憶体。 文後雜 是壞方式則是將所輸入之影像資料，不論 此方式並不需二二畫f之資：均進行補償後輸出，雖然 像資料& ^ ^之畫素補償記憶體，但由於所有之影 像貝枓均經過補償，雖然解決損壞晝素問顳一： 之景多像資料I ^ t ^ 、 也k成所輸出 广原有之影像資料間存在一定之差異。 如何在不增加額外記憶 晝素，並對其圮姅> ρ π ~ 之1f况下，圮錄損壞之 錄之㈣資料進行補償，即成為改進之重點。 【發明内容】 損壞的即是在提供-種影像輸出裝置，具有 本發明提升輸出影像之品質。 用額外之記憶體來 】出凌置，並不需要使 己錄知壞晝素，因此可降低成本。 1232056 本發明再一目的為提供一 外記憶體來記錄損壞晝素之/輸出裝置，在不使用額 償來提升輸出影像品質’。、、況下，可針對損壞畫素進行補 根據上述之目的，本發明提供 包括一 FPN校正表， 種衫像輸出裝置，至少 保存有在建立FPN护 壞畫素之資料，-PRNU校正I 乂正表日守判斷為損 正表時判斷為損壞書辛。去, 子有在建立PRNU校 會先藉由-校正—表和二=受 示），接著如第3圖所示—損 進^扠正（如第2圖所 後之影像資料進行補償，1出旦員區塊會將所有校正 正表和FPN校正表中^X擇區塊會根據刪U校 經補償之影像資料或是補償後之資料貝枓，選擇輸出未 【實施方式】 在不限制本發明之精神及 實施例，介紹本發明之實扩.二觀圍之下’以下即以-發明之精神後，當可岸用太 埝技*者，在瞭解本 ⑽中。由於傳統之影像輸出褒置 同之衫像產 來對損壞畫素記錄以待後續 ：要求一額外之記憶體 rr便續補彳貝，要不就是對 全面性補償，上述兩者一要七、& 疋對所有晝素進仃 σ ^ ^要求額外記憶裝置，一爭塑於屮與 像品質’因此本發明即是 1〜響輸出衫 像輸出裝置，俜利用值试+進上述缺點。根據本發明之影 FPN和prNU校正表f 、书曰便用 杆於X丄 （eCtl〇n table)來對不平均之影像進 仃扠正，本發明即是在對為 丁 J <〜诼進 、又先7L件進行校正之同時，將偵測 1232056 到異常之受光元件_供$你—< 干併5己錄在FPN和？1^1;校正記錄 如此，在做影像補償時，即 记錄表中， _進行補償，而不需額外使用一記憶體。…素之 每個文光體在相同的光線照射下，應該有相 出，然而由於受光體本身於 二 °之輪 產、使用過程及材料之差里初 或與產品之組裝生 材科之差異都可能造成bad pixel產生，甚去 =處理將影響影像品f會造成對光反應彼 = 光體對光之反應與整體平均之差里 L個別文 , J冬差/、稱為像素反應不一勒 …〇to ReSponse N〇n-Unif〇rmity，pRNu)，而另一方 受光體亦會受到來白冰臾 ’ 卜界之衫響而造成讀取影像時產生雜 说，此稱為固定型式雜^ · ” 士，上、十、不二 ㈣Ν〇1%卿），一般而 0 上述兩值應該越低越好。 “般為了解決光源、鏡頭及受光體本身特性等因辛所 1式雜錢像素反應4-狀情形，會進行如第 圖所不之平面校正法（F1 1 at fleld Correctlon)，利用減法和 ^ 1對母一個畫素進行F_pRNu校正。在進行校正 =儲二先建^卿和刚11校正表，其中卿和PRNU校正表 於—記憶體中’此記憶體可為—隨機存取記憶體，例 能:糟由，描一張黑紙或遮蔽鏡頭使受光體處於不受光狀 二建立每一個受光體受到雜訊之影響之FPN表，在正常情 =I黑、紙之掃描值應為0，但受雜訊影響造成每個受光體 髀盥：：〜127。另外藉由掃描一張白紙，建立每-個受光 2、 “丨值之PRNU表，在正常情形下白紙之掃描值應 ^ ，但由於受光體本身材料之差異，造成每個受光體之 1232056 PRNU值介於128〜255。因此當一資料由受光體讀取時’此 資料會先經過FPN校正202，先將雜訊影響過渡掉’再經過 PRNU校正204，消除受光體間之差異。 本發明是在做受光元件校正之同時’增加受光元件之 判斷，並將偵測到異常之受光元件一併記錄在FPN和PRNU 校正記錄表，便不需額外之記憶體。例如，在正常情況下， FPN校正表是藉由掃描一張黑紙所建立，因此表中每一個值 應介於0〜127間，若有一值超過127，即可斷言此為損壞畫 素，因此本發明之方法，會在原來之FPN校正記錄表記錄下 此損壞晝素。同樣的，PRNU校正記錄表是藉由掃描一張白 紙所建立，因此表中之每一值應介於128〜255間’若有一 值低於128，即可斷言此為損壞畫素，因此本發明之方法， 即會在原來之PRNU校正記錄表中記錄下此損壞畫素，供後 續作晝素補償。藉由將損壞畫素之位址記錄在PRNU和FPN 校正表中，即可不使用額外之記憶裝置。值得注意的是，上 述判斷損壞畫素之關鍵值，如FPN值之127和PRNU值之128 可由使用者根據所要求之影像品質更改，而記錄損壞晝素之 方法，例如可將原來掃描到之FPN值和PRNU值置換為0儲 存，來代表此畫素為損壞，然，亦可使用其他之值代表。 參閱第3圖所示為本發明根據FPN和PRNU校正表之内 容決定其輸出之影像輸出裝置。當一資料由受光體讀取時， 此資料會先作一平面校正，藉由一 FPN校正表302，將雜訊 過濾掉，再經過PRNU校正表304，消除受光體間之差異， 其中FPN校正表中儲存有在建立FPN校正表時判斷為損壞 1232056 旦素之貝枓，而PRNU校正表 JL # M ^ 4 门樣儲存有在建立PRNU校 正表日守d斷為損壞畫素之資料。 & — 入德，指捸佥i 凡父後之影像資料輸 、展旦素補你區塊3 10會將所有校正後 不論是由壞的畫素或是好的書辛 和象、’、 玛後眘+4· i # , —京貝枓均進仃補償，補償後之 和像貝枓再傳送到輸出選擇區 之影像資料會經由-延遲模电33〇值而另—方面未經補償 碟，mm ^ ^㈣未經補償之影像資料進行延 ^出擇之影像f料和補償後之影像資料同時傳輸 :L= 此時輸出選擇區塊320會根據_校 遲ί:3:ί表中所記錄之損壞晝素資料，選擇輸出由延 =二Τ之未經補償影像資料或是由損壞畫素補償區 塊3 1 0補彳員後之影像資料。 辛之:方法，並不需要額外記憶體來記錄損壞畫 素之位址，且根據PRNU校正表和fpn校 ::，了素j料來決定需不需要做補償，因此不會二^ 造YfM象失7 =之使用而將所有資料均進行補償，反而 心 /、再加以，本發明是使用原來PRNU校正矣 和FPN校正表進行記錄，並不會改變原本之架構。 雖d本發月已以一較佳實施例揭露如上，然其並 限定本發明，任何熟習此技藝者，在不 範圍内，當可作各種之更動盥列说m L ^ *月之精神和 4德.由-奎Ϊ 發明之保護範圍 田見後附之申。月專利範圍所界定者為準。 【圖式簡單說明】 10 1232056 為讓本發明之上述和其他目的、特徵、和優點能更明顯 易懂，下文特舉一較佳實施例，並配合所附圖式，作詳細說 明如下： 第1圖為傳統影像處理裝置； 第2圖為根據本發明之影像資料校正裝置；以及 第3圖為依據本發明較佳實施例的影像處理裝置。 【元件代表符號簡單說明】 100晝素補償記憶體 110和3 10損壞晝素補償區塊 120和320輸出選擇區塊 1 3 0和3 3 0延遲模組 102、202 和 302 FPN 校正表 104、204 和 3 04 PRNU 校正表 111232. Description of the invention [Technical field to which the invention belongs] The present invention relates to an image output device, and more particularly to an image output device with an image compensation function. [Prior art] Charge Coupled Devices (CCDs) and complementary = oxygen half-field correction crystals (CMOS) are often used as the light receiver on the scanning head of a scanner or a digital camera. Among them, when the CCDs feel the light on the surface, they will reflect the charge on the element, and the signals generated by all the photosensitive elements on the entire CCDs form a complete picture. CMOS is mainly a semiconductor made of two elements, silicon and germanium, so that it can be used in CM0s = = semiconductors with negative and positive charge levels. The motors produced by these two complementary effects can be used. Process the wafer record and interpret it as an image. Therefore, when these receivers have defects and bad pixels, they will affect the image quality. F CCD A㈣ The solution is to first set the bits of the damaged pixels. The address is recorded on the top: when the recording image is read by 0, the average recorded value of its adjacent pixels is replaced by ㈣ping ^, that is, two points per thousand are used to damage the day pixels. Refer to Figure 1 for the traditional image wheel. ^ ^ When read by the receiver, this data is like the output processing system. When the data is filtered out, it will be filtered by pR before 102. At the end of the difference, the image_out A is positive 104 again, eliminating the difference between the photoreceptors. Among them, the day pixel compensation memory HH) is used to record the address of the damaged pixel after the pixel detection. When the image data is transmitted, the damaged day element compensation block 110 will input the The image data, whether it is a bad pixel or good pixel data, will be compensated. The compensated image data will be transmitted to the output selection block 120. On the other hand, the uncompensated image data will also pass a delay. Module 130 is transmitted to the output selection block 12. At this time, the rotation of selection block 120 will compensate for the damage recorded in the memory 100 according to the pixel ^ The address of the pixel 'selection output is delayed by module 13 () Uncompensated image transmitted; material or image data compensated by damaged pixel compensation block 丨 丨 〇 In the above _ :::: processing, an additional pixel compensation memory is needed. The disadvantage of the traditional method shown in Figure 1 is that it requires complex procedures and additional memory. The miscellaneous method is to use the input image data, regardless of this method. Data: all output after compensation, although the pixel compensation memory like data & ^ ^, but because all the image shells have been compensated, although the damage to the day and time is resolved, the scene is multi-image data I ^ t ^, There is also a gap between the original and exported video data. It becomes the focus of improvement how to compensate for the corrupted recorded data without adding extra memory to the day element and 1 圮 姅 f ρ π ~ 1f. [Inventive Content] Damaged That is to provide an image output device with the present invention to improve the quality of the output image. Using additional memory to output the data does not need to make the recorded information known, so it can reduce costs. 1232056 The present invention again One purpose is to provide an external memory to record the damaged day / output device, without using compensation to improve the quality of the output image '., And, in addition, can compensate for the damaged pixel. According to the above purpose, the present invention provides Includes an FPN correction table, a shirt-like image output device, and at least the information for establishing FPN protection pixels, and -PRNU correction I when the watch is judged to be damaged by the watch, is judged to be damaged. Go, Ziyou establishes a PRNU school conference by first-correction-table and two = received), and then as shown in Figure 3-loss advance ^ fork Zheng (as the image data after Figure 2 to compensate, 1 The member block will select all correction tables and FPN correction tables ^ X. The selected block will be deleted based on the U-corrected image data or compensated data. The output is not selected. [Implementation method] Without restriction The spirit and embodiments of the present invention introduce the practical expansion of the present invention. Below the second view, the following is the spirit of the invention, when you can use the Taiji technique *, understand this book. Because of the traditional video The output of the same shirt is produced to record the damaged pixels for subsequent follow-up: if an additional memory rr is required, it will be replenished, or it will be a comprehensive compensation. The above two are seven. &Amp; All daytime photosets require additional memory devices, which strive to shape the image and image quality. Therefore, the present invention is a 1 ~ 4 output shirt image output device, which uses the value test to add the above disadvantages. According to the shadow of the present invention FPN and prNU correction table f, Shu Yue will use the rod X 丄 (eCtl0n table) to unequal The image is straightened. The present invention is to correct the light receiving element that is 1232056 to the abnormality at the same time when it is calibrated and then 7L, it will detect 1232056 to the abnormal light-receiving element_for $ 你 — < dry and 5 It has been recorded in FPN and? 1 ^ 1; the correction record is so, when image compensation is performed, that is, in the record table, _ is used for compensation without using an additional memory .... Each of the literary bodies is in the same Under the light, there should be phase out, but because the photoreceptor itself is in the production of two degrees, the process of use and the difference between the material or the difference between the product and the biomaterials assembly may cause the bad pixel, even = processing Will affect the image quality f will cause the response to light = the difference between the light body's response to light and the overall average, the difference between the individual text, J winter difference /, called pixel response heterogeneity ... 〇To ReSponse Non-Unif〇 rmity, pRNu), while the other receiver will also be subject to the sound of Bai Bing'an's shirt, which will cause noise when reading the image. This is called a fixed pattern. (㈣Ν〇1% 卿), generally 0 and the above two values should be as low as possible. "Generally in order to solve the light The characteristics of the lens, the receiver, and the characteristics of the light receiver, such as the type 1 miscellaneous pixel, show a 4-shape situation, and the plane correction method (F1 1 at fleld Correctlon) shown in the figure will be performed. F_pRNu correction. After calibration = Chu Erxian and Qing 11 correction tables, where Qing and PRNU correction tables are in -memory 'this memory can be-random access memory, for example: bad, draw a black Paper or shielding lens keeps the receiver in a state of no light. Establish an FPN table where each receiver is affected by noise. Under normal conditions = I black, the scanning value of paper should be 0, but each receiver is affected by noise. Body wash toilet :: ~ 127. In addition, by scanning a piece of white paper, create a PRNU table for each light-receiving 2, "丨 value. Under normal circumstances, the scanning value of white paper should be ^, but due to the difference in the material of the light-receiver itself, 1232056 PRNU of each light-receiver The value is between 128 and 255. Therefore, when a data is read by the light receiver, 'this data will first undergo FPN correction 202, and the influence of noise will be transitioned out', and then PRNU correction 204 will be performed to eliminate the difference between the light receivers. It is to increase the judgment of the light-receiving element at the same time as the light-receiving element calibration, and record the abnormal light-receiving element in the FPN and PRNU calibration record table, without additional memory. For example, under normal circumstances, The FPN correction table is created by scanning a piece of black paper. Therefore, each value in the table should be between 0 and 127. If a value exceeds 127, it can be asserted that this is a damaged pixel. Therefore, the method of the present invention will The damaged daylight was recorded in the original FPN correction record sheet. Similarly, the PRNU correction record sheet was created by scanning a blank sheet of paper, so each value in the table should be between 128 ~ 255. At 128, It can be asserted that this is a damaged pixel, so the method of the present invention will record the damaged pixel in the original PRNU correction record table for subsequent compensation of day pixels. By recording the address of the damaged pixel in PRNU And FPN correction table, it is not necessary to use additional memory devices. It is worth noting that the above key values for judging damaged pixels, such as 127 for FPN and 128 for PRNU, can be changed by the user according to the required image quality, and The method of recording the damaged day element, for example, the original scanned FPN value and PRNU value can be replaced with 0 to store this pixel to indicate that this pixel is damaged, but other values can also be used to represent it. See Figure 3 for the illustration Invented an image output device that determines its output according to the contents of the FPN and PRNU correction tables. When a data is read by the receiver, this data is first corrected for a plane, and a FPN correction table 302 is used to filter out noise. Then pass the PRNU correction table 304 to eliminate the difference between the photoreceptors. The FPN correction table stores the shellfish which is judged to be damaged when the FPN correction table is created. The PRNU correction table JL # M ^ 4 is stored in the sample. In the establishment of the PRNU correction table, the date and time of d are broken to the data of damaged pixels. &Amp; — Enter Germany, which means that the image data of the queen ’s father is lost, and the block is filled up. From bad pixels or good book Xinhexiang, ', Mahoushen + 4 · i #, —Jingbei 仃 are all compensated, and the compensated sum is transmitted to the image data in the output selection area. It will be delayed by the value of the delay mode and the value of 33. The other aspect is the uncompensated disc, mm ^ ^ ㈣ uncompensated image data is delayed ^ the selected image f material and the compensated image data are transmitted at the same time: L = at this time The output selection block 320 will choose to output the uncompensated image data with delay = 2T or the compensation block based on the damaged pixels based on the damaged day pixel data recorded in the _School Delay ί: 3: ί table. 3 1 0 Image data after filling up the staff. Xin Zhi: method, does not require additional memory to record the address of the damaged pixel, and according to the PRNU correction table and fpn calibration :: The element j material is used to determine whether compensation is needed, so YfM will not be created. As for the loss of 7 = all the data are compensated for, but instead they are / / added. The present invention uses the original PRNU correction 矣 and FPN correction table for recording without changing the original structure. Although d this month has been disclosed as above with a preferred embodiment, but it does not limit the invention. Anyone skilled in this art can, within the scope, make various changes. The spirit of the month and 4. The scope of protection invented by Germany and Kui Tian is described in the attached application. The monthly patent scope shall prevail. [Brief description of the drawings] 10 1232056 In order to make the above and other objects, features, and advantages of the present invention more comprehensible, a preferred embodiment is given below in conjunction with the accompanying drawings to make a detailed description as follows: 1 is a conventional image processing apparatus; FIG. 2 is an image data correction apparatus according to the present invention; and FIG. 3 is an image processing apparatus according to a preferred embodiment of the present invention. [A brief description of the components' representative symbols] 100 daylight compensation memory 110 and 3 10 damaged daylight compensation block 120 and 320 output selection block 1 3 0 and 3 3 0 delay modules 102, 202 and 302 FPN correction table 104, 204 and 3 04 PRNU correction table 11