201216697 六、發明說明： 【發明所屬之技術領域】 本發明侧於-像處理方法是—轉原始影 行亮度特徵混合與階層合併而得到高動態影像的方法。/ 【先前技術】 隨著科技的蓬勃發展以及現代人生活品質的提昇，數位影像 裝置，如·數位相機(Digital Still Camera)已逐漸成為現代人±_ 必需。隨著研發技狀辭，·械之魏亦㈣地推陳出新， 成為市場主流。 以數位相機為例，當娜影像為高反差的場景(如場景中同時 、有天氣晴朗的天空與陰蔽處的建築物)時，礙於數位相機本身 軟a更體之輸出限制，大部分動態範圍的場景會被犧轉，而無 a =地將整购景的動態(Dy_iefan齡狄來。也就 =說’如「第1A圖」所示’其中影像之高亮度區域(如天空的雲 从、低讀區域(如圖巾下方的大樓)於成像照壯相當朦聽，並 很難以不失真地情況下被呈現在照片中。 '』已知現有技術有根據被拍攝場景的亮度分佈而改變其 :色調映射(tone mapping)的做法。然而此種做法之效果十分 的門題、 種做法亦會造成色彩不自然且過於藝術化(artifact) 照片是、=此之外，當數位相機連續拍攝數張照片時，由於該些 並無法攝’且其巾間隔的時間非常有限。因此，現有技術 士同％景’連續拍攝多張照片的情況，將數張照片以 201216697 色調映射的做法合併。因此，習知的做法不但無法有效得到去龜 術化的照片，同時也增加使用者於操作上的不便。 β 【發明内容】 馨於以上關題’本發贿供—種影像處理方法，藉以將影 像感測器所擷取到的原始影像進行處理，除了能有效保有原始參 像之亮度特徵外，亦考量_免人工化、色彩_失真等因素了 藉以解決前述問題。 本發贿f郷縣理松，胁—触影舰置。影像 處理方法包括： 擷取-原始影像；自原始影像獲得一高亮影像、一中亮影像 與:低祕像；根據高亮、中亮影像與低亮影像決定對應的 二=重陣列、—中亮權重陣列與—低亮權重陣列；以及根據 :t 亮影像、低亮影像與對應的高亮權重陣列、中亮權 重陣列與低雜重陣顺得-高動態影像。 發明，之影像處理方法，可自原始影像獲取對應 偵測程^，之!^影像、中亮影像與低亮影像 =2 :亮度判斷程序，決定其個別所佔之相對權重。最後 =權重加總轉到-張可把砰亮度特徵賴合紗之高動態 詳細的特徵、實作與功效，兹配合圖式作最佳實施例 【實施方式】 5、參‘、、、帛IB ®」’為根據本發明實施例之數位影像裝 201216697 置的結構示意圖。數位影像裝置1〇〇可以是數位相機(Digital Still Camera)、數位攝影機(Digital Video Camera)、網路攝影機 (Webcamcorder)或是整合有數位影像擷取功能之電子產品，如··手 機、個人數位助理(personal Digital Assistant, PDA )等。數位影像裝 置100包括影像感測器1〇2、感測控制器1〇4、微處理器1〇6、影 像訊號處理單元108、編解碼器110、暫時儲存器112、輸入輸出 單元114、顯示引擎單元116與結果儲存器118。數位影像裝置1〇〇 可藉由其内部各個元件(如上述)之分區處理，而將影像顯示於顯示 裝置200上。根據本發明實施例之影像感測器1〇2係以適於數位 相機的影像感測器為例，但並不以此為限◦常見的影像感測器1〇2 可以疋但不限於電何輕合元件(Charge_C0Upied Device，CCD)。 衫像感測器102具有複數個感應像素(pixeis)，以接收景像傳 來的光線，並利用光電轉換而將該景像轉換成對應的影像資料。 為了能得到彩色感應之絲，複數__感應像素被群組成一 濾鏡樣板(Filtering Pattern)。濾鏡樣板可以是並不限於拜耳樣板 (Bayer Pattern，或稱貝爾圖樣）。依據本發明實施例，係為採用拜 耳樣板做為濾鏡樣板之實施例，但其他種類之樣板亦可以用來實 施本發明。 以「第2圖」為例，每四個相鄰的感應像素2心，22&，22说，228 即構成-麟耳樣板20。數姆像裝置⑽即可藉由每一個拜耳 樣板20内之感應像素22R，22Gr，22Gb，22B，接收景像傳來的光線， 並利用光電轉換而娜-對應的原始影像。為便於後續之說明， 以下即就原始影像中之單一拜耳樣板2〇作以下之解釋與說明。 201216697 根據本發明一實施例之感應像素22R，22Gr，22Gb，22B各具有對應 之14位元(bit)’但不以此為限，可視不同之影像感測器1〇2而定， 亦可為12位元、1〇位元、16位元等，以下係以感應像素具有14 位元進行說明。 明參考第3A圖」，係為根據本發明一實施例之影像處理方 法之流程圖。其係適用於前述數位影像裝置1〇〇，其中影像感測器 102於-拜耳魏2〇巾，具有概佩絲素22R,m 並可利贼絲素22R，22Gn22Gb,22B據哺換由—景像傳來的光線 並榻取一原始影像。此影像處理方法包括： 步驟S302 :擷取一原始影像； 步驟S3〇4:自原始影像獲得一高亮影像、-中亮影像與-低 亮影像； 步驟S306 :根據高亮影像、中亮影像與低亮影像決定對應的 -高亮權重陣列、-中亮權重陣列與一低亮權重陣列；及 根據高亮影像、中亮影像、低亮影像與對應的高 重陣列、中亮權重陣列與低亮權重陣列獲得-高動態影像。 其中上述步驟S302至S308係由「笛m因 ’象 所執行。 第1Β圖」中之微處理器106 關於步驟S304 ’請配合參閱「第 .,示斗Α圖」至「第4D圖 為自原始影像獲得一高亮影像、一 圆」係 圖。如「第4Α ®丨所干^ 冗衫像與一低亮影像之示意 之 '、感應像素22^22(^22^ 22B各且右斜廡 14位元，步驟S3G4係為執行—位 ’ ‘、有對應201216697 VI. Description of the Invention: [Technical Field of the Invention] The present invention is a method for obtaining a high dynamic image by the image processing method, which is a combination of brightness feature blending and hierarchical merging. / [Prior Art] With the rapid development of technology and the improvement of the quality of life of modern people, digital video devices, such as Digital Still Camera, have gradually become the modern human standard. With the development of technical tactics, the Wei Wei (four) of the machine has been promoted and become the mainstream of the market. Taking a digital camera as an example, when the image of Na Na is a high-contrast scene (such as a scene with a clear sky and a sheltered building), the digital camera itself has a softer output limit. The dynamic range of the scene will be sacrificed, and no a = ground will be the dynamics of the whole purchase (Dy_iefan age Dilai. Also = say 'as shown in Figure 1A') where the high-brightness area of the image (such as the sky The cloud from the low-reading area (the building below the towel) is quite pleasing in the image, and it is difficult to be presented in the photo without distortion. '』 It is known that the prior art has a brightness distribution according to the scene being photographed. And change it: tone mapping. However, the effect of this method is very important, and the method will also cause the color to be unnatural and too artistic. The photo is, =, in addition, when the digital camera When several photos are taken continuously, it is impossible to take pictures because of the fact that the time interval between the towels is very limited. Therefore, in the case where the existing technicians and the % scenes continuously take multiple photos, the photos of several photos are mapped with 201216697 tone. merge Therefore, the conventional practice can not only effectively obtain the photo of the turtle, but also increase the user's inconvenience in operation. β [Summary of the invention] Xin is in the above-mentioned topic, the bribe supply-image processing method. The original image captured by the image sensor is processed, in addition to effectively retaining the brightness characteristics of the original image, it also considers factors such as _free artificialization, color _ distortion and the like to solve the aforementioned problem. County Lisong, threat-touching ship. Image processing methods include: capture - original image; obtain a highlight image from the original image, a medium bright image and: low secret image; according to highlight, medium bright image and low The bright image determines the corresponding two=rear array, the medium-light weight array and the low-light weight array; and according to: t bright image, low-light image and corresponding highlight weight array, medium-light weight array and low-heavy weight array In the invention, the image processing method can obtain the corresponding detection process ^ from the original image, ^ image, medium bright image and low light image = 2: brightness judgment program, determine its individual occupation For the weights. Finally = weights plus totals to - Zhang can be the high dynamic detailed features, implementation and efficacy of the 砰 特征 特征 特征 , , , , , , , , , 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳 最佳帛 ® ® 为 为 为 为 为 为 帛 帛 帛 帛 帛 帛 帛 帛 帛 帛 帛 帛 帛 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 (Webcamcorder) or an electronic product that integrates digital image capture functions, such as a mobile phone or a personal digital assistant (PDA). The digital image device 100 includes an image sensor 1 2, a sensing controller 1〇4, a microprocessor 1〇6, an image signal processing unit 108, a codec 110, a temporary storage 112, an input and output unit 114, and a display. Engine unit 116 and result store 118. The digital video device 1 显示 can display the image on the display device 200 by the partition processing of its internal components (described above). The image sensor 1〇2 according to the embodiment of the present invention is exemplified by an image sensor suitable for a digital camera, but is not limited thereto. A common image sensor 1〇2 may be, but not limited to, an electric How to combine components (Charge_C0Upied Device, CCD). The shirt image sensor 102 has a plurality of pixels (pixeis) for receiving light from the scene and converting the scene into corresponding image data by photoelectric conversion. In order to obtain the color sensing filament, the complex __ sensing pixels are grouped into a Filtering Pattern. The filter template can be, but not limited to, a Bayer Pattern (or Bell Pattern). Embodiments in accordance with the present invention are embodiments in which a Bayer template is used as a filter template, but other types of templates may be used to implement the present invention. Taking "Fig. 2" as an example, every four adjacent sensing pixels 2, 22 & 22, 228 constitutes the lining template 20. The digital image device (10) can receive the light from the scene by the sensing pixels 22R, 22Gr, 22Gb, 22B in each of the Bayer templates 20, and utilize the photoelectric conversion to match the original image. For the convenience of the following description, the following explains and explains the single Bayer sample 2 in the original image. 201216697 According to an embodiment of the invention, the sensing pixels 22R, 22Gr, 22Gb, 22B each have a corresponding 14 bits, but not limited thereto, depending on different image sensors 1〇2, The following is a description of the 12-bit, 1-bit, 16-bit, etc., in which the sensing pixel has 14 bits. Referring to Figure 3A, there is shown a flow chart of an image processing method in accordance with an embodiment of the present invention. The image sensor 102 is applied to the above-mentioned digital image device, wherein the image sensor 102 is in the Bayer Wei 2 towel, has a plain silk fibroin 22R, m and can be thief silk fibroin 22R, 22Gn22Gb, 22B according to the feeding - The light from the scene and the original image. The image processing method includes: Step S302: capturing an original image; Step S3〇4: obtaining a highlight image, a medium bright image, and a low light image from the original image; Step S306: according to the highlighted image and the bright image Corresponding to the low-light image determination - the highlight weight array, the --light weight array and the low-light weight array; and according to the highlight image, the medium-bright image, the low-bright image and the corresponding high-heavy array, the medium-light weight array and Low-light weight arrays get high motion images. The above steps S302 to S308 are performed by the "drum m" image. The microprocessor 106 in the first figure is related to the step S304', please refer to "the first, the bucket map" to the "4D picture is the self. The original image gets a highlight image and a circle. For example, "The 4th Α 丨 丨 ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ Corresponding
Ko獲得如「第4B圖」之高亮影、「第^圓以自原始影像 第4C圖」之中亮影像 201216697 S〇與「第4D圖」之低亮影像&。舉例而言，感應像素 22R，22Gn22Gb，22BS具有14位元，因此位元_呈序即包括：自原 始影像Κ〇之各個感縣素中對應的一十位元區間獲得中亮影像 S^〇 ’自原始影像Κ〇之各個感應像素中對應之一高位元區間獲得高 党影像S]與自原始影像κ〇之各個感應像素中對應的一低位元區 間獲得低亮影像Sj。 其中，中位元區間可以取在各個感應像素之14位元中的第2 至9個位元，高位元區間可以取在各個感應像素之14位元中的第 3至10個位it，低位元區間可以取在各個感應像素之14位元中的 第1至8個位元。由於所取得之位元位置獨，因此，高亮影像 S:1、中亮影像S。與低絲像Sl可分聰有原始影像^之較亮區 域之特徵、中亮度(I域之特徵與較暗區域之特徵。然前述中位元 區間、高位元區間及低位元區間之位元取樣範關為舉例，並非 用以限制本發明之細，且可分娜有縣影像&之㈣亮度範 圍之影像數量可視實際需求而決定，非限定於三張。 更明確地說’在高亮影像S]中，大部分高亮度的區域特徵會 被保留’而低亮度的區域的特徵則會被犧牲。_，於低亮影像 Si中’大部分低亮度的(1域特徵會被保留，而高亮度的區域的特 徵則會被犧牲。至於巾絲像SG，即可保留原始影像巾相對中亮 度的區域特徵，而不在此區域的亮度值均被犧牲。於此需說明的 疋，中位元區間、尚位元區間與低位元區間並不以此為限，可依 據數位影像裝置1GG所要呈現之縣^蚊所取錢像素之位元 區間。其次，上述位元篩選程序中，高亮影像Si、中亮影像s〇 201216697 與低免影像Si之成像順序亦無先後之限制。 ，續請配合「第4B圖」至「第4D圖」一併閱覽之。由於高亮 一 影像S-!、中免影像^與低売影像8!係為自原始影像^執行前述 位元筛選程序後，所獲得之影像資料。因此高亮影像&、中亮影 像S〇與低受影像SA同-拜耳樣板2〇内分別各具有感應像素 22R.1,22G,l522Gb.1,22B.i ' 22R0,22Gr0,22Gb0,22B0 ^ 22R1,22Grl,22Gbl, 22B1。 M於步驟S3G6，舉例而言，在對比高亮影像I、中亮影像 So與低亮影像\於同-拜耳樣板2〇内之同一光遽鏡(以紅光為例） 時’感應像素22R-b22R()，22R1分別具有其各自對應於高亮影像s^、 中亮影像S〇與低亮影像Sl中之邊緣值仏如知。因此，於合併 感應像素時，其分別所佔之觀％ “W⑽I可為 ----—-、--R0 傲 ED1 ew+erq+er1 er-1+erq+Eri 。同樣地，感應像素 2m' u2Gb()’22Gbl 與 22βι，22β〇,22βι 皆會具有其各 •自對應高亮影像心、中亮影像S〇與低亮影像Si之邊緣值 Egm，E⑽Egh、丑㈤知知與EB小EB0，Ebi。因此，在對比高亮影 像~、中亮影像s。與低亮影像Si之同一拜耳樣板2〇下，可依上 述之邊緣制程序決定出對應高亮影像％的一高亮權重陣列 〜、對應中亮影像SG的-中亮權重陣列^與對應低亮影像Si的 -低亮權重陣列Al。其中高亮權重_ ^、中亮權重陣列^與 低亮權重陣列Al分別如「第5A圖」至「第5c圖」所示。 由於高亮影像h、巾亮影像Sq與低細象&可各自保有原 始影像Κ〇之較亮區域、中亮度區域與較暗區域之特徵，且其各自 201216697 即可=可由上述邊緣偵麵而決定。因此步驟· :::二蝴 _ 令-像S =影像~乘以其相對應之高亮權重陣列〜加上 以對應之中亮權重陣列A〇加上低亮影像^乘 一= 目職之低__ Αι，以得_示於「第 不裝置200之高動態影像。 其次铺本發㈣—實關，步驟纖蚊高亮權重 络占A·1中冗權重陣列A〇與低亮權重陣列Al的做法，並不以邊 、.彖偵測程序為限，亦可為—亮度嶋程核—彩度靖程序。 綜上，根據本發明實施例之影像處理方法，可自原始影像獲 取對應於獨碰舰之影像㈣，再透過邊賴測程序、亮度 觸程序或彩度判斷程序，決定各影像資料所佔之相對權重。最 後再以權重加總法得到一張可把不同亮度特徵融合咖〇n)起來之 尚動態影像。 此外’為了使得最終輸出之高動態影像的色彩可以更自然， 且去除多餘人X化之痕跡，根據本發明實施例之影像處理方法， 步驟S308更可以下述實施例實現之： 盘驟S308之第二實施你p 「第3B圖」係為關於「第3A圖」之步驟獅，其第二實施 例之流程圖。由於高絲像％、巾細象&與低亮影像^係為 感應像素對應於8位元之影像資料，故「第m圖」中之微處理器 106可先針對南免影像S l、中亮影像％與低亮影像\進行一階層 柔化程序，如步驟S310所示’以獲得對應於高亮影像^的一高 201216697 主柔影像仏、對應於中亮影像s〇的一中主柔影像與對應於低 亮影像3!的一低主柔影像Ml。 、 、續請配合參閱「第6A圖」與「第6B圖」，以高主柔影像 為例，關於步，驟S310之階層柔化程序，較佳地可以是：將高 "亮影像S-1中各個拜耳樣板20内的同一光遽鏡(以紅光為例)進行處 理程序(如内插法），形成單一色光感應像素⑷。爾後，再以此 方法類推’獲得其餘經階層柔化程序後之感應像素 .UUh ’峰算得較接近於原高亮影像I之高主柔影像 -1並以此方法自中冗景>像§〇獲得其對應的中主柔影像與 2低亮影像Si獲得其對應的低主柔影像Μι。因此，舉例而言，在 高亮影像S·!、中亮影像sG與低亮影像Si之影像資料的解析度為 _萬畫素(8M pixels)的情況下’其對應之高主柔影像吣、中主 柔影像M0與低主柔影像％係為2M pixels。 關於步驟S312,如「第6B圖」至「第6D圖」所示，高主柔 影像吣、中主柔影像Mq與低主柔影像Μι可再根據前述之邊緣 •偵測程序、亮度判斷程序或彩度判斷程序，決定各自對應的高主 柔權重陣列、中主柔權重陣列與低主柔權重陣列。接著，如步驟 S31〇/f不，再依照權重加總法，並以微處理器⑽修正各點之感 應像素後得到-放大回影像資料為8Mpixds之柔化影像。 如「第3C圖」之流程圖所示，柔化影像可再經由邊緣制程 序、亮度判斷程序或彩度判斷程序與一仿真影像進行比較，取得 其各自對應之權重比例，再依權重加總法獲得最終輸出之高動態 影像。其中仿真影像即為未經過階層柔化程序前之高亮影像S i、 11 201216697 中冗景》像S〇與低亮影像Si與其相對應之高亮權重陣列A_i、中亮 權重陣列A〇與低亮權重陣列Al經權重加總法後所產生之影像資 料。 因此，根據本發明實施例之影像處理方法，更可依據上述實 訑例，以進行階層合併(Hierarchical)程序，但並不以上述實施例之 做法為限。設計者於影像處理程序中，可自行決定所要合併的層 級數量及影像資料，續取—階躲合後更高動態、更接近真實 衫度，且可有效去除人工化之影像資料。 雖然本發明以前述之較佳實施例揭露如上，然其並非用以限 定本2明’任何熟習相像技藝者，在不脫離本發明之精神和範圍 内，當可作些許之更動與潤飾，因此本發明之專利保護範圍須視 本說明書所附之申請專利範圍所界定者為準。 【圖式簡單說明】 第1A圖係為習知一影像感測器在接收具有高反差場景時所 呈現的影像相片； 第1B圖係為根據本發明實施彳狀數位影像裝置的結構示音 圖； " 第2圖係為根據本發明實施例之拜耳樣板示意圖； 第3A圖係為根據本發明實施例之流程圖； 第3B圖係為根據本發明步驟S3〇8第二實施例之流程圖； 第3C圖係為根據第圖之流程圖； -第4A圖係為根據本發明實施例之原始影像於單一拜耳樣板 12 201216697 第4B ®係為根據本發明實施例之高亮影像於單-拜耳樣板 不意圖，Ko obtained a high-bright image such as "Picture 4B" and a "Bright image from the original image 4C". The low-definition image & 201216697 S and "4D". For example, the sensing pixels 22R, 22Gn22Gb, 22BS have 14 bits, so the bit_sequence includes: obtaining the medium-light image S^〇 from the corresponding one-tenth interval in each of the original image frames. A high-resolution image S is obtained from a corresponding one of the high-level sections of each of the original pixels of the original image, and a low-light image Sj is obtained from a corresponding low-order section of each of the sensing pixels of the original image. Wherein, the median interval can be taken from the 2nd to 9th bits of the 14-bit of each sensing pixel, and the high-order interval can be taken from the 3rd to 10th bit of the 14-bit of each sensing pixel, the low bit The meta-interval can take the first to eighth bits of the 14-bit of each sensing pixel. Since the position of the obtained bit is unique, the highlighted image S: 1, the bright image S. And the low-filament image Sl can distinguish the characteristics of the brighter region of the original image, the medium brightness (the characteristics of the I domain and the characteristics of the darker region. However, the bits of the median interval, the high-order interval and the low-order interval) The sampling range is an example, and is not intended to limit the details of the present invention, and the number of images in the brightness range of the county image can be determined according to actual needs, and is not limited to three. More specifically, 'at high In the bright image S], most of the high-brightness area features will be preserved' while the low-brightness area features will be sacrificed. _, in the low-light image Si 'most of the low-brightness (1 domain feature will be retained However, the characteristics of the high-brightness area are sacrificed. As for the towel line SG, the area characteristics of the original image towel relative to the medium brightness can be preserved, and the brightness values in this area are not sacrificed. The median interval, the still bit interval and the low bit interval are not limited thereto, and may be based on the bit interval of the money pixel of the county mosquito to be presented by the digital image device 1GG. Secondly, in the above bit filter program, Highlight image Si, medium bright shadow There is no limit to the order of imaging such as s〇201216697 and low-free image Si. Please continue to read it together with "4B" to "4D". Because of the high-definition image S-! And the low-definition image 8 is the image data obtained after the above-mentioned bit filter program is executed from the original image. Therefore, the highlight image & the bright image S〇 is the same as the low-image SA-Bayer model 2〇 Each has therein sensing pixels 22R.1, 22G, l522Gb.1, 22B.i ' 22R0, 22Gr0, 22Gb0, 22B0 ^ 22R1, 22Grl, 22Gbl, 22B1. M in step S3G6, for example, in contrast highlight images I, the bright image So and the low-light image \ in the same - Bayer template 2 之 within the same pupil mirror (in the case of red light) when the 'inductive pixels 22R-b22R (), 22R1 have their respective corresponding to highlight The edge values in the image s^, the medium bright image S〇 and the low-light image S1 are as known. Therefore, when combining the sensing pixels, the respective % of view “W(10)I can be ------, -- R0 proud ED1 ew+erq+er1 er-1+erq+Eri. Similarly, the sensing pixel 2m' u2Gb()'22Gbl and 22βι, 22β〇, 22βι will have their own • self-correspondence Bright image heart, medium bright image S〇 and low light image Si edge value Egm, E (10) Egh, ugly (five) know and EB small EB0, Ebi. Therefore, in contrast highlight image ~, medium bright image s. and low light image The same Bayer sample of Si can be used to determine a highlight weight array corresponding to the highlighted image %, the medium-light weight array corresponding to the bright image SG, and the corresponding low-light image Si. - Low-light weight array Al. The highlight weight _ ^, the medium-light weight array ^ and the low-light weight array Al are respectively shown in "5A map" to "5c map". Since the highlight image h, the towel bright image Sq and the low fine image & can each retain the characteristics of the brighter region, the medium luminance region and the darker region of the original image, and their respective 201216697 can be And decided. Therefore, the steps ::::2 _ 令 - like S = image ~ multiplied by its corresponding highlight weight array ~ plus the corresponding light weight array A 〇 plus low light image ^ multiply one = position Low __ Αι, to get _ shown in the "No Motion 200 high dynamic image. Secondly, the distribution of the hair (4) - real off, the step of the mosquito highlights the weight of the A·1 redundant weight array A 〇 and low light weight The method of the array Al is not limited to the edge detection process, and may be a brightness process core-color saturation process. In summary, the image processing method according to the embodiment of the invention may be obtained from the original image. Corresponding to the image of the single-handed ship (4), the relative weight of each image data is determined by the edge measurement program, the brightness touch program or the chroma judgment program. Finally, a weight summation method is used to obtain a different brightness characteristic. In addition, in order to make the color of the high-motion image of the final output more natural, and to remove the trace of the extra person, the image processing method according to the embodiment of the present invention, step S308 can be further The following embodiment is implemented: You two embodiments p "FIG. 3B" based on a step of "FIG. 3A" lion, a second embodiment of a flow diagram. Since the high-filament image %, the towel fine image & and the low-light image ^ are the sensing pixels corresponding to the 8-bit image data, the microprocessor 106 in the "mth image" can first target the south image S l , Bright image % and low-light image\ perform a level softening process, as shown in step S310, to obtain a high 201216697 main soft image corresponding to the highlighted image ^, corresponding to the middle bright image s〇 The image and a low main soft image M1 corresponding to the low light image 3!. For the continuation, please refer to "6A" and "6B". Taking the high-flexible image as an example, the hierarchical softening procedure for step S310 may preferably be: high "bright image S The same pupil mirror (in the case of red light) in each Bayer template 20 in -1 performs a processing procedure (such as interpolation) to form a single color light sensing pixel (4). Then, in this way, the analog pixel after the rest of the softening procedure is obtained. The UUh' peak is calculated to be closer to the high-elastic image-1 of the original highlighted image I and is used in this method. § 〇 obtain its corresponding central soft image and 2 low light image Si to obtain its corresponding low main soft image Μι. Therefore, for example, in the case where the resolution of the image data of the highlight image S·!, the medium bright image sG, and the low-light image Si is _ megapixel (8M pixels), the corresponding high main soft image 吣The main soft image M0 and the low main soft image % are 2M pixels. In step S312, as shown in "Fig. 6B" to "6D", the high main soft image 吣, the middle main soft image Mq and the low main soft image 可 ι can be further based on the aforementioned edge detection program, brightness judging program Or the chroma judgment program determines the respective high main soft weight array, medium main soft weight array and low main soft weight array. Then, if the step S31 〇 / f is not, according to the weight summation method, and the microprocessor (10) corrects the sensing pixels of each point to obtain - the enlarged image data is 8Mpixds soft image. As shown in the flowchart of "3C", the softened image can be compared with a simulated image by an edge program, a brightness judgment program or a chroma judgment program, and the weight ratios of the respective corresponding weights are obtained, and then the weights are added according to the weights. The method obtains a high dynamic image of the final output. The simulated image is the highlight image before the level softening procedure S i, 11 201216697, the scenes like the S〇 and the low-light image Si and the corresponding highlight weight array A_i, the medium-light weight array A〇 The low-light weight array Al is weighted and added to the image data generated by the total method. Therefore, the image processing method according to the embodiment of the present invention can be further subjected to a Hierarchical program according to the above embodiment, but is not limited to the above embodiment. In the image processing program, the designer can decide the number of layers and image data to be merged, and continue to take the higher dynamics, closer to the real shirt, and effectively remove the artificial image data. While the present invention has been described above in terms of the preferred embodiments thereof, it is not intended to limit the scope of the invention, and may be modified and retouched without departing from the spirit and scope of the invention. The patent protection scope of the present invention is defined by the scope of the patent application attached to the specification. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1A is a pictorial image presented by a conventional image sensor when receiving a scene with high contrast; FIG. 1B is a structural diagram of a digital image device according to the present invention. &2; Fig. 2 is a schematic diagram of a Bayer sample according to an embodiment of the present invention; Fig. 3A is a flow chart according to an embodiment of the present invention; and Fig. 3B is a flow of a second embodiment of step S3 8 according to the present invention; Figure 3C is a flow chart according to the figure; - Figure 4A is an original image according to an embodiment of the present invention on a single Bayer sample 12 201216697 4B ® is a highlight image according to an embodiment of the present invention - Bayer model is not intended,
不意圖， 單一拜耳樣板 第4D ®係為根據本發明實施例之低亮影像於單-拜耳樣板 示意圖； ’ 第5A圖係為根據本發明實施例之高_重陣列示意圖； 第5B圖係為根據本發明實施例之巾雜重陣列示意圖. 第5C圖係為根據本發明實施例之低亮權重陣列示意圖； 第6A圖係為根據本發明實施例之高亮影像於單一拜耳樣板 示意圖； 第6B 板示意圖； 圖係為根據本發明實施例之高主柔影像於單一拜耳樣 第6C圖係為根據本發明實施例之中主柔影像於單一拜耳樣 鲁 板不意圖；以及 第6D 板示意圖。 圖係為根據本發明實施例之低主柔影像於單一拜耳樣 【主要元件符號說明】 20 拜耳樣板 100 數位影像裝置 102 影像感測器 104 感測控制器 13 106 201216697 108 110 112 114 116 118 200 S-lIt is not intended that the single Bayer sample 4D® is a schematic diagram of a low-bright image in a single-Bayer model according to an embodiment of the present invention; '5A is a schematic diagram of a high-heavy array according to an embodiment of the present invention; FIG. 5C is a schematic diagram of a low-light weight array according to an embodiment of the present invention; FIG. 6A is a schematic diagram of a high-bright image in a single Bayer template according to an embodiment of the present invention; 6B board schematic diagram; diagram is a high main soft image according to an embodiment of the present invention in a single Bayer sample 6C diagram is not intended to be a single Bayer sample in accordance with an embodiment of the present invention; and a 6D board diagram . The figure is a low main soft image according to an embodiment of the present invention in a single Bayer sample. [Main component symbol description] 20 Bayer template 100 Digital image device 102 Image sensor 104 Sensing controller 13 106 201216697 108 110 112 114 116 118 200 Sl
So S, M.i M〇 M! K〇 22^,22〇Γ,22〇^22β 22^.ι,22〇γ-ι,22〇^ι,22β-ι 22R〇,22Gr〇,22Gb〇,22B〇 22Ri ,22〇ri ,22〇bi ,22b \ 24r. i ,24Gr. j 24Gb.! 24b. i 24R〇,24Gr〇,24Gb〇,24B〇 24ri ,24Gri ,24Gbi ,24Bi 微處理器 影像訊號處理單元 編解碼器 暫時儲存器 輸入輸出單元 顯示引擎單元 結果儲存器 顯示裝置 高亮影像 中亮影像 低亮影像 南主柔影像 中主柔影像 低主柔影像 原始影像 感應像素 感應像素 感應像素 感應像素 感應像素 感應像素 感應像素 201216697So S, Mi M〇M! K〇22^,22〇Γ,22〇^22β 22^.ι,22〇γ-ι,22〇^ι,22β-ι 22R〇,22Gr〇,22Gb〇,22B 〇22Ri , 22〇ri , 22〇bi , 22b \ 24r. i ,24Gr. j 24Gb.! 24b. i 24R〇, 24Gr〇, 24Gb〇, 24B〇24ri, 24Gri, 24Gbi, 24Bi microprocessor image signal processing Unit codec temporary storage unit input and output unit display engine unit result memory display device highlight image bright image low light image south main soft image main soft image low main soft image original image sensing pixel sensing pixel sensing pixel sensing pixel sensing Pixel sensing pixel sensing pixel 201216697
Wr.^Wro^Wr! 權重 WGr-l,WGr〇,WGrl 權重 WGb-l，WGb0，WGbl 權重 Wb-1,Wb〇,Wbi 權重 A-i 高亮權重陣列 A〇 中亮權重陣列 Ai 低亮權重陣列 15Wr.^Wro^Wr! Weight WGr-l, WGr〇, WGrl Weight WGb-l, WGb0, WGbl Weight Wb-1, Wb〇, Wbi Weight Ai Highlight Weight Array A 〇 Light Weight Array Ai Low Light Weight Array 15