201121336 六、發明說明： 【發明所屬之技術領域】 本發明係關於一種高動態範圍合成影像之色彩校正 方法，更詳而言之，係有關於藉由拍攝裝置取得不同曝光 程度所拍攝到的影像，並找出同色彩於不同明度下的關連 性，以作為對該拍攝裝置所取得的高動態範圍合成影像進 行色彩校正之技術。 I【先前技術】 目前多數顯示設備所能呈現動態範圍（Dynamic Range) 有限，導致許多影像細節無法清楚呈現，所以透過高動態 範圍（High Dynamic Range, HDR)顯示技術更顯得重要，因 此，高動態範圍合成影像技術也逐漸被用來呈現較佳品質 之影像。 所謂的動態範圍係指影像中最暗與最亮的對比值，一 般來說，電腦在顯示數位影像時，若要達到如真實環境之 籲明暗對比表現，其困難度較高，常因無法同時兼顧曝光不 足導致過暗區域或曝光過度導致過亮區域的情況，而造成 影像中部份影像無法完整顯示出來；對此，高動態範圍影 像合成技術則是針對前述數位影像表現之難處進行改善， 利用多張影像曝光程度不同的數位影像資料，疊合出具有 明暗範圍較廣的合成影像，以增加影像顯示範圍的影像處 理技術。 目前高動態範圍影像合成的作法，係利用複數張不同 曝光程度的影像進行合成，在連續的曝光影像中，影像含 4 111354 201121336 有曝光過度或曝光不足的像素點，當然也有曝光恰當的像 素點，因此，由多張影像中包找尋曝光恰當之像素點來進 行合成，以得到所謂的高動態範圍合成影像。惟，依目前 高動態範圍影像合成方式，多數著重在以恢復影像細節資 訊為主要目的，對於合成影像色彩表現部分，頂多只以白 平衡或依比例計算等簡單方式進行處理，然而，針對曝光 程度不同之影像，其進光量不同恐造成影像彩度受影響或 I是色相變化，進而導致高動態範圍影像合成後色調可能發 生變化，甚至有失真情況，使得影像呈現品質不夠完善。 因此，如何提供一種高動態範圍合成影像之色彩校正 方法，使得高動態範圍合成影像在影像合成後其色彩表現 更接近真實，以減少傳統高動態範圍影像合成時，僅考量 以清楚呈現為目的而導致成像色彩品質不佳等問題，實為 目前亟欲解決之技術課題。 【發明内容】 馨鑒於上述習知技術之缺點，本發明係提出一種高動態 範圍合成影像之色彩校正方法，其目的在藉由建立色彩偏 移模型（color distortion model ’ CDM)以及將欲校正之高 動態範圍合成影像透過色彩校正矩陣（color calibration matrix，Mcc )，以得到更符合真實色彩之高動態範圍合成 影像。 本發明提供一種高動態範圍合成影像之色彩校正方 法，其包括以下步驟：（1)令拍攝裝置在相同光源環境下 以不同曝光程度拍攝色票之影像，以取得該影像於色彩空 5 111354 201121336 間（Color Space)之色度座標；（2)利用該色度座標建立 色彩偏移模型；（3)透過該色彩偏移模型以迴歸計算方式 取得色彩校正矩陣；以及（4)藉由該色彩校正矩陣對該拍 攝裝置所取得的高動態範圍合成影像進行色彩校正，以調 整該高動態範圍合成影像之色彩。 其中，該色彩偏移模型係由色票之原始色彩RGB值 轉換成色度座標所組成，而該色度座標係為該影像於不同 曝光程度之色度值。另外，該色彩校正矩陣係為該色票於 不同曝光程度之明度、彩度及色相之變化關係。 此外，該步驟（3)係依據該色票於不同曝光程度之 原始色彩的明度值，透過該色彩偏移模型得到對應的目標 色度值，且將該目標色度值與該色票原始色彩之色度值透 過二次多元迴歸方式以得該色彩校正矩陣。而該步驟（4) 係透過該色彩校正矩陣内各色彩對應關係，依據該高動態 範圍合成影像之各像素點的明度值，以對該合成影像進行 φ色彩校正。 於一變化例中，經由本發明色彩校正後之高動態範圍 合成影像，復利用 CIECAM02色外貌模式（Color Appearance Model，CAM )調整經色彩校正後之該高動態 範圍合成影像於顯示器上所呈現之外貌色彩。 相較於習知技術，本發明之高動態範圍合成影像之色 彩校正方法，係藉由建立色彩偏移模型以作為色彩校正依 據，使得高動態範圍合成影像之色彩呈現接近真實，以降 低傳統高動態合成影像於色彩表現方面之不足，而將高動 6 111354 201121336 態範圍合成影像透過色彩校正以提升其呈現的色彩品質。 再者，經色彩修正後的該高動態範圍合成影像復透過色外 貌模式調整其於數位顯示裝置之外貌色彩，俾讓使用者透 過顯示裝置觀看影像時，能更接近人眼所觀察到影像色彩。 【實施方式】 以下藉由特定的具體實例說明本發明之技術内容，熟 悉此技藝之人士可由本說明書所揭示之内容輕易地瞭解本 發明之其他優點與功效。 ® 如第1圖所示，係用以說明本發明之高動態範圍合成 影像之色彩校正方法之流程圖。如圖所示，首先於步驟S1 中，係令拍攝裝置在同一光源環境下以不同曝光程度拍攝 色票之影像，取得該些影像於色彩空間内之色度座標。此 步驟之目的是為了取得該拍攝裝置所拍攝到的影像於不同 曝光程度下，每種色彩所產生的色度變化，以作為後續色 彩校正之依據；而該些影像於色彩空間内之色度座標的取 鲁得方式係將色票置於光源均勻的同一拍攝箱内，以不同曝 光程度拍攝複數張影像，因同一場景於不同曝光程度所拍 攝所呈現色彩會產生變化，藉此得到複數張同一場景不同 曝光程度的影像，最後，再透過該些於不同曝光程度之影 像，取得色票中各種顏色於色彩空間内的對應色度座標。 該色票可為GretagMacbeth色票、PANTONE色票、 DIC色票、RAL色票或被攝影及繪圖用於校正顏色之色 票等任一種，以作為拍攝後用於判斷色彩差異之依據。 其中，該色彩空間係一種透過X軸、Y軸與Z軸所構 7 111354 201121336 成之三維空間，例如透過紅色（Red )、綠色（Green )、藍 色（Blue)三原色（即簡稱RGB)作為X、Y和Z座標軸 來表示色彩。亦可透過色相（hue )、彩度（saturation )和 明度（lightness)等另一種色彩表示要素作為X、Y和Z 座標軸來表示色彩空間。每種可能的顏色於色彩空間内都 有唯一的位置；於此需說明，本發明所採用的色彩空間不 限定，而可為CIELAB色彩空間、CIEXYZ色彩空間、HSB I色彩空間、RGB色彩空間或用於顯示色度之色彩空間等任 一種0 接著，於步驟S2中，利用色度座標建立色彩偏移模 型。由於所有色度座標中包括各種顏色之色票於不同曝光 程度下所呈現之色度值，因此，於本實施例中，可將色彩 空間内各色度座標數據透過三次迴歸多項式計算，以建立 色彩偏移模型，換句話說，該色彩偏移模型之建構，係由 該色票之原始色彩RGB值轉換成色度座標，再將該色度座 鲁標之色彩關係建構成色彩偏移模型。 之後，於步驟S3中，則透過色彩偏移模型取得色彩 校正矩陣。主要係依據色票内各色彩於不同曝光程度下的 明度值，透過色彩偏移模型計算比對，以找出符合的目標 色度值，接著，再將目標色度值與色票各種色彩色度值透 過二次多元迴歸方式計算以得色彩校正矩陣，換句話說， 該色彩校正矩陣係為該色票於不同曝光程度之明度、彩度 以及色相變化關係。 最後，於步驟S4中，透過色彩校正矩陣作為該拍攝 8 111354 201121336 裝置對所其取得的高動態範圍合成影像進行色彩校正，以 調整該高動態範圍合成影像之色彩。本步驟係由色彩偏移 模型得到色彩校正矩陣内各色彩對應關係，此時，將欲進 行校正的高動態範圍合成影像之各像素點的明度值，透過 該色彩校正矩陣進行計算，以得到高動態範圍合成影像校 正後色彩。 如第2a圖所示，係用以詳細說明第1圖所示之本發 明高動態範圍合成影像之色彩校正方法中步驟S3的流程 圖，即進一步說明取得色彩校正矩陣的方法，首先，於步 驟S31中，係依據色票於不同曝光程度之原始色彩的明度 值，透過色彩偏移模型得到對應的目標色度值，由於色彩 不同曝光程度下其明度值不同，也造成色彩呈現上有所差 異，因此，藉由色彩偏移模型中各色彩於不同明度值下的 色度值，以找出相對應的目標色度值，而該目標色度值係 最符合人眼所觀看到的色彩。 接著，於步驟S32中，將步驟S31所取得目標色度值 與色票上各原始色彩色度值建立對應關聯性，且透過二次 多元迴歸方式估計出色彩校正矩陣，以作為高動態範圍合 成影像色彩校正之用。 如第2b圖所示，係用以詳細說明第1圖所示之本發 明高動態範圍合成影像之色彩校正方法之步驟S4的流程 圖，即進一步說明進行色彩校正的處理，首先於步驟S.41 中，係依據高動態範圍合成影像之各像素點的明度值，透 過色彩校正矩陣得到對應的色彩，藉由色彩於色彩校正矩 9 111354 201121336 陣内不同明度值之關聯性，求得最接近人眼所觀察到之色 彩。 接著，於步驟S42中，將所得到對應色彩，取代高動 悲範圍合成衫像之色彩，完成色彩校正程序，使該高^雜 範圍合成影像所呈現之色彩符合人眼所視之感覺。 茲以一具體實施例來說明本發明高動態範圍合成影 像之色彩校正方法。 φ 首先，依據不同曝光程度下色彩關聯性來建立色彩偏 移模型’其作法係把GretagMacbeth色票（註：該 GretagMacbeth色票係一種包括24種顏色之色票）置於D65 標準光源對色燈箱内，透過數位相機拍攝在不同曝光程度 下的複數張影像，且將所有數據轉換成CIELAB色彩空間 内的色度座標。第3圖所示係包括24種顏色於不同曝光程 度所造成色度分佈，其中，三個基本座標係包含表示顏色 的亮度L，L=0為黑色，而L= 100為白色，於紅色和綠 #色之間的位置a# (a+負值為綠色，正值為紅色）以及在黃 色和藍色之間的位置(1/負值為藍色，正值為黃色）。 接著，透過三次迴歸的多項式（式1)計算出在不同 曝光程度下，該GretagMacbeth色票中24種色彩於CIELAB 色彩空間内之明度、彩度、色相等變化值，以構成色彩偏 務模型，作為色彩校正之依據。 a. = m, -i- τηΊ .L] + m, ,L, + , ! : ； ’ ' ,1 = 1,2,3,-, 24 (1) bi = m, tL^ -r -i- mz iLi -r ^ 其中，I、<3、6表示同一色塊於不同曝光程度下的色 10 111354 201121336 • 度值，m表示參數，ζ·表示GretagMacbeth色票的24種顏 4 色。 透過前述將GretagMacbeth色票24種原始色彩於不同 曝光程度下的RGB值轉換為CIELAB色彩空間所表示色 度座標，以形成如第3圖所示之CIELAB色彩空間，也就 是透過該CIELAB色彩空間可得知每一種色彩在不同明度 下（依據曝光程度不同）之色度以及其關聯性，因此，欲 ¥校正的高動態範圍合成影像中任一像素點透過該關聯性， 即可找出所需要的目標色度值，在此所謂之目標色度值係 指最接近人眼所視之顏色，因此，使用者可訂定最符合自 己期望或觀看到色彩的明度值，以作為校正依據。 接著，依據該色彩偏移模型來建立色彩校正矩陣。其 作法係將數位像機所拍攝多張GretagMacbeth色票於同一 場景不同曝光程度下的低動態範圍影像，並將該些影像隨 機選取進行高動態範圍影像合成，以取得各明暗程度不同 鲁的合成影像樣本，其主要目的係為了分析不同明暗程度之 合成影像的色彩變化差異。 將前述合成影像原始色彩的明度值L透過色彩偏移 模型，可取得對應相同色彩的目標色度值，最後，將該目 標色度值及前述合成影像原始色彩的色度值透過二次多元 迴歸計算，以tf算出色彩校正矩陣，如下述式（2)至式（4) 所表示。 11 111354 (2) V201121336 L" a" b" -Mcc b' L'a， L'b' a'b'201121336 VI. Description of the Invention: [Technical Field] The present invention relates to a color correction method for a high dynamic range synthetic image, and more particularly, to an image captured by a camera at different exposure levels And find the correlation between the same color and different brightness, as a technique for color correction of the high dynamic range synthesized image obtained by the camera. I [Prior Art] At present, most display devices have a limited dynamic range, which makes many image details unclear, so it is more important to display through High Dynamic Range (HDR) display technology. Therefore, high dynamics Range synthetic imaging techniques are also increasingly being used to present images of better quality. The so-called dynamic range refers to the darkest and brightest contrast value in the image. Generally speaking, when the computer displays the digital image, if it is to achieve the contrast between the dark and the dark, the difficulty is higher, often because the simultaneous Considering the lack of exposure, the over-dark area or the over-exposure leads to the over-bright area, and some images in the image cannot be completely displayed. For this, the high-dynamic range image synthesis technology improves the difficulty of the above-mentioned digital image. By using digital image data with different exposure levels of multiple images, a composite image with a wide range of light and dark is superimposed to increase the image processing range of the image display range. At present, the method of high dynamic range image synthesis is to use a plurality of images with different exposure levels for synthesis. In continuous exposure images, the image contains 4 111354 201121336 with overexposed or underexposed pixels, and of course there are also exposed pixels. Therefore, it is synthesized by finding a pixel with appropriate exposure from a plurality of images to obtain a so-called high dynamic range synthesized image. However, according to the current high dynamic range image synthesis method, most of the focus is on restoring image detail information. For the color expression part of synthetic image, at most, it is only processed in a simple way such as white balance or proportional calculation. However, for exposure For images of different degrees, the amount of light entering may cause the image chroma to be affected or I to change the hue, which may cause the hue of the high dynamic range image to be changed or even distorted, resulting in insufficient image quality. Therefore, how to provide a color correction method for a high dynamic range synthetic image, so that the color performance of the high dynamic range synthetic image is more realistic after image synthesis, so as to reduce the traditional high dynamic range image synthesis, only for the purpose of clear presentation The problems that lead to poor image color quality are the technical issues that are currently being solved. SUMMARY OF THE INVENTION In view of the above disadvantages of the prior art, the present invention provides a color correction method for a high dynamic range synthetic image, the purpose of which is to establish a color distortion model (CDM) and to correct it. The high dynamic range synthesized image passes through a color calibration matrix (Mcc) to obtain a high dynamic range synthesized image that is more in line with real colors. The invention provides a color correction method for a high dynamic range synthetic image, which comprises the following steps: (1) causing the image capturing device to take an image of the color ticket at different exposure levels under the same light source environment to obtain the image in the color space 5 111354 201121336 a color coordinate of the color space; (2) using the chromaticity coordinate to establish a color shift model; (3) obtaining a color correction matrix by regression calculation using the color shift model; and (4) by the color The correction matrix performs color correction on the high dynamic range synthesized image obtained by the photographing device to adjust the color of the high dynamic range synthesized image. The color shift model is composed of the original color RGB value of the color ticket converted into a chromaticity coordinate, and the chromaticity coordinate is a chromaticity value of the image at different exposure levels. In addition, the color correction matrix is a relationship between the brightness, chroma, and hue of the color ticket at different exposure levels. In addition, the step (3) obtains a corresponding target chromaticity value according to the brightness value of the original color of the color ticket at different exposure levels, and obtains the target chromaticity value and the original color of the color ticket. The chromaticity values are obtained by quadratic multiple regression to obtain the color correction matrix. And the step (4) is to perform φ color correction on the synthesized image according to the color value of each pixel of the high dynamic range synthetic image by using the color correspondence in the color correction matrix. In a variant, the CIECAM02 Color Appearance Model (CAM) is used to adjust the color-corrected high-dynamic range synthesized image on the display through the color-corrected high-dynamic range synthesized image of the present invention. Appearance color. Compared with the prior art, the color correction method of the high dynamic range synthetic image of the present invention is used as a color correction basis by establishing a color shift model, so that the color rendering of the high dynamic range synthetic image is close to reality, thereby reducing the traditional high. Dynamically synthesizing images in terms of color performance, and moving the synthetic image of the high-motion 6 111354 201121336 range through color correction to enhance the color quality of the image. Furthermore, the color-corrected image of the high dynamic range composite image is adjusted to the appearance of the digital display device, so that when the user views the image through the display device, the image color can be closer to the human eye. . [Embodiment] The technical contents of the present invention are described below by way of specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in the present specification. As shown in Fig. 1, it is a flow chart for explaining the color correction method of the high dynamic range synthesized image of the present invention. As shown in the figure, first, in step S1, the imaging device is caused to capture images of the color tickets at different exposure levels under the same light source environment, and obtain the chromaticity coordinates of the images in the color space. The purpose of this step is to obtain the chromaticity changes produced by each color of the image captured by the camera at different exposure levels, as a basis for subsequent color correction; and the chromaticity of the images in the color space. The method of taking the coordinates is to place the color ticket in the same shooting box with uniform light source, and take multiple images at different exposure levels. The color of the same scene will be changed at different exposure levels, thereby obtaining multiple sheets. The images of different exposure levels in the same scene, and finally, the corresponding chromaticity coordinates of the various colors in the color space in the color space are obtained through the images of different exposure levels. The color ticket may be any one of a GretagMacbeth color ticket, a PANTONE color ticket, a DIC color ticket, a RAL color ticket, or a color ticket used for correcting colors by photography and drawing, as a basis for judging color difference after shooting. The color space is a three-dimensional space formed by the X-axis, the Y-axis, and the Z-axis, for example, by red (Red), green (Green), and blue (Blue) (hereinafter referred to as RGB). The X, Y, and Z coordinate axes represent colors. Another color representation element such as hue, saturation, and lightness can also be used as the X, Y, and Z coordinate axes to represent the color space. Each possible color has a unique position in the color space; it should be noted that the color space used in the present invention is not limited, but may be CIELAB color space, CIEXYZ color space, HSB I color space, RGB color space or Any one of color spaces for displaying chromaticity, etc. Next, in step S2, a color shift model is established using chromaticity coordinates. Since all the chromaticity coordinates include the chromaticity values of the color tickets of the various colors at different exposure levels, in this embodiment, the chromaticity coordinate data in the color space can be calculated by the cubic regression polynomial to establish the color. The offset model, in other words, the color shift model is constructed by converting the original color RGB values of the color ticket into chromaticity coordinates, and then constructing the color shift model by the color relationship of the chromaticity seat. Thereafter, in step S3, the color correction matrix is acquired through the color shift model. Mainly based on the brightness value of each color in the color ticket at different exposure levels, the color shift model is used to calculate the alignment to find the matching target chromaticity value, and then the target chromaticity value and the color ticket color color The degree value is calculated by the quadratic multiple regression method to obtain a color correction matrix. In other words, the color correction matrix is the relationship between brightness, chroma, and hue of the color ticket at different exposure levels. Finally, in step S4, the color correction matrix is used as the camera to perform color correction on the high dynamic range synthesized image obtained by the device to adjust the color of the high dynamic range synthesized image. In this step, the color correspondence model obtains the color correspondence relationship in the color correction matrix. At this time, the brightness value of each pixel of the high dynamic range synthesized image to be corrected is calculated through the color correction matrix to obtain a high value. Dynamic range synthetic image corrected color. As shown in FIG. 2a, it is a flowchart for explaining step S3 in the color correction method of the high dynamic range synthesized image of the present invention shown in FIG. 1 to further explain the method for obtaining the color correction matrix. First, in the step. In S31, according to the brightness value of the original color of the color ticket at different exposure levels, the corresponding target chromaticity value is obtained through the color shift model, and the brightness value is different due to different brightness levels under different exposure degrees. Therefore, by using the chromaticity values of the colors in different color values in the color shift model, the corresponding target chromaticity values are found, and the target chromaticity values are most consistent with the colors observed by the human eye. Next, in step S32, the target chromaticity value obtained in step S31 is associated with each original color chromaticity value on the color ticket, and the color correction matrix is estimated by the quadratic multiple regression method as a high dynamic range synthesis. Image color correction. As shown in FIG. 2b, it is a flowchart for explaining in detail the step S4 of the color correction method of the high dynamic range synthesized image of the present invention shown in FIG. 1, that is, the process of performing color correction is further explained, first in step S. In 41, according to the brightness value of each pixel of the high dynamic range synthetic image, the corresponding color is obtained through the color correction matrix, and the closestness is obtained by the correlation of different brightness values in the color correction moment 9 111354 201121336 array. The color observed by the eye. Next, in step S42, the color corresponding to the high-sorage range is combined with the color of the shirt image to complete the color correction process, so that the color represented by the high-range synthetic image conforms to the perception of the human eye. A color correction method for a high dynamic range synthetic image of the present invention will be described in a specific embodiment. φ First, the color shift model is established according to the color correlation under different exposure levels. The method is to put the GretagMacbeth color ticket (Note: the GretagMacbeth color ticket is a color ticket including 24 colors) in the D65 standard light source color light box. Inside, a plurality of images at different exposure levels are captured by a digital camera, and all data is converted into chromaticity coordinates in the CIELAB color space. Figure 3 shows the chromaticity distribution of 24 colors at different exposure levels, where the three basic coordinates contain the brightness L representing the color, L = 0 is black, and L = 100 is white, in red and The position between green #color a# (a+ negative value is green, positive value is red) and the position between yellow and blue (1/negative value is blue, positive value is yellow). Then, through the polynomial of the three regressions (Formula 1), the brightness, chroma, and color equalization values of the 24 colors in the GIELAB color space of the GretagMacbeth color ticket are calculated to form a color shifting model at different exposure levels. As the basis for color correction. a. = m, -i- τηΊ .L] + m, ,L, + , ! : ; ' ' ,1 = 1,2,3,-, 24 (1) bi = m, tL^ -r -i - mz iLi -r ^ where I, <3,6 represent the color of the same patch at different exposure levels. 10 111354 201121336 • Degree value, m for the parameter, ζ· indicates the 24 color 4 colors of the GretagMacbeth color ticket. Converting the RGB values of the 24 original colors of the GretagMacbeth color ticket to the chromaticity coordinates represented by the CIELAB color space through the foregoing to form the CIELAB color space as shown in FIG. 3, that is, through the CIELAB color space. Knowing the chromaticity of each color under different brightness (depending on the degree of exposure) and its relevance, therefore, any pixel in the high dynamic range synthetic image to be corrected by the correlation can find out what is needed The target chromaticity value, the so-called target chromaticity value refers to the color that is closest to the human eye. Therefore, the user can set the brightness value that best matches the desired or viewed color as the basis for correction. Then, a color correction matrix is established according to the color shift model. The method is to take a plurality of GretagMacbeth color tickets of a digital camera in a low dynamic range image with different exposure levels in the same scene, and randomly select the images for high dynamic range image synthesis to obtain a composite of different brightness levels. The main purpose of the image sample is to analyze the difference in color variation of the composite image with different brightness levels. The brightness value L of the original color of the synthetic image is transmitted through the color shift model to obtain a target chromaticity value corresponding to the same color, and finally, the target chromaticity value and the chromaticity value of the original color of the synthesized image are transmitted through a quadratic multiple regression. Calculated, the color correction matrix is calculated as tf, as expressed by the following equations (2) to (4). 11 111354 (2) V201121336 L"a"b" -Mcc b' L'a, L'b' a'b'
LaLa
Mcc·· Ί>2 '5 ^1,6 ^1,7 \8 ^1,9 ^1,10 2>] ni,2 ru, „ ^ „ ’ ^,4 气 5 «2,6 W2,7 W2,8 n2,9 W2,10 ^3,3 ^3·4 ^3,5 ^3,6 ^3,7 ^3,8 ^3,9 ^3,10 «3 (3)Mcc·· Ί>2 '5 ^1,6 ^1,7 \8 ^1,9 ^1,10 2>] ni,2 ru, „ ^ „ ' ^,4 gas 5 «2,6 W2,7 W2,8 n2,9 W2,10 ^3,3 ^3·4 ^3,5 ^3,6 ^3,7 ^3,8 ^3,9 ^3,10 «3 (3)
U :«』+ n”a, + „ 6,+ '.乂，+'〆2 +'〆 +'7ZV+ni，0,+'々’ + '10 ^3"= ni.]L'+ n3 0a'n ~ A +^2.5^+^2/+^2,71^+^2,8^^+^2,9^6+^2,10 (4) 3 n3,4^ + nisa'2 + ni6b'2 + ni7L'a' + n3gL'b' + n3ga'b' + n310 其中，上i、 ^ ，，a二 、α’及6’表示合成影像原色彩的色度值，1”、 β及，則表示明厗杜 ^ TS ., a度值上，代入色彩偏移模型中相同顏色的 夕項式所切的目料度值。 ' =上述方式，藉由該色彩偏移模型中不同明度值所 A# 度差異產生對應關聯性，再透過高動態範圍合成 像素點㈣度值求得所期待的色度值，以取代原高 恶範圍合成影像内各像素點之色彩，使該影像色彩更符 合肉眼所視之顏色。 經本發明之高動態範圍合成影像之色彩校正方法而 几成色彩校正後之高動態範圍合成影像，復可利用 CIECAM02色外貌模式以調整於顯示器上所呈現該高動態 範圍合成影像之色彩。簡單來說，藉由CIECAM02色外貌 Π1354 12 201121336 „ 模式調整，可使得影像隨著不同環境光源變化而模擬人眼 t 色適應後的色彩修正結果，俾使經色外貌模式調整後之影 像與環境光源改變後所擷取的影像達到色外貌一致。 其作法如下：將某一光源下顯示的影像其像素RGB 值轉換至XYZ三刺激值，以作為光源環境下樣本三刺激值 X!、Y!及Z!，同時紀錄該光源下影像白點的三刺激值Xwl、 Ywl及Zwl，將上述資料經由CIECAM02色外貌模式計算 以取得知覺相關屬性，因此，當光源有所改變時則透過該 ® CIECAM02可求得新的三刺激值X2、Y2及Z2，該色外貌 模式轉換說明圖如第4圖所示，其中，在兩種不同光源下， 透過CIECAM02色外貌模式以樣本三刺激值與影像白點 的三刺激值找出另一種與光源對應的三刺激值，以作為顯 示器顯示調整之用，俾使人眼觀看顯示器所呈現影像時能 更接近人眼所觀看到的實際色彩。 綜上所述，本發明提出一種高動態範圍合成影像之色 籲彩校正方法，相較於習知缺點，本發明提出藉由不同曝光 程度之影像於色彩空間内各色彩的色彩座標，以建構出色 彩偏移模組，接著把各色彩之明度值透過色彩偏移模組找 到對應相同顏色的目標色度值，最後把目標色度值與原色 彩色度值比較計算求得色彩校正模組。透過上述方式，可 將高動態範圍合成影像進行色彩調整，俾使該高動態範圍 合成影像呈現上更符合人眼所看到的真實色彩，減低傳統 高動態範圍影像合成後顏色失真的狀態；另外，透過 CIECAM02色外貌模式讓修正後的高動態範圍合成影像顯 13 Π1354 201121336 示於顯示器時，能顯示出與人眼所視實物色与 像，以達到色外貌一致的效果。 / ^ 非用雜朗本㈣之_及其功效，而 财發明。任何”此項賤之人切可在不違 月本赉明之精神及範疇下，對上每 _ 變。因此，本發明之權利保護範 二進订修飾與改 範圍所列。 隻㈣應、如後述之申請專利U :«』+ n”a, + „ 6,+ '.乂,+'〆2 +'〆+'7ZV+ni,0,+'々' + '10 ^3"= ni.]L'+ N3 0a'n ~ A +^2.5^+^2/+^2,71^+^2,8^^+^2,9^6+^2,10 (4) 3 n3,4^ + nisa' 2 + ni6b'2 + ni7L'a' + n3gL'b' + n3ga'b' + n310 where i, ^, a, a, and 6' represent the chromaticity values of the original color of the synthetic image, 1" , β and , then denotes the explicit value of the eigen-type of the same color in the color shift model. ' = the above method, by the color shift model The difference in the A# degree difference between the different brightness values produces a corresponding correlation, and then the desired chromaticity value is obtained by synthesizing the pixel point (four) degree value of the high dynamic range, so as to replace the color of each pixel in the original high evil range synthetic image, so that The color of the image is more in line with the color as seen by the naked eye. The high dynamic range synthetic image of the high dynamic range synthetic image of the present invention and the high dynamic range synthetic image after the color correction can be adjusted by using the CIECAM02 color appearance mode to adjust the display on the display. High dynamic range synthetic image color. Simply put, with CIECAM02 color Π 1354 12 201121336 „ Mode adjustment, which can simulate the color correction result of the human eye t color adaptation with the change of different ambient light sources, so that the image captured by the image and ambient light source after the adjustment of the color appearance mode reaches The color appearance is consistent. The method is as follows: converting the pixel RGB value of the image displayed under a certain light source to the XYZ tristimulus value as the sample tristimulus values X!, Y! and Z! in the light source environment, and simultaneously recording the white point of the image under the light source. The tristimulus values Xwl, Ywl and Zwl are calculated by the CIECAM02 color appearance mode to obtain the perceptual correlation property. Therefore, when the light source is changed, the new tristimulus values X2, Y2 and Z2 can be obtained through the CIECAM02. The color appearance mode conversion explanatory diagram is as shown in FIG. 4, wherein, under two different light sources, the CEMIAM02 color appearance mode is used to find another light source corresponding to the sample tristimulus value and the image white point tristimulus value. The tristimulus value is used as a display display adjustment, so that the human eye can view the image displayed by the display closer to the actual color viewed by the human eye. In summary, the present invention provides a method for correcting a color gamut of a high dynamic range synthetic image. Compared with the conventional disadvantages, the present invention proposes to construct a color coordinate of each color in a color space by images of different exposure levels. The color shifting module is then used to find the target chromaticity value corresponding to the same color through the color shifting module, and finally calculate the color correcting module by comparing the target chromaticity value with the primary color chromaticity value. Through the above method, the high dynamic range synthetic image can be color-adjusted, so that the high dynamic range synthetic image is more in line with the real color seen by the human eye, and the state of color distortion after the conventional high dynamic range image synthesis is reduced; Through the CIECAM02 color appearance mode, the corrected high dynamic range synthetic image display 13 Π 1354 201121336 can be displayed on the display, can display the physical color and image as seen by the human eye, in order to achieve the same color appearance. / ^ Non-use of the rare (4) _ and its efficacy, and financial invention. Anyone who has "the 贱 切 切 切 切 切 切 不 不 不 不 不 不 不 不 不 不 不 不 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此 因此Patent application described later
【圖式簡單說明】 範圍合成影像之色彩校正方 第1圖係本發明之高動態 法之步驟流程圖；[Simple diagram of the figure] Color correction method of the range synthesis image Fig. 1 is a flow chart of the steps of the high dynamic method of the present invention;
弟2a圖係第】圖所示之本發明高動態 之色彩校正方法中㈣S3詳細流程圖；u 第2b圖係第1圖所示之本發明高動態 之色彩校正方法中步驟S4詳細流程圖；。 圍合成影像 圍合成影像 第3圖係本發明高動態範 之色彩空間；以及 圍合成影像之色彩校正方法 法而以圖係經本翻高雜範圍合絲像之色彩校正方 '凡成色办校正後之高動態範圍合成氛暴、# $ CIECAM02色外貌模式轉換之說明圖。^ 【主要元件符號說明】 步驟 步驟 步驟 S1 〜S4 S31-S32 S41〜S42 111354 142a is a detailed flow chart of the high dynamic color correction method of the present invention shown in the figure. (4) S3 detailed flowchart; u 2b is a detailed flowchart of step S4 in the high dynamic color correction method of the present invention shown in FIG. 1; . The third image of the synthetic image surrounding synthetic image is the color space of the high dynamic range of the present invention; and the color correction method of the synthetic image is corrected by the color correction method of the image of the high-definition range of the image. High dynamic range synthetic atmosphere, # CIECAM02 color appearance mode conversion diagram. ^ [Main component symbol description] Step Steps Steps S1 to S4 S31-S32 S41 to S42 111354 14